Shashank Agnihotri
Computer Networks – Page 1
Queueing theory
Queueing theory is the mathematical study of waiting lines, or queues. The theory enables
mathematical analysis of several related processes, including arriving at the (back of the) queue,
waiting in the queue (essentially a storage process), and being served at the front of the queue.
The theory permits the derivation and calculation of several performance measures including the
average waiting time in the queue or the system, the expected number waiting or receiving
service, and the probability of encountering the system in certain states, such as empty, full,
having an available server or having to wait a certain time to be served.
Queueing theory has applications in diverse fields,[1] including telecommunications,[2] traffic
engineering, computing[3] and the design of factories, shops, offices and hospitals.[4]
Overview
The word queue comes, via French, from the Latin cauda, meaning tail. The spelling "queueing"
over "queuing" is typically encountered in the academic research field. In fact, one of the flagship
journals of the profession is named "Queueing Systems".
Queueing theory is generally considered a branch of operations research because the results
are often used when making business decisions about the resources needed to provide service.
It is applicable in a wide variety of situations that may be encountered in business, commerce,
industry, healthcare,[5] public service and engineering. Applications are frequently encountered
in customer service situations as well as transport and telecommunication. Queueing theory is
directly applicable to intelligent transportation systems, call
centers, PABXs,networks, telecommunications, server queueing, mainframe computer of
telecommunications terminals, advanced telecommunications systems, and traffic flow.
Notation for describing the characteristics of a queueing model was first suggested by David G.
Kendall in 1953. Kendall's notation introduced an A/B/C queueing notation that can be found in
all standard modern works on queueing theory, for example, Tijms.[6]
The A/B/C notation designates a queueing system having A as interarrival time distribution, B as
service time distribution, and C as number of servers. For example, "G/D/1" would indicate a
General (may be anything) arrival process, a Deterministic (constant time) service process and a
single server. More details on this notation are given in the article about queueing models.
History
Agner Krarup Erlang, a Danish engineer who worked for the Copenhagen Telephone Exchange,
published the first paper on queueing theory in 1909.[7]
Shashank Agnihotri
Computer Networks – Page 2
David G. Kendall introduced an A/B/C queueing notation in 1953. Important work on queueing
theory used in modern packet switchingnetworks was performed in the early 1960s by Leonard
Kleinrock.
Application to telephony
The public switched telephone network (PSTN) is designed to accommodate the offered traffic
intensity with only a small loss. Theperformance of loss systems is quantified by their grade of
service, driven by the assumption that if sufficient capacity is not available, the call is refused
and lost.[8] Alternatively, overflow systems make use of alternative routes to divert calls via
different paths — even these systems have a finite traffic carrying capacity.[8]
However, the use of queueing in PSTNs allows the systems to queue their customers' requests
until free resources become available. This means that if traffic intensity levels exceed available
capacity, customer's calls are not lost; customers instead wait until they can be served.[9] This
method is used in queueing customers for the next available operator.
A queueing discipline determines the manner in which the exchange handles calls from
customers.[9] It defines the way they will be served, the order in which they are served, and the
way in which resources are divided among the customers.[9][10] Here are details of four queueing
disciplines:
First in first out
This principle states that customers are served one at a time and that the customer that has
been waiting the longest is served first.[10]
Last in first out
This principle also serves customers one at a time, however the customer with the shortest
waiting time will be served first.[10] Also known as a stack.
Processor sharing
Customers are served equally. Network capacity is shared between customers and they all
effectively experience the same delay.[10]
Priority
Customers with high priority are served first.[10]
Shashank Agnihotri
Computer Networks – Page 3
Queueing is handled by control processes within exchanges, which can be modelled using state
equations.[9][10] Queueing systems use a particular form of state equations known as a Markov
chain that models the system in each state.[9] Incoming traffic to these systems is modelled via
a Poisson distribution and is subject to Erlang’s queueing theory assumptions viz.[8]

Pure-chance traffic – Call arrivals and departures are random and independent events.[8]

Statistical equilibrium – Probabilities within the system do not change.[8]
Full availability – All incoming traffic can be routed to any other customer within the
network.[8]


Congestion is cleared as soon as servers are free.[8]
Classic queueing theory involves complex calculations to determine waiting time, service time,
server utilization and other metrics that are used to measure queueing performance. [9][10]
Queueing networks
Networks of queues are systems which contain an arbitrary, but finite, number m of queues.
Customers, sometimes of different classes,[11]travel through the network and are served at the
nodes. The state of a network can be described by a vector
, where ki is the number
of customers at queue i. In open networks, customers can join and leave the system, whereas in
closed networks the total number of customers within the system remains fixed.
The first significant results in this area were Jackson networks, for which an efficient product
form equilibrium distribution exists and the mean value analysis which allows average metrics
such as throughput and sojourn times to be computed.[12]
Role of Poisson process, exponential distributions
A useful queueing model represents a real-life system with sufficient accuracy and is analytically
tractable. A queueing model based on thePoisson process and its companion exponential
probability distribution often meets these two requirements. A Poisson process models random
events (such as a customer arrival, a request for action from a web server, or the completion of
the actions requested of a web server) as emanating from a memoryless process. That is, the
length of the time interval from the current time to the occurrence of the next event does not
depend upon the time of occurrence of the last event. In the Poisson probability distribution, the
observer records the number of events that occur in a time interval of fixed length. In the
(negative) exponential probability distribution, the observer records the length of the time interval
between consecutive events. In both, the underlying physical process is memoryless.
Shashank Agnihotri
Computer Networks – Page 4
Models based on the Poisson process often respond to inputs from the environment in a manner
that mimics the response of the system being modeled to those same inputs. The analytically
tractable models that result yield both information about the system being modeled and the form
of their solution. Even a queueing model based on the Poisson process that does a relatively
poor job of mimicking detailed system performance can be useful. The fact that such models
often give "worst-case" scenario evaluations appeals to system designers who prefer to include
a safety factor in their designs. Also, the form of the solution of models based on the Poisson
process often provides insight into the form of the solution to a queueing problem whose detailed
behavior is poorly mimicked. As a result, queueing models are frequently modeled as Poisson
processes through the use of the exponential distribution.
Limitations of queueing theory
The assumptions of classical queueing theory may be too restrictive to be able to model realworld situations exactly. The complexity of production lines with product-specific characteristics
cannot be handled with those models. Therefore specialized tools have been developed to
simulate, analyze, visualize and optimize time dynamic queueing line behavior.
For example; the mathematical models often assume infinite numbers of customers, infinite
queue capacity, or no bounds on inter-arrival or service times, when it is quite apparent that
these bounds must exist in reality. Often, although the bounds do exist, they can be safely
ignored because the differences between the real-world and theory is not statistically significant,
as the probability that such boundary situations might occur is remote compared to the expected
normal situation. Furthermore, several studies show the robustness of queueing models outside
their assumptions. In other cases the theoretical solution may either prove intractable or
insufficiently informative to be useful.
Alternative means of analysis have thus been devised in order to provide some insight into
problems that do not fall under the scope of queueing theory, although they are often scenariospecific because they generally consist of computer simulations or analysis of experimental
data. See network traffic simulation.
Shashank Agnihotri
Computer Networks – Page 5
Birth–death process
The birth–death process is a special case of continuous-time Markov process where the states
represent the current size of a population and where the transitions are limited to births and
deaths. Birth–death processes have many applications in demography, queueing
theory,performance engineering, or in biology, for example to study the evolution of bacteria.
When a birth occurs, the process goes from state n to n + 1. When a death occurs, the process
goes from state n to state n − 1. The process is specified by birth rates
and death
rates
.
Examples
A pure birth process is a birth–death process where
A pure death process is a birth–death process where
for all
for all
A (homogeneous) Poisson process is a pure birth process where
.
.
for all
M/M/1 model and M/M/c model, both used in queueing theory, are birth–death processes used
to describe customers in an infinite queue.
Use in queueing theory
In queueing theory the birth–death process is the most fundamental example of a queueing
model, the M/M/C/K/ /FIFO (in completeKendall's notation) queue. This is a queue with
Poisson arrivals, drawn from an infinite population, and C servers with exponentially
distributedservice time with K places in the queue. Despite the assumption of an infinite
population this model is a good model for various telecommunication systems.
M/M/1 queue
The M/M/1 is a single server queue with an infinite buffer size. In a non-random environment the
birth–death process in queueing models tend to be long-term averages, so the average rate of
Shashank Agnihotri
arrival is given as and the average service time as
aM/M/1 queue when,
Computer Networks – Page 6
. The birth and death process is
The difference equations for the probability that the system is in state k at time t are,
M/M/C queue
The M/M/C is multi-server queue with C servers and an infinite buffer. This differs from
the M/M/1 queue only in the service time which now becomes,
and
with
M/M/1/K queue
The M/M/1/K queue is a single server queue with a buffer of size K. This queue has applications
in telecommunications, as well as in biology when a population has a capacity limit. In
telecommunication we again use the parameters from the M/M/1 queue with,
In biology, particularly the growth of bacteria, when the population is zero there is no ability to
grow so,
Additionally if the capacity represents a limit where the population dies from over population,
The differential equations for the probability that the system is in state k at time t are,
Shashank Agnihotri
Computer Networks – Page 7
Equilibrium
A queue is said to be in equilibrium if the limit
exists. For this to be the case,
must be zero.
Using the M/M/1 queue as an example, the steady state (equilibrium) equations are,
If
and
for all
(the homogenous case), this can be reduced to
Limit behaviour
In a small time
, only three types of transitions are possible: one death, or one birth, or no
birth nor death. If the rate of occurrences (per unit time) of births is and that for deaths is ,
then the probabilities of the above transitions are
,
, and
respectively.
For a population process, "birth" is the transition towards increasing the population by 1 while
"death" is the transition towards decreasing the population size by 1.
Shashank Agnihotri
Computer Networks – Page 8
Shashank Agnihotri
Computer Networks – Page 9
Protocol
In information technology, a protocol is the special set of rules that end points in a
telecommunication connection use when they communicate. Protocols specify interactions
between the communicating entities.
Protocols exist at several levels in a telecommunication connection. For example, there are
protocols for the data interchange at the hardware device level and protocols for data
interchange at the application program level. In the standard model known as Open Systems
Interconnection (OSI), there are one or more protocols at each layer in the telecommunication
exchange that both ends of the exchange must recognize and observe. Protocols are often
described in an industry or international standard.

Networking Tutorials and Guides

Telecom Routing and Switching

IP Telephony Systems
The TCP/IP Internet protocols, a common example, consist of:

Transmission Control Protocol (TCP), which uses a set of rules to exchange messages
with other Internet points at the information packet level

Internet Protocol (IP), which uses a set of rules to send and receive messages at the
Internet address level

Additional protocols that include the Hypertext Transfer Protocol (HTTP) and File Transfer
Protocol (FTP), each with defined sets of rules to use with corresponding programs elsewhere
on the Internet
There are many other Internet protocols, such as the Border Gateway Protocol (BGP) and the
Dynamic Host Configuration Protocol (DHCP).
The word protocol comes from the Greek protocollon, meaning a leaf of paper glued to a
manuscript volume that describes the contents.
Shashank Agnihotri
Computer Networks – Page 10
OSI model
OSI model
7. Application layer
NNTP · SIP · SSI · DNS · FTP · Gopher · HTTP · NFS · NTP · SMPP · SMTP ·SNMP ·
Telnet · DHCP · Netconf · RTP · SPDY · (more)
6. Presentation layer
MIME · XDR · TLS · SSL
5. Session layer
Named pipe · NetBIOS · SAP · PPTP ·SOCKS
4. Transport layer
TCP · UDP · SCTP · DCCP · SPX
3. Network layer
IP (IPv4, IPv6) · ICMP · IPsec · IGMP ·IPX · AppleTalk
2. Data link layer
ATM · SDLC · HDLC · ARP · CSLIP ·SLIP · GFP · PLIP · IEEE 802.2 · LLC ·L2TP · IE
EE 802.3 · Frame Relay · ITU-T G.hn DLL · PPP · X.25 · Network switch
1. Physical layer
EIA/TIA-232 · EIA/TIA-449 · ITU-T VSeries · I.430 · I.431 · POTS · PDH ·SONET/SDH · PON · OTN · DSL ·IEEE 802.3 · IE
EE 802.11 · IEEE 802.15 · IEEE 802.16 ·
Shashank Agnihotri
Computer Networks – Page 11
IEEE 1394 · ITU-T G.hn PHY · USB · Bluetooth · Hubs
The Open Systems Interconnection (OSI) model is a product of the Open Systems
Interconnection effort at the International Organization for Standardization. It is a prescription of
characterizing and standardizing the functions of a communications system in terms
ofabstraction layers. Similar communication functions are grouped into logical layers. A layer
serves the layer above it and is served by the layer below it.
For example, a layer that provides error-free communications across a network provides the
path needed by applications above it, while it calls the next lower layer to send and receive
packets that make up the contents of that path. Two instances at one layer are connected by a
horizontal connection on that layer.
History
Work on a layered model of network architecture was started and the International Organization
for Standardization (ISO) began to develop its OSI framework architecture. OSI had two major
components: an abstract model of networking, called the Basic Reference Model or seven-layer
model, and a set of specific protocols.
The concept of a seven-layer model was provided by the work ofCharles Bachman, Honeywell
Information Services. Various aspects of OSI design evolved from experiences with the
ARPANET, the fledgling Internet, NPLNET, EIN, CYCLADESnetwork and the work in IFIP
WG6.1. The new design was documented in ISO 7498 and its various addenda. In this model, a
Shashank Agnihotri
Computer Networks – Page 12
networking system was divided into layers. Within each layer, one or more entities implement its
functionality. Each entity interacted directly only with the layer immediately beneath it, and
provided facilities for use by the layer above it.
Protocols enabled an entity in one host to interact with a corresponding entity at the same layer
in another host. Service definitions abstractly described the functionality provided to an (N)-layer
by an (N-1) layer, where N was one of the seven layers of protocols operating in the local host.
The OSI standards documents are available from the ITU-T as the X.200-series of
recommendations.[1] Some of the protocol specifications were also available as part of the ITU-T
X series. The equivalent ISO and ISO/IEC standards for the OSI model were available from ISO,
but only some of them without fees.[2]
Description of OSI layers
According to recommendation X.200, there are seven layers, labeled 1 to 7, with layer 1 at the
bottom. Each layer is generically known as an N layer. An "N+1 entity" (at layer N+1) requests
services from an "N entity" (at layer N).
At each level, two entities (N-entity peers) interact by means of the N protocol by
transmitting protocol data units (PDU).
A Service Data Unit (SDU) is a specific unit of data that has been passed down from an OSI
layer to a lower layer, and which the lower layer has not yet encapsulated into a protocol data
unit (PDU). An SDU is a set of data that is sent by a user of the services of a given layer, and is
transmitted semantically unchanged to a peer service user.
The PDU at a layer N is the SDU of layer N-1. In effect the SDU is the 'payload' of a given PDU.
That is, the process of changing an SDU to a PDU, consists of an encapsulation process,
performed by the lower layer. All the data contained in the SDU becomes encapsulated within
the PDU. The layer N-1 adds headers or footers, or both, to the SDU, transforming it into a PDU
of layer N-1. The added headers or footers are part of the process used to make it possible to
get data from a source to a destination.
OSI Model
Data unit
Host
Data
layers
Layer
7. Application
Function
Network process to application
6. Presentation Data representation, encryption and decryption,
convert machine dependent data to machine
Shashank Agnihotri
Computer Networks – Page 13
independent data
Segments
5. Session
Interhost communication, managing sessions
between applications
4. Transport
End-to-end connections, reliability and flow control
Packet/Datagram 3. Network
Media
Frame
layers
Bit
Path determination andlogical addressing
2. Data link
Physical addressing
1. Physical
Media, signal and binary transmission
Some orthogonal aspects, such as management and security, involve every layer.
Security services are not related to a specific layer: they can be related by a number of layers,
as defined by ITU-T X.800 Recommendation.[3]
These services are aimed to improve the CIA triad (confidentiality, integrity, and availability) of
transmitted data. Actually the availability of communication service is determined by network
design and/or network management protocols. Appropriate choices for these are needed to
protect against denial of service.
Layer 1: physical layer
The physical layer defines electrical and physical specifications for devices. In particular, it
defines the relationship between a device and atransmission medium, such as a copper or fiber
optical cable. This includes the layout
of pins, voltages, cable specifications, hubs,repeaters, network adapters, host bus
adapters (HBA used in storage area networks) and more.
The major functions and services performed by the physical layer are:

Establishment and termination of a connection to a communications medium.

Participation in the process whereby the communication resources are effectively shared
among multiple users. For example, contentionresolution and flow control.

Modulation, or conversion between the representation of digital data in user equipment
and the corresponding signals transmitted over a communications channel. These are signals
operating over the physical cabling (such as copper and optical fiber) or over a radio link.
Shashank Agnihotri
Computer Networks – Page 14
Parallel SCSI buses operate in this layer, although it must be remembered that the
logical SCSI protocol is a transport layer protocol that runs over this bus. Various physical-layer
Ethernet standards are also in this layer; Ethernet incorporates both this layer and the data link
layer. The same applies to other local-area networks, such as token ring, FDDI, ITUT G.hn and IEEE 802.11, as well as personal area networks such as Bluetooth and IEEE
802.15.4.
Layer 2: data link layer
The data link layer provides the functional and procedural means to transfer data between
network entities and to detect and possibly correct errors that may occur in the physical layer.
Originally, this layer was intended for point-to-point and point-to-multipoint media, characteristic
of wide area media in the telephone system. Local area network architecture, which included
broadcast-capable multiaccess media, was developed independently of the ISO work in IEEE
Project 802. IEEE work assumed sublayering and management functions not required for WAN
use. In modern practice, only error detection, not flow control using sliding window, is present in
data link protocols such as Point-to-Point Protocol (PPP), and, on local area networks, the IEEE
802.2 LLC layer is not used for most protocols on the Ethernet, and on other local area
networks, its flow control and acknowledgment mechanisms are rarely used. Sliding window flow
control and acknowledgment is used at the transport layer by protocols such as TCP, but is still
used in niches where X.25 offers performance advantages.
The ITU-T G.hn standard, which provides high-speed local area networking over existing wires
(power lines, phone lines and coaxial cables), includes a complete data link layer which provides
both error correction and flow control by means of a selective repeat Sliding Window Protocol.
Both WAN and LAN service arrange bits, from the physical layer, into logical sequences called
frames. Not all physical layer bits necessarily go into frames, as some of these bits are purely
intended for physical layer functions. For example, every fifth bit of the FDDI bit stream is not
used by the layer.
WAN protocol architecture
Connection-oriented WAN data link protocols, in addition to framing, detect and may correct
errors. They are also capable of controlling the rate of transmission. A WAN data link layer might
implement a sliding window flow control and acknowledgment mechanism to provide reliable
delivery of frames; that is the case for Synchronous Data Link Control (SDLC) and HDLC, and
derivatives of HDLC such as LAPB andLAPD.
Shashank Agnihotri
Computer Networks – Page 15
IEEE 802 LAN architecture
Practical, connectionless LANs began with the pre-IEEE Ethernet specification, which is the
ancestor of IEEE 802.3. This layer manages the interaction of devices with a shared medium,
which is the function of a media access control (MAC) sublayer. Above this MAC sublayer is the
media-independent IEEE 802.2 Logical Link Control (LLC) sublayer, which deals with
addressing and multiplexing on multiaccess media.
While IEEE 802.3 is the dominant wired LAN protocol and IEEE 802.11 the wireless LAN
protocol, obsolescent MAC layers include Token Ring and FDDI. The MAC sublayer detects but
does not correct errors.
Layer 3: network layer
The network layer provides the functional and procedural means of transferring variable
length data sequences from a source host on one network to a destination host on a different
network, while maintaining the quality of service requested by the transport layer (in contrast to
the data link layer which connects hosts within the same network). The network layer performs
network routing functions, and might also perform fragmentation and reassembly, and report
delivery errors. Routers operate at this layer, sending data throughout the extended network and
making the Internet possible. This is a logical addressing scheme – values are chosen by the
network engineer. The addressing scheme is not hierarchical.
The network layer may be divided into three sublayers:
1.
Subnetwork access – that considers protocols that deal with the interface to networks,
such as X.25;
2.
Subnetwork-dependent convergence – when it is necessary to bring the level of a transit
network up to the level of networks on either side
3.
Subnetwork-independent convergence – handles transfer across multiple networks.
An example of this latter case is CLNP, or IPv7 ISO 8473. It manages
the connectionless transfer of data one hop at a time, from end system to ingress router, router
to router, and from egress router to destination end system. It is not responsible for reliable
delivery to a next hop, but only for the detection of erroneous packets so they may be discarded.
In this scheme, IPv4 and IPv6 would have to be classed with X.25 as subnet access protocols
because they carry interface addresses rather than node addresses.
A number of layer-management protocols, a function defined in the Management Annex, ISO
7498/4, belong to the network layer. These include routing protocols, multicast group
Shashank Agnihotri
Computer Networks – Page 16
management, network-layer information and error, and network-layer address assignment. It is
the function of the payload that makes these belong to the network layer, not the protocol that
carries them.
Layer 4: transport layer
The transport layer provides transparent transfer of data between end users, providing reliable
data transfer services to the upper layers. The transport layer controls the reliability of a given
link through flow control, segmentation/desegmentation, and error control. Some protocols are
state- and connection-oriented. This means that the transport layer can keep track of the
segments and retransmit those that fail. The transport layer also provides the acknowledgement
of the successful data transmission and sends the next data if no errors occurred.
OSI defines five classes of connection-mode transport protocols ranging from class 0 (which is
also known as TP0 and provides the least features) to class 4 (TP4, designed for less reliable
networks, similar to the Internet). Class 0 contains no error recovery, and was designed for use
on network layers that provide error-free connections. Class 4 is closest to TCP, although TCP
contains functions, such as the graceful close, which OSI assigns to the session layer. Also, all
OSI TP connection-mode protocol classes provide expedited data and preservation of record
boundaries. Detailed characteristics of TP0-4 classes are shown in the following table:[4]
Feature Name
TP0 TP1 TP2 TP3 TP4
Connection oriented network
Yes Yes Yes Yes Yes
Connectionless network
No No No No Yes
Concatenation and separation
No Yes Yes Yes Yes
Segmentation and reassembly
Yes Yes Yes Yes Yes
Error Recovery
No Yes Yes Yes Yes
Reinitiate connection (if an excessive number of PDUs are
unacknowledged)
No Yes No Yes No
Shashank Agnihotri
Computer Networks – Page 17
Multiplexing and demultiplexing over a single virtual circuit
No No Yes Yes Yes
Explicit flow control
No No Yes Yes Yes
Retransmission on timeout
No No No No Yes
Reliable Transport Service
No Yes No Yes Yes
Perhaps an easy way to visualize the transport layer is to compare it with a Post Office, which
deals with the dispatch and classification of mail and parcels sent. Do remember, however, that
a post office manages the outer envelope of mail. Higher layers may have the equivalent of
double envelopes, such as cryptographic presentation services that can be read by the
addressee only. Roughly speaking, tunneling protocols operate at the transport layer, such as
carrying non-IP protocols such as IBM's SNA or Novell's IPX over an IP network, or end-to-end
encryption with IPsec. While Generic Routing Encapsulation (GRE) might seem to be a networklayer protocol, if the encapsulation of the payload takes place only at endpoint, GRE becomes
closer to a transport protocol that uses IP headers but contains complete frames or packets to
deliver to an endpoint. L2TP carries PPP frames inside transport packet.
Although not developed under the OSI Reference Model and not strictly conforming to the OSI
definition of the transport layer, theTransmission Control Protocol (TCP) and the User Datagram
Protocol (UDP) of the Internet Protocol Suite are commonly categorized as layer-4 protocols
within OSI.
Layer 5: session layer
The session layer controls the dialogues (connections) between computers. It establishes,
manages and terminates the connections between the local and remote application. It provides
for full-duplex, half-duplex, or simplex operation, and establishes checkpointing, adjournment,
termination, and restart procedures. The OSI model made this layer responsible for graceful
close of sessions, which is a property of theTransmission Control Protocol, and also for session
checkpointing and recovery, which is not usually used in the Internet Protocol Suite. The session
layer is commonly implemented explicitly in application environments that use remote procedure
calls. On this level, Inter-Process_(computing) communication happen (SIGHUP, SIGKILL, End
Process, etc.).
Layer 6: presentation layer
Shashank Agnihotri
Computer Networks – Page 18
The presentation layer establishes context between application-layer entities, in which the
higher-layer entities may use different syntax and semantics if the presentation service provides
a mapping between them. If a mapping is available, presentation service data units are
encapsulated into session protocol data units, and passed down the stack.
This layer provides independence from data representation (e.g., encryption) by translating
between application and network formats. The presentation layer transforms data into the form
that the application accepts. This layer formats and encrypts data to be sent across a network. It
is sometimes called the syntax layer.[5]
The original presentation structure used the basic encoding rules of Abstract Syntax Notation
One (ASN.1), with capabilities such as converting an EBCDIC-coded text file to an ASCII-coded
file, or serialization of objects and other data structures from and to XML.
Layer 7: application layer
The application layer is the OSI layer closest to the end user, which means that both the OSI
application layer and the user interact directly with the software application. This layer interacts
with software applications that implement a communicating component. Such application
programs fall outside the scope of the OSI model. Application-layer functions typically include
identifying communication partners, determining resource availability, and synchronizing
communication. When identifying communication partners, the application layer determines the
identity and availability of communication partners for an application with data to transmit. When
determining resource availability, the application layer must decide whether sufficient network or
the requested communication exist. In synchronizing communication, all communication
between applications requires cooperation that is managed by the application layer. Some
examples of application-layer implementations also include:

On OSI stack:

FTAM File Transfer and Access Management Protocol

X.400 Mail

Common management information protocol (CMIP)

On TCP/IP stack:

Hypertext Transfer Protocol (HTTP),

File Transfer Protocol (FTP),

Simple Mail Transfer Protocol (SMTP)

Simple Network Management Protocol (SNMP).
Cross-layer functions
Shashank Agnihotri
Computer Networks – Page 19
This "datagram service model" reference in MPLS may be confusing or unclear to
readers. Please help clarify the "datagram service model" reference in MPLS;
suggestions may be found on the talk page
There are some functions or services that are not tied to a given layer, but they can affect more
than one layer. Examples include the following:

security service (telecommunication)[3] as defined by ITU-T X.800 Recommendation.

management functions, i.e. functions that permit to configure, instantiate, monitor,
terminate the communications of two or more entities: there is a specific application layer
protocol, common management information protocol (CMIP) and its corresponding
service, common management information service (CMIS), they need to interact with every layer
in order to deal with their instances.

Multiprotocol Label Switching (MPLS) operates at an OSI-model layer that is generally
considered to lie between traditional definitions of layer 2 (data link layer) and layer 3 (network
layer), and thus is often referred to as a "layer-2.5" protocol. It was designed to provide a unified
data-carrying service for both circuit-based clients and packet-switching clients which provide a
datagram service model. It can be used to carry many different kinds of traffic, including IP
packets, as well as native ATM, SONET, and Ethernet frames.

ARP is used to translate IPv4 addresses (OSI layer 3) into Ethernet MAC addresses (OSI
layer 2).
Interfaces
Neither the OSI Reference Model nor OSI protocols specify any programming interfaces, other
than as deliberately abstract service specifications. Protocol specifications precisely define the
interfaces between different computers, but the software interfaces inside computers, known
as network sockets are implementation-specific.
For example Microsoft Windows' Winsock, and Unix's Berkeley sockets and System V Transport
Layer Interface, are interfaces between applications (layer 5 and above) and the transport (layer
4). NDIS and ODI are interfaces between the media (layer 2) and the network protocol (layer 3).
Interface standards, except for the physical layer to media, are approximate implementations of
OSI service specifications.
Examples
Shashank Agnihotri
Computer Networks – Page 20
Layer
OSI protocols
#
7
6
5
4
3
TCP/IP protocols
Name
Application
Signal
ing
Syste
m 7[6]
AppleTalk
INAP,
NNTP, SIP, SSI,DNS, FTP
FTAM, X.400,X.500, D
MAP,T AFP, ZIP,RTMP,
,Gopher, HTTP,NFS, NTP,
AP,ROSE, RTSE,ACS
CAP,I
DHCP,SMPP, SMTP,SNM
E[7]CMIP[8]
SUP,T NBP
P, Telnet,RIP, BGP
UP
Presentation
ISO/IEC 8823, X.226,
MIME, SSL, TLS,XDR
ISO/IEC 9576-1, X.236
Session
ISO/IEC 8327, X.225, Sockets. Session
ISO/IEC 9548-1, X.235 establishment inTCP, RTP
Transport
ISO/IEC 8073, TP0,
TP1, TP2, TP3, TP4
TCP, UDP, SCTP,DCCP
(X.224), ISO/IEC 8602,
X.234
Network
ISO/IEC 8208,X.25 (P
LP),
ISO/IEC 8878,X.223,
ISO/IEC 84731, CLNPX.233.
IPX
SNA
UMTS
Misc. examples
RIP,
APPC
HL7, Modbus
SAP
AFP
TDI, ASCII, EBCDIC,MIDI, MPE
G
ASP,ADSP,PAP
Named
pipes, NetBIOS,SAP, half
duplex, full
duplex, simplex, RPC,SOCKS
NWLink
DDP,SPX
ATP
IP, IPsec, ICMP,IGMP, OS SCCP,
IPX
PF
MTP (TokenTalkorEther
Talk)
DLC?
NBF
RRC (Radio
Resource
Control)Pack
NBF, Q.931, IS-IS
et Data
Convergenc
e
Protocol (PD
CP)
Shashank Agnihotri
Computer Networks – Page 21
and BMC(Br
oadcast/Mult
icast
Control)
2
1
Data Link
Physical
ISO/IEC 7666,X.25 (L
APB),Token Bus,
X.222, ISO/IEC 88022 LLC Type 1 and 2[9]
X.25 (X.21bis,EIA/TIA232,EIA/TIA-449,EIA530,G.703)[9]
PPP, SBTV SLIP,PPTP
802.3 (Ethernet),802.11a/b/g/n
MAC/LLC,802.1Q
(VLAN), ATM,HDP, FDDI, Fibre
Channel, Frame
LLC (Logical
Relay,HDLC, ISL, PPP, Q.921,T
IEEE
Link
LocalTalk,
oken
MTP,
802.3frami
Control), MA
AppleTalk Remote
SDLC
Ring, CDP, NDPARP (maps
Q.710
ng,Etherne
C(Media
Access,PPP
layer 3 to layer 2 address), ITU-T
t II framing
Access
G.hn DLL
Control)
CRC, Bit stuffing, ARQ,Data
Over Cable Service Interface
Specification
(DOCSIS), interface bonding
RS-232,RSMTP, 422,STP,
Q.710
PhoneNet
UMTS
Twina
Physical
x
layer or L1
RS-232, Full
duplex,RJ45, V.35, V.34, I.430,I.
431, T1, E1, 10BASET, 100BASETX, POTS,SONET, SDH, DSL,8
02.11a/b/g/n PHY, ITU-T G.hn
PHY, Controller Area
Network, Data Over Cable
Service Interface Specification
(DOCSIS)
Shashank Agnihotri
Computer Networks – Page 22
Comparison with TCP/IP model
In the TCP/IP model of the Internet, protocols are deliberately not as rigidly designed into strict
layers as in the OSI model.[10] RFC 3439contains a section entitled "Layering considered
harmful." However, TCP/IP does recognize four broad layers of functionality which are derived
from the operating scope of their contained protocols, namely the scope of the software
application, the end-to-end transport connection, the internetworking range, and the scope of the
direct links to other nodes on the local network.
Even though the concept is different from the OSI model, these layers are nevertheless often
compared with the OSI layering scheme in the following way: The Internet application
layer includes the OSI application layer, presentation layer, and most of the session layer. Its
end-to-end transport layer includes the graceful close function of the OSI session layer as well
as the OSI transport layer. The internetworking layer (Internet layer) is a subset of the OSI
network layer (see above), while the link layer includes the OSI data link and physical layers, as
well as parts of OSI's network layer. These comparisons are based on the original seven-layer
protocol model as defined in ISO 7498, rather than refinements in such things as the internal
organization of the network layer document.
The presumably strict peer layering of the OSI model as it is usually described does not present
contradictions in TCP/IP, as it is permissible that protocol usage does not follow the hierarchy
implied in a layered model. Such examples exist in some routing protocols (e.g., OSPF), or in
the description of tunneling protocols, which provide a link layer for an application, although the
tunnel host protocol may well be a transport or even an application layer protocol in its own right.
Shashank Agnihotri
Computer Networks – Page 23
Data Link Layer (Layer 2)
The second-lowest layer (layer 2) in the OSI Reference Model stack is the data link layer, often
abbreviated “DLL” (though that abbreviation has other meanings as well in the computer world).
The data link layer, also sometimes just called the link layer, is where many wired and wireless
local area networking (LAN) technologies primarily function. For example, Ethernet, Token Ring,
FDDI and 802.11 (“wireless Ethernet” or “Wi-Fi’) are all sometimes called “data link layer
technologies”. The set of devices connected at the data link layer is what is commonly
considered a simple“network”, as opposed to an internetwork.
Data Link Layer Sublayers: Logical Link Control (LLC) and Media Access Control (MAC)
The data link layer is often conceptually divided into two sublayers: logical link control
(LLC) and media access control (MAC). This split is based on the architecture used in the IEEE
802 Project, which is the IEEE working group responsible for creating the standards that define
many networking technologies (including all of the ones I mentioned above except FDDI). By
separating LLC and MAC functions, interoperability of different network technologies is made
easier, as explained in our earlier discussion of networking model concepts.
Data Link Layer Functions
The following are the key tasks performed at the data link layer:
Logical Link Control (LLC): Logical link control refers to the functions required for the
establishment and control of logical links between local devices on a network. As mentioned
above, this is usually considered a DLL sublayer; it provides services to the network layer above
it and hides the rest of the details of the data link layer to allow different technologies to work
seamlessly with the higher layers. Most local area networking technologies use the IEEE 802.2
LLC protocol.
o
Media Access Control (MAC): This refers to the procedures used by devices to control
access to the network medium. Since many networks use a shared medium (such as a single
network cable, or a series of cables that are electrically connected into a single virtual medium) it
is necessary to have rules for managing the medium to avoid conflicts. For example. Ethernet
uses the CSMA/CD method of media access control, while Token Ring uses token passing.
o
Data Framing: The data link layer is responsible for the final encapsulation of higherlevel messages into framesthat are sent over the network at the physical layer.
o
Addressing: The data link layer is the lowest layer in the OSI model that is concerned
with addressing: labeling information with a particular destination location. Each device on a
network has a unique number, usually called ahardware address or MAC address, that is used
by the data link layer protocol to ensure that data intended for a specific machine gets to it
properly.
o
Error Detection and Handling: The data link layer handles errors that occur at the lower
levels of the network stack. For example, a cyclic redundancy check (CRC) field is often
employed to allow the station receiving data to detect if it was received correctly.
o
Shashank Agnihotri
Computer Networks – Page 24
Physical Layer Requirements Definition and Network Interconnection Device Layers
As I mentioned in the topic discussing the physical layer, that layer and the data link layer are
very closely related. The requirements for the physical layer of a network are often part of the
data link layer definition of a particular technology. Certain physical layer hardware and encoding
aspects are specified by the DLL technology being used. The best example of this is the
Ethernet standard, IEEE 802.3, which specifies not just how Ethernet works at the data link
layer, but also its various physical layers.
Since the data link layer and physical layer are so closely related, many types of hardware are
associated with the data link layer. Network interface cards (NICs) typically implement a specific
data link layer technology, so they are often called “Ethernet cards”, “Token Ring cards”, and so
on. There are also a number of network interconnection devices that are said to “operate at layer
2”, in whole or in part, because they make decisions about what to do with data they receive by
looking at data link layer frames. These devices include most bridges, switches and barters,
though the latter two also encompass functions performed by layer three.
Some of the most popular technologies and protocols generally associated with layer 2 are
Ethernet, Token Ring, FDDI (plus CDDI), HomePNA, IEEE 802.11, ATM, and TCP/IP's Serial
Link Interface Protocol (SLIP) and Point-To-Point Protocol (PPP).
Key Concept: The second OSI Reference Model layer is the data link layer. This is the
place where most LAN and wireless LAN technologies are defined. Layer two is responsible
for logical link control, media access control, hardware addressing, error detection and
handling, and defining physical layer standards. It is often divided into the logical link control
(LLC) and media access control (MAC) sublayers, based on the IEEE 802 Project that uses that
architecture.
The Data-Link layer is the protocol layer in a program that handles the moving of data in and out
across a physical link in a network. The Data-Link layer is layer 2 in the Open Systems
Interconnect (OSI) model for a set of telecommunication protocols.
The Data-Link layer contains two sublayers that are described in the IEEE-802 LAN standards:

Media Access Control (MAC)

Logical Link Control (LLC)
The Data-Link layer ensures that an initial connection has been set up, divides output data into
data frames, and handles the acknowledgements from a receiver that the data arrived
successfully. It also ensures that incoming data has been received successfully by analyzing bit
patterns at special places in the frames.
Shashank Agnihotri
Computer Networks – Page 25
Physical Layer (Layer 1)
The lowest layer of the OSI Reference Model is layer 1, the physical layer; it is commonly
abbreviated “PHY”. The physical layer is special compared to the other layers of the model,
because it is the only one where data is physically moved across the network interface. All of the
other layers perform useful functions to create messages to be sent, but they must all be
transmitted down the protocol stack to the physical layer, where they are actually sent out over
the network.
Note: The physical layer is also “special” in that it is the only layer that really does not
apply specifically to TCP/IP. Even in studying TCP/IP, however, it is still important to
understand its significance and role in relation to the other layers where TCP/IP protocols
reside.
Understanding the Role of the Physical Layer
The name “physical layer” can be a bit problematic. Because of that name, and because of what
I just said about the physical layer actually transmitting data, many people who study networking
get the impression that the physical layer is only about actual network hardware. Some people
may say the physical layer is “the network interface cards and cables”. This is not actually the
case, however. The physical layer defines a number of network functions, not just hardware
cables and cards.
A related notion is that “all network hardware belongs to the physical layer”. Again, this isn't
strictly accurate. All hardware must have some relation to the physical layer in order to send
data over the network, but hardware devices generally implement multiple layers of the OSI
model, including the physical layer but also others. For example, an Ethernet network interface
card performs functions at both the physical layer and the data link layer.
Physical Layer Functions
The following are the main responsibilities of the physical layer in the OSI Reference Model:
Definition of Hardware Specifications: The details of operation of cables, connectors,
wireless radio transceivers, network interface cards and other hardware devices are generally a
function of the physical layer (although also partially the data link layer; see below).
o
Encoding and Signaling: The physical layer is responsible for various encoding and
signaling functions that transform the data from bits that reside within a computer or other device
into signals that can be sent over the network.
o
Data Transmission and Reception: After encoding the data appropriately, the physical
layer actually transmits the data, and of course, receives it. Note that this applies equally to
wired and wireless networks, even if there is no tangible cable in a wireless network!
o
Topology and Physical Network Design: The physical layer is also considered the
domain of many hardware-related network design issues, such as LAN and WAN topology.
o
Shashank Agnihotri
Computer Networks – Page 26
In general, then, physical layer technologies are ones that are at the very lowest level and deal
with the actual ones and zeroes that are sent over the network. For example, when considering
network interconnection devices, the simplest ones operate at the physical layer: repeaters,
conventional hubs and transceivers. These devices have absolutely no knowledge of the
contents of a message. They just take input bits and send them as output. Devices like switches
and routers operate at higher layers and look at the data they receive as being more than
voltage or light pulses that represent one or zero.
Relationship Between the Physical Layer and Data Link Layer
It's important to point out that while the physical layer of a network technology primarily defines
the hardware it uses, the physical layer is closely related to the data link layer. Thus, it is not
generally possible to define hardware at the physical layer “independently” of the technology
being used at the data link layer. For example, Ethernet is a technology that describes specific
types of cables and network hardware, but the physical layer of Ethernet can only be isolated
from its data link layer aspects to a point. While Ethernet cables are “physical layer”, for
example, their maximum length is related closely to message format rules that exist at the data
link layer.
Furthermore, some technologies perform functions at the physical layer that are normally more
closely associated with the data link layer. For example, it is common to have the physical layer
perform low-level (bit level) repackaging of data link layer frames for transmission. Error
detection and correction may also be done at layer 1 in some cases. Most people would
consider these “layer two functions”.
In many technologies, a number of physical layers can be used with a data link layer. Again
here, the classic example is Ethernet, where dozens of different physical layer implementations
exist, each of which uses the same data link layer (possibly with slight variations.)
Physical Layer Sublayers
Finally, many technologies further subdivide the physical layer into sublayers. In order to
increase performance, physical layer encoding and transmission methods have become more
complex over time. The physical layer may be broken into layers to allow different network
media to be supported by the same technology, while sharing other functions at the physical
layer that are common between the various media. A good example of this is the physical layer
architecture used for Fast Ethernet, Gigabit Ethernet and 10-Gigabit Ethernet.
Note: In some contexts, the physical layer technology used to convey bits across a
network or communications line is called a transport method. Don't confuse this with the
functions of the OSI transport layer (layer 4).
Key Concept: The lowest layer in the OSI Reference Model is the physical layer. It is the
realm of networking hardware specifications, and is the place where technologies reside that
perform data encoding, signaling, transmission and reception functions. The physical layer is
closely related to the data link layer.
Shashank Agnihotri
Computer Networks – Page 27
The ALOHA protocol
Pure ALOHA
Pure ALOHA protocol. Boxes indicate frames. Shaded boxes indicate frames which have
collided.
The first version of the protocol (now called "Pure ALOHA", and the one implemented in
ALOHAnet) was quite simple:

If you have data to send, send the data

If the message collides with another transmission, try resending "later"
Note that the first step implies that Pure ALOHA does not check whether the channel is busy
before transmitting. The critical aspect is the "later" concept: the quality of the backoff scheme
chosen significantly influences the efficiency of the protocol, the ultimate channel capacity, and
the predictability of its behavior.
To assess Pure ALOHA, we need to predict its throughput, the rate of (successful) transmission
of frames. (This discussion of Pure ALOHA's performance follows Tanenbaum. [9]) First, let's
make a few simplifying assumptions:

All frames have the same length.

Stations cannot generate a frame while transmitting or trying to transmit. (That is, if a
station keeps trying to send a frame, it cannot be allowed to generate more frames to send.)

The population of stations attempts to transmit (both new frames and old frames that
collided) according to a Poisson distribution.
Let "T" refer to the time needed to transmit one frame on the channel, and let's define "frametime" as a unit of time equal to T. Let "G" refer to the mean used in the Poisson distribution over
transmission-attempt amounts: that is, on average, there are G transmission-attempts per frametime.
Shashank Agnihotri
Computer Networks – Page 28
Overlapping frames in the pure ALOHA protocol. Frame-time is equal to 1 for all frames.
Consider what needs to happen for a frame to be transmitted successfully. Let "t" refer to the
time at which we want to send a frame. We want to use the channel for one frame-time
beginning at t, and so we need all other stations to refrain from transmitting during this time.
Moreover, we need the other stations to refrain from transmitting between t-T and t as well,
because a frame sent during this interval would overlap with our frame.
For any frame-time, the probability of there being k transmission-attempts during that frame-time
is:
Comparison of Pure Aloha and Slotted Aloha shown on Throughput vs. Traffic Load plot.
The average amount of transmission-attempts for 2 consecutive frame-times is 2G. Hence, for
any pair of consecutive frame-times, the probability of there being ktransmission-attempts during
those two frame-times is:
Therefore, the probability (
) of there being zero transmission-attempts between tT and t+T (and thus of a successful transmission for us) is:
Shashank Agnihotri
Computer Networks – Page 29
The throughput can be calculated as the rate of transmission-attempts multiplied by the
probability of success, and so we can conclude that the throughput (
) is:
The maximum throughput is 0.5/e frames per frame-time (reached when G = 0.5), which is
approximately 0.184 frames per frame-time. This means that, in Pure ALOHA, only about 18.4%
of the time is used for successful transmissions.
Slotted ALOHA
Slotted ALOHA protocol. Boxes indicate frames. Shaded boxes indicate frames which are in the
same slots.
An improvement to the original ALOHA protocol was "Slotted ALOHA", which introduced discrete
timeslots and increased the maximum throughput.[10] A station can send only at the beginning of
a timeslot, and thus collisions are reduced. In this case, we only need to worry about the
transmission-attempts within 1 frame-time and not 2 consecutive frame-times, since collisions
can only occur during each timeslot. Thus, the probability of there being zero transmissionattempts in a single timeslot is:
the probability of k packets is:
The throughput is:
The maximum throughput is 1/e frames per frame-time (reached when G = 1), which is
approximately 0.368 frames per frame-time, or 36.8%.
Shashank Agnihotri
Computer Networks – Page 30
Slotted ALOHA is used in low-data-rate tactical satellite communications networks by military
forces, in subscriber-based satellite communications networks, mobile telephony call setup, and
in the contactless RFID technologies.
Other Protocols
The use of a random access channel in ALOHAnet led to the development of Carrier Sense
Multiple Access (CSMA), a 'listen before send' random access protocol which can be used when
all nodes send and receive on the same channel. The first implementation of CSMA
wasEthernet, and CSMA was extensively modeled in.[11]
ALOHA and the other random-access protocols have an inherent variability in their throughput
and delay performance characteristics. For this reason, applications which need highly
deterministic load behavior often used polling or token-passing schemes (such as token ring)
instead of contention systems. For instance ARCNET was popular in embedded data
applications in the 1980s.
Design
Network architecture
Two fundamental choices which dictated much of the ALOHAnet design were the two-channel
star configuration of the network and the use of random accessing for user transmissions.
The two-channel configuration was primarily chosen to allow for efficient transmission of the
relatively dense total traffic stream being returned to users by the central time-sharing computer.
An additional reason for the star configuration was the desire to centralize as many
communication functions as possible at the central network node (the Menehune), minimizing
the cost of the original all-hardware terminal control unit (TCU) at each user node.
The random access channel for communication between users and the Menehune was
designed specifically for the traffic characteristics of interactive computing. In a conventional
communication system a user might be assigned a portion of the channel on either a frequencydivision multiple access (FDMA) or time-division multiple access (TDMA) basis. Since it was well
known that in time-sharing systems [circa 1970], computer and user data are bursty, such fixed
assignments are generally wasteful of bandwidth because of the high peak-to-average data
rates that characterize the traffic.
To achieve a more efficient use of bandwidth for bursty traffic, ALOHAnet developed the random
access packet switching method that has come to be known as a pure ALOHA channel. This
approach effectively dynamically allocates bandwidth immediately to a user who has data to
Shashank Agnihotri
Computer Networks – Page 31
send, using the acknowledgment/retransmission mechanism described earlier to deal with
occasional access collisions. While the average channel loading must be kept below about 10%
to maintain a low collision rate, this still results in better bandwidth efficiency than when fixed
allocations are used in a bursty traffic context.
Two 100 kHz channels in the experimental UHF band were used in the implemented system,
one for the user-to-computer random access channel and one for the computer-to-user
broadcast channel. The system was configured as a star network, allowing only the central node
to receive transmissions in the random access channel. All user TCUs received each
transmission made by the central node in the broadcast channel. All transmissions were made in
bursts at 9600 bit/s, with data and control information encapsulated in packets.
Each packet consisted of a 32-bit header and a 16-bit header parity check word, followed by up
to 80 bytes of data and a 16-bit parity check word for the data. The header contained address
information identifying a particular user so that when the Menehune broadcast a packet, only the
intended user's node would accept it.
Remote units
The original user interface developed for the system was an all-hardware unit called an
ALOHAnet Terminal Control Unit (TCU), and was the sole piece of equipment necessary to
connect a terminal into the ALOHA channel. The TCU was composed of a UHF antenna,
transceiver, modem, buffer and control unit. The buffer was designed for a full line length of 80
characters, which allowed handling of both the 40 and 80 character fixed-length packets defined
for the system. The typical user terminal in the original system consisted of a Teletype Model
33 or a dumb CRT user terminal connected to the TCU using a standard RS-232C interface.
Shortly after the original ALOHA network went into operation, the TCU was redesigned with one
of the first Intel microprocessors, and the resulting upgrade was called a PCU (Programmable
Control Unit).
Additional basic functions performed by the TCU's and PCU’s were generation of a cyclic-paritycheck code vector and decoding of received packets for packet error-detection purposes, and
generation of packet retransmissions using a simple random interval generator. If an
acknowledgment was not received from the Menehune after the prescribed number of automatic
retransmissions, a flashing light was used as an indicator to the human user. Also, since the
TCU's and PCU’s did not send acknowledgments to the Menehune, a steady warning light was
displayed to the human user when an error was detected in a received packet. Thus it can be
seen that considerable simplification was incorporated into the initial design of the TCU as well
as the PCU, making use of the fact that it was interfacing a human user into the network.
Shashank Agnihotri
Computer Networks – Page 32
The Menehune
The central node communications processor was an HP 2100 minicomputer called the
Menehune, which is the Hawaiian language word for “imp”, or dwarf people,[12] and was named
for its similar role to the original ARPANET Interface Message Processor (IMP) which was being
deployed at about the same time. In the original system, the Menehune forwarded correctlyreceived user data to the UH central computer, an IBM System 360/65 time-sharing system.
Outgoing messages from the 360 were converted into packets by the Menehune, which were
queued and broadcast to the remote users at a data rate of 9600 bit/s. Unlike the half-duplex
radios at the user TCUs, the Menehune was interfaced to the radio channels with full-duplex
radio equipment.
Later developments
In later versions of the system, simple radio relays were placed in operation to connect the main
network on the island of Oahu to other islands in Hawaii, and Menehune routing capabilities
were expanded to allow user nodes to exchange packets with other user nodes, theARPANET,
and an experimental satellite network. More details are available in [3] and in the technical
reports listed in the Further Reading section below.
Shashank Agnihotri
Computer Networks – Page 33
Carrier sense multiple access
Carrier Sense Multiple Access (CSMA) is a probabilistic Media Access Control (MAC) protocol
in which a node verifies the absence of other traffic before transmitting on a shared transmission
medium, such as an electrical bus, or a band of the electromagnetic spectrum.
"Carrier Sense" describes the fact that a transmitter uses feedback from a receiver that detects
a carrier wave before trying to send. That is, it tries to detect the presence of an
encoded signal from another station before attempting to transmit. If a carrier is sensed, the
station waits for the transmission in progress to finish before initiating its own transmission. In
other words, CSMA is based on the principle "sense before transmit" or "listen before talk".
"Multiple Access" describes the fact that multiple stations send and receive on the medium.
Transmissions by one node are generally received by all other stations using the medium.
Contents
[hide]

1 Protocol modifications

2 CSMA access modes

3 References

4 See also
[edit]Protocol modifications
Carrier sense multiple access with collision detection (CSMA/CD) is a modification of CSMA.
CSMA/CD is used to improve CSMA performance by terminating transmission as soon as a
collision is detected, and reducing the probability of a second collision on retry.
Carrier sense multiple access with collision avoidance (CSMA/CA) is a modification of CSMA.
Collision avoidance is used to improve the performance of CSMA by attempting to be less
"greedy" on the channel. If the channel is sensed busy before transmission then the
transmission is deferred for a "random" interval. This reduces the probability of collisions on the
channel.
[edit]CSMA access modes
1-persistent
When the sender (station) is ready to transmit data, it checks if the physical medium is busy. If
so, it senses the medium continually until it becomes idle, and then it transmits a piece of data
(a frame). In case of a collision, the sender waits for a random period of time and attempts to
transmit again. 1-persistent CSMA is used in CSMA/CD systems including Ethernet.
Shashank Agnihotri
Computer Networks – Page 34
P-persistent
This is a sort of trade-off between 1 and non-persistent CSMA access modes. When the sender
is ready to send data, it checks continually if the medium is busy. If the medium becomes idle,
the sender transmits a frame with a probability p. If the station chooses not to transmit (the
probability of this event is 1-p), the sender waits until the next available time slot and transmits
again with the same probability p. This process repeats until the frame is sent or some other
sender starts transmitting. In the latter case the sender monitors the channel, and when idle,
transmits with a probability p, and so on. p-persistent CSMA is used in CSMA/CA systems
including WiFiand other packet radio systems.
Non-persistent
Non persistent CSMA is less aggressive compared to P persistent protocol. In this protocol,
before sending the data, the station senses the channel and if the channel is idle it starts
transmitting the data. But if the channel is busy, the station does not continuously sense it but
instead of that it waits for random amount of time and repeats the algorithm. Here the algorithm
leads to better channel utilization but also results in longer delay compared to 1 –persistent.
O-persistent
Each station is assigned a transmission order by a supervisor station. When medium goes idle,
stations wait for their time slot in accordance with their assigned transmission order. The station
assigned to transmit first transmits immediately. The station assigned to transmit second waits
one time slot (but by that time the first station has already started transmitting). Stations monitor
the medium for transmissions from other stations and update their assigned order with each
detected transmission (i.e. they move one position closer to the front of the queue).[1] Opersistent CSMA is used by CobraNet, LonWorks and the controller area network.
Shashank Agnihotri
Computer Networks – Page 35
Token bus network
Token passing in a Token bus network
Token bus is a network implementing the token ring protocol over a "virtual ring" on a coaxial
cable. A token is passed around the network nodes and only the node possessing the token may
transmit. If a node doesn't have anything to send, the token is passed on to the next node on the
virtual ring. Each node must know the address of its neighbour in the ring, so a special protocol
is needed to notify the other nodes of connections to, and disconnections from, the ring.
Token bus was standardized by IEEE standard 802.4. It is mainly used for industrial
applications. Token bus was used by GM (General Motors) for their Manufacturing Automation
Protocol (MAP) standardization effort. This is an application of the concepts used in token
ring networks. The main difference is that the endpoints of the bus do not meet to form a
physical ring. The IEEE 802.4 Working Group is disbanded. In order to guarantee the packet
delay and transmission in Token bus protocol, a modified Token bus was proposed in
Manufacturing Automation Systems and flexible manufacturing system (FMS).
Shashank Agnihotri
Computer Networks – Page 36
Token ring
Internet protocol suite
Application layer

DHCP
DHCPv6























DNS
FTP
HTTP
IMAP
IRC
LDAP
MGCP
NNTP
NTP
POP
RPC
RTP
RTSP
SIP
SMTP
SNMP
SOCKS
SSH
Telnet
TLS/SSL
XMPP
(more)
Transport layer


TCP
UDP
DCCP
SCTP

RSVP


Shashank Agnihotri




Computer Networks – Page 37
RIP
BGP
ECN
(more)
Internet layer









IP
IPv4
IPv6
ICMP
ICMPv6
IGMP
OSPF
IPsec
(more)
Link layer

ARP/InARP


NDP

Tunnels

L2TP

PPP
Media access control

Ethernet

DSL

ISDN

FDDI

(more)
Two examples of token ring networks: a) Using a single MAU b) Using several MAUs connected
to each other
Shashank Agnihotri
Computer Networks – Page 38
Token ring network
IBM hermaphroditic connector with locking clip
An IBM 8228 MAU
Madge 4/16Mbps TokenRing ISA NIC
Token ring local area network (LAN) technology is a local area network protocol which resides
at the data link layer (DLL) of the OSI model. It uses a special three-byte frame called a token
that travels around the ring. Token-possession grants the possessor permission to transmit on
the medium. Token ring frames travel completely around the loop.
Initially used only in IBM computers, it was eventually standardized with protocol IEEE 802.5.
Description
Stations on a token ring LAN are logically organized in a ring topology with data being
transmitted sequentially from one ring station to the next with a control token circulating around
Shashank Agnihotri
Computer Networks – Page 39
the ring controlling access. This token passing mechanism is shared by ARCNET, token bus,
and FDDI, and has theoretical advantages over the stochastic CSMA/CD of Ethernet.
Physically, a token ring network is wired as a star, with 'hubs' and arms out to each station and
the loop going out-and-back through each.
Cabling is generally IBM "Type-1" shielded twisted pair, with unique hermaphroditic connectors,
commonly referred to as IBM data connectors. The connectors have the disadvantage of being
quite bulky, requiring at least 3 x 3 cm panel space, and being relatively fragile.
Initially (in 1985) token ring ran at 4 Mbit/s, but in 1989 IBM introduced the first 16 Mbit/s token
ring products and the 802.5 standard was extended to support this. In 1981, Apollo
Computerintroduced their proprietary 12 Mbit/s Apollo token ring (ATR) and Proteon introduced
their 10 Mbit/s ProNet-10 token ring network in 1984. However, IBM token ring was not
compatible with ATR or ProNet-10.
Each station passes or repeats the special token frame around the ring to its nearest
downstream neighbour. This token-passing process is used to arbitrate access to the shared
ring media. Stations that have data frames to transmit must first acquire the token before they
can transmit them. Token ring LANs normally use differential Manchester encoding of bits on the
LAN media.
IBM popularized the use of token ring LANs in the mid 1980s when it released its IBM token ring
architecture based on active MAUs (Media Access Unit, not to be confused with Medium
Attachment Unit) and the IBM Structured Cabling System. The Institute of Electrical and
Electronics Engineers (IEEE) later standardized a token ring LAN system as IEEE 802.5.[1]
Token ring LAN speeds of 4 Mbit/s and 16 Mbit/s were standardized by the IEEE 802.5 working
group. An increase to 100 Mbit/s was standardized and marketed during the wane of token ring's
existence while a 1000 Mbit/s speed was actually approved in 2001, but no products were ever
brought to market.[2]
When token ring LANs were first introduced at 4 Mbit/s, there were widely circulated claims that
they were superior to Ethernet,[3] but these claims were fiercely debated.[4][5]
With the development of switched Ethernet and faster variants of Ethernet, token ring
architectures lagged behind Ethernet, and the higher sales of Ethernet allowed economies of
scale which drove down prices further, and added a compelling price advantage.
Token ring networks have since declined in usage and the standards activity has since come to
a standstill as 100Mbps switched Ethernet has dominated the LAN/layer 2 networking market.
[edit]Token frame
Shashank Agnihotri
Computer Networks – Page 40
When no station is transmitting a data frame, a special token frame circles the loop. This special
token frame is repeated from station to station until arriving at a station that needs to transmit
data. When a station needs to transmit data, it converts the token frame into a data frame for
transmission. Once the sending station receives its own data frame, it converts the frame back
into a token. If a transmission error occurs and no token frame, or more than one, is present, a
special station referred to as the Active Monitor detects the problem and removes and/or
reinserts tokens as necessary (see Active and standby monitors). On 4 Mbit/s Token Ring, only
one token may circulate; on 16 Mbit/s Token Ring, there may be multiple tokens.
The special token frame consists of three bytes as described below (J and K are special nondata characters, referred to as code violations).
Token priority
Token ring specifies an optional medium access scheme allowing a station with a high-priority
transmission to request priority access to the token.
8 priority levels, 0–7, are used. When the station wishing to transmit receives a token or data
frame with a priority less than or equal to the station's requested priority, it sets the priority bits to
its desired priority. The station does not immediately transmit; the token circulates around the
medium until it returns to the station. Upon sending and receiving its own data frame, the station
downgrades the token priority back to the original priority.
Token ring frame format
A data token ring frame is an expanded version of the token frame that is used by stations to
transmit media access control (MAC) management frames or data frames from upper layer
protocols and applications.
Token Ring and IEEE 802.5 support two basic frame types: tokens and data/command frames.
Tokens are 3 bytes in length and consist of a start delimiter, an access control byte, and an end
delimiter. Data/command frames vary in size, depending on the size of the Information field.
Data frames carry information for upper-layer protocols, while command frames contain control
information and have no data for upper-layer protocols.
Shashank Agnihotri
Computer Networks – Page 41
Data/Command Frame
SD
AC
FC
DA
SA
PDU from LLC (IEEE 802.2) CRC
8 bits 8 bits 8 bits 48 bits 48 bits up to 18200x8 bits
ED
FS
32 bits 8 bits 8 bits
Starting Delimiter
consists of a special bit pattern denoting the beginning of the frame. The bits from most
significant to least significant are J,K,0,J,K,0,0,0. J and K are code violations. Since Manchester
encoding is self clocking, and has a transition for every encoded bit 0 or 1, the J and K codings
violate this, and will be detected by the hardware.Both the Starting Delimiter and Ending
Delimiter fields are used to mark frame boundaries
J
K
0
J
K
0
0
0
1 bit 1 bit 1 bit 1 bit 1 bit 1 bit 1 bit 1 bit
Access Control
this byte field consists of the following bits from most significant to least significant bit order:
P,P,P,T,M,R,R,R. The P bits are priority bits, T is the token bit which when set specifies that this
is a token frame, M is the monitor bit which is set by the Active Monitor (AM) station when it sees
this frame, and R bits are reserved bits.
+
Bits 0–2
3
4
5–7
0
Priority
Token
Monitor
Reservation
Frame Control
a one byte field that contains bits describing the data portion of the frame contents which
indicates whether the frame contains data or control information. In control frames, this byte
specifies the type of control information.
+
Bits 0–1
0 Frame type
Bits 2–7
Control Bits
Shashank Agnihotri
Computer Networks – Page 42
Frame type – 01 indicates LLC frame IEEE 802.2 (data) and ignore control bits; 00 indicates
MAC frame and control bits indicate the type ofMAC control frame
Destination address
a six byte field used to specify the destination(s) physical address .
Source address
Contains physical address of sending station . It is six byte field that is either the local assigned
address (LAA) or universally assigned address (UAA) of the sending station adapter.
Data
a variable length field of 0 or more bytes, the maximum allowable size depending on ring speed
containing MAC management data or upper layer information.Maximum length of 4500 bytes
Frame Check Sequence
a four byte field used to store the calculation of a CRC for frame integrity verification by the
receiver.
Ending Delimiter
The counterpart to the starting delimiter, this field marks the end of the frame and consists of the
following bits from most significant to least significant: J,K,1,J,K,1,I,E. I is the intermediate frame
bit and E is the error bit.
J
K
1
J
K
1
I
E
1 bit 1 bit 1 bit 1 bit 1 bit 1 bit 1 bit 1 bit
Frame Status
a one byte field used as a primitive acknowledgement scheme on whether the frame was
recognized and copied by its intended receiver.
A
C
0
0
A
C
0
0
1 bit 1 bit 1 bit 1 bit 1 bit 1 bit 1 bit 1 bit
Shashank Agnihotri
Computer Networks – Page 43
A = 1 , Address recognized C = 1 , Frame copied
Token Frame
Start Delimiter Access Control End Delimiter
8 bits
8 bits
8 bits
Abort Frame
SD
ED
8 bits 8 bits
Used to abort transmission by the sending station
Active and standby monitors
Every station in a token ring network is either an active monitor (AM) or standby monitor (SM)
station. However, there can be only one active monitor on a ring at a time. The active monitor is
chosen through an election or monitor contention process.
The monitor contention process is initiated when

a loss of signal on the ring is detected.

an active monitor station is not detected by other stations on the ring.

a particular timer on an end station expires such as the case when a station hasn't seen a
token frame in the past 7 seconds.
When any of the above conditions take place and a station decides that a new monitor is
needed, it will transmit a "claim token" frame, announcing that it wants to become the new
monitor. If that token returns back to the sender, it is OK for it to become the monitor. If some
other station tries to become the monitor at the same time then the station with the highest MAC
address will win the election process. Every other station becomes a standby monitor. All
stations must be capable of becoming an active monitor station if necessary.The active monitor
performs a number of ring administration functions. The first function is to operate as the master
Shashank Agnihotri
Computer Networks – Page 44
clock for the ring in order to provide synchronization of the signal for stations on the wire.
Another function of the AM is to insert a 24-bit delay into the ring, to ensure that there is always
sufficient buffering in the ring for the token to circulate. A third function for the AM is to ensure
that exactly one token circulates whenever there is no frame being transmitted, and to detect a
broken ring. Lastly, the AM is responsible for removing circulating frames from the ring
.Token
ring insertion process
Token ring stations must go through a 5-phase ring insertion process before being allowed to
participate in the ring network. If any of these phases fail, the token ring station will not insert into
the ring and the token ring driver may report an error.
Phase 0 (Lobe Check) — A station first performs a lobe media check. A station
is wrapped at the MSAU and is able to send 2000 test frames down its transmit pair which will

loop back to its receive pair. The station checks to ensure it can receive these frames without
error.
Phase 1 (Physical Insertion) — A station then sends a 5 volt signal to the MSAU to open
the relay.


Phase 2 (Address Verification) — A station then transmits MAC frames with its own MAC
address in the destination address field of a token ring frame. When the frame returns and if the
Address Recognized (AR) and Frame Copied (FC) bits in the frame-status are set to 0
(indicating that no other station currently on the ring uses that address), the station must
participate in the periodic (every 7 seconds) ring poll process. This is where stations identify
themselves on the network as part of the MAC management functions.
Phase 3 (Participation in ring poll) — A station learns the address of its Nearest Active
Upstream Neighbour (NAUN) and makes its address known to its nearest downstream

neighbour, leading to the creation of the ring map. Station waits until it receives an AMP or SMP
frame with the AR and FC bits set to 0. When it does, the station flips both bits (AR and FC) to 1,
if enough resources are available, and queues an SMP frame for transmission. If no such frames
are received within 18 seconds, then the station reports a failure to open and de-inserts from the
ring. If the station successfully participates in a ring poll, it proceeds into the final phase of
insertion, request initialization.
Phase 4 (Request Initialization) — Finally a station sends out a special request to a
parameter server to obtain configuration information. This frame is sent to a special functional
address, typically a token ring bridge, which may hold timer and ring number information the new
station needs to know.

Shashank Agnihotri
Computer Networks – Page 45
IEEE 802
IEEE 802 refers to a family of IEEE standards dealing with local area networks and metropolitan
area networks.
More specifically, the IEEE 802 standards are restricted to networks carrying variable-size
packets. (By contrast, in cell relay networks data is transmitted in short, uniformly sized units
called cells. Isochronous networks, where data is transmitted as a steady stream of octets, or
groups of octets, at regular time intervals, are also out of the scope of this standard.) The
number 802 was simply the next free number IEEE could assign,[1] though “802” is sometimes
associated with the date the first meeting was held — February 1980.
The services and protocols specified in IEEE 802 map to the lower two layers (Data Link and
Physical) of the seven-layer OSI networking reference model. In fact, IEEE 802 splits the OSI
Data Link Layer into two sub-layers named Logical Link Control (LLC) and Media Access
Control (MAC), so that the layers can be listed like this:

Data link layer

LLC Sublayer

MAC Sublayer

Physical layer
The IEEE 802 family of standards is maintained by the IEEE 802 LAN/MAN Standards
Committee (LMSC). The most widely used standards are for the Ethernet family, Token Ring,
Wireless LAN, Bridging and Virtual Bridged LANs. An individual Working Group provides the
focus for each area.
Working groups
Name
Description
IEEE 802.1
Bridging (networking) and Network Management
IEEE 802.2
LLC
IEEE 802.3
Ethernet
Note
inactive
Shashank Agnihotri
Computer Networks – Page 46
IEEE 802.4
Token bus
disbanded
IEEE 802.5
Defines the MAC layer for a Token Ring
inactive
IEEE 802.6
MANs
disbanded
IEEE 802.7
Broadband LAN using Coaxial Cable
disbanded
IEEE 802.8
Fiber Optic TAG
disbanded
IEEE 802.9
Integrated Services LAN
disbanded
IEEE 802.10
Interoperable LAN Security
disbanded
IEEE 802.11
a/b/g/n
Wireless LAN (WLAN) & Mesh (Wi-Fi certification)
IEEE 802.12
100BaseVG
IEEE 802.13
Unused[2]
IEEE 802.14
Cable modems
IEEE 802.15
Wireless PAN
IEEE 802.15.1
Bluetooth certification
IEEE 802.15.2
IEEE 802.15 and IEEE 802.11 coexistence
IEEE 802.15.3
High-Rate wireless PAN
disbanded
disbanded
Shashank Agnihotri
Computer Networks – Page 47
IEEE 802.15.4
Low-Rate wireless PAN (e.g., ZigBee, WirelessHART, MiWi,
etc.)
IEEE 802.15.5
Mesh networking for WPAN
IEEE 802.16
Broadband Wireless Access (WiMAX certification)
IEEE 802.16.1
Local Multipoint Distribution Service
IEEE 802.17
Resilient packet ring
IEEE 802.18
Radio Regulatory TAG
IEEE 802.19
Coexistence TAG
IEEE 802.20
Mobile Broadband Wireless Access
IEEE 802.21
Media Independent Handoff
IEEE 802.22
Wireless Regional Area Network
IEEE 802.23
Emergency Services Working Group
New (March,
2010)
Shashank Agnihotri
Computer Networks – Page 48
Medium Access Sublayer
Unit-3 Medium Access Sublayer
Structure:
3.0 Objectives
3.1 LAN and WAN
3.2 ALOHA Protcols
3.3 LAN Protocols
3.4 IEEE 802 Standards for LANs
3.5 Fiber Optic Networks
3.6 Summary
3.7 Self Assessment Questions
3.8 Terminal Questions
3.9 Answers to Self Assessment Questions
3.10 Answers to Terminal Questions
3.0 Objectives
This unit provides the reader the necessary theory for understanding the Medium Access (MAC)
sublayer of the data link layer.
After completion of this unit you will be able to:
· Define LAN and MAN
· Describe the channel allocation mechanisms used in various LANs and MANs
· Describe ALOHA protocols
· Compare and Contrast various LAN protocols
· Explain various IEEE standards for LANs
3.1 LAN and WAN
Shashank Agnihotri
Computer Networks – Page 49
i) Static Channel Allocation in LAN and MAN
ii) Dynamic Channel Allocation in LAN and MAN
As the data link layer is overloaded, it is split into MAC and LLC sub layers. MAC sub-layer is
the bottom part of the data link layer. Medium access control is often used as a synonym
to multiple access protocol, since the MAC sub layer provides the protocol and control
mechanisms that are required for a certain channel access method. This unit deals with broadcast
networks and their protocols.
In any broadcast network, the key issue is how to determine who gets to use the channel when
there is a competition. When only one single channel is available, determining who should get
access to the channel for transmission is a very complex task. Many protocols for solving the
problem are known and they form the contents of this unit.
Thus unit provides an insight of a channel access control mechanism that makes it possible for
several terminals or network nodes to communicate within a multipoint network. The MAC layer
is essentially important in local area networks(LAN’s), many of which use a multi-access channel
as the basis for communication. WAN’s in contrast use a point to point networks.
To get a head start, let us define LANs and MANs.
Definition: A Local Area Network (LAN) is a network of systems spread over small geographical
area, for example a network of computers within a building or small campus.
The owner of a LAN may be the same organization within which the LAN network is set up. It has
higher data rates i.e. in scales of Mbps (Rates at which the data are transferred from one system to
another) because the systems to be spanned are very close to each other in proximity.
Definition: A WAN (Wide Area Network) typically spans a set of countries that have data rates
less than 1Mbps, because of the distance criteria.
The LANs may be owned by multiple organizations since the spanned distance is spread over
some countries.
i) Static Channel Allocation in LAN and MAN
Before going for the exact theory behind the methods of channel allocations, we need to
understand the base behind this theory, which is given below:
The channel allocation problem
We can classify the channels as static and dynamic. The static channel is where the number of
users are stable and the traffic is not bursty. When the number of users using the channel keeps on
varying the channel is considered as a dynamic channel. The traffic on these dynamic channels
also keeps on varying. For example: In most computer systems, the data traffic is extremely
Shashank Agnihotri
Computer Networks – Page 50
bursty. We see that in this system, the peak traffic to mean traffic ratios of 1000:1 are common.
· Static channel allocation
The usual way of allocating a single channel among the multiple users is frequency division
multiplexing (FDM). If there are N users, the bandwidth allocated is split into N equal sized
portions. FDM is simple and efficient technique for small number of users. However when the
number of senders is large and continuously varying or the traffic is bursty, FDM is not suitable.
The same arguments that apply to FDM also apply to TDM. Thus none of the static channels
allocation methods work well with bursty traffic we explore the dynamic channels.
· Dynamic channels allocation in LAN’s and MAN’s
Before discussing the channel allocation problems that is multiple access methods we will see the
assumptions that we are using so that the analysis will become simple.
Assumptions:
1. The Station Model:
The model consists of N users or independent stations. Stations are sometimes called terminals.
The probability of frame being generated in an interval of length ∆t is λ.Δt, where λ is a constant
and defines the arrival rate of new frames. Once the frame has been generated, the station is
blocked and does nothing until the frame has been successfully transmitted.
2. Single Channel Assumption:
A single channel is available for all communication. All stations can transmit using this single
channel. All can receive from this channel. As far as the hard is concerned, all stations are
equivalent. It is possible the software or the protocols used may assign the priorities to them.
3. Collisions:
If two frames are transmitted simultaneously, they overlap in time and the resulting signal is
distorted or garbled. This event or situation is called a collision. We assume that all stations can
detect collisions. A collided frame must be retransmitted again later. Here we consider no other
errors for retransmission other than those generated because of collisions.
4. Continuous Time
For a continuous time assumption we mean, that the frame transmission on the channel can begin
any instant of time. There is no master clock dividing the time into discrete intervals.
5. Slotted Time
In case of slotted time assumption, the time is divided into discrete slots or intervals. The frame
Shashank Agnihotri
Computer Networks – Page 51
transmission on the channel begins only at the start of a slot. A slot may contain 0, 1, or more
frames. The 0 frame transmission corresponds to idle slot, 1 frame transmission corresponds to
successful transmission, and more frame transmission corresponds to a collision.
6. Carrier Sense
Using this facility the users can sense the channel. i.e. the stations can tell if the channel is in use
before trying to use it. If the channel is sensed as busy, no station will attempt to transmit on the
channel unless and until it goes idle.
7. No Carrier Sense:
This assumption implies that this facility is not available to the stations. i.e. the stations cannot tell
if the channel is in use before trying to use it. They just go ahead and transmit. It is only after
transmission of the frame they determine whether the transmission was successful or not.
The first assumption states that the station is independent and work is generated at a constant rate.
It also assumes that each station has only one program or one user. Thus when the station is
blocked no new work is generated. The single channel assumption is the heart of this station model
and this unit. The collision assumption is also very basic. Two alternate assumptions about time
are discussed. For a given system only one assumption about time holds good, i.e. either the
channel is considered to be continuous time based or slotted time based. Also a channel can be
sensed or not sensed by the stations. Generally LANs can sense the channel but wireless networks
cannot sense the channel effectively. Also stations on wired carrier sense networks can terminate
their transmission prematurely if they discover collision. But in case of wireless networks collision
detection is rarely done.
3.2 ALOHA Protocols
In 1970s, Norman Abramson and his colleagues at University of Hawaii devised a new and elegant
method to solve the channel allocation problem. Their work has been extended by many
researchers since then. His work is called the ALOHA system which uses ground-based radio
broadcasting. This basic idea is applicable to any system in which uncoordinated users are
competing for the use of a shared channel.
Pure or Un-slotted Aloha
The ALOHA network was created at the University of Hawaii in 1970 under the leadership of
Norman Abramson. The Aloha protocol is an OSI layer 2 protocol for LAN networks
with broadcast topology.
The first version of the protocol was basic:
· If you have data to send, send the data
· If the message collides with another transmission, try resending it later
Shashank Agnihotri
Computer Networks – Page 52
Figure 3.1: Pure ALOHA
Figure 3.2: Vulnerable period for the node: frame
The Aloha protocol is an OSI layer 2 protocol used for LAN. A user is assumed to be always in
two states: typing or waiting. The station transmits a frame and checks the channel to see if it was
successful. If so the user sees the reply and continues to type. If the frame transmission is not
successful, the user waits and retransmits the frame over and over until it has been successfully
sent.
Let the frame time denote the amount of time needed to transmit the standard fixed length frame.
We assume the there are infinite users and generate the new frames according Poisson distribution
with the mean N frames per frame time.
· If N>1 the users are generating the frames at higher rate than the channel can handle. Hence all
frames will suffer collision.
· Hence the range for N is
0<N<1
· If N>1 there are collisions and hence retransmission frames are also added with the new frames
for transmissions.
Let us consider the probability of k transmission attempts per frame time. Here the transmission of
Shashank Agnihotri
Computer Networks – Page 53
frames includes the new frames as well as the frames that are given for retransmission. This total
traffic is also poisoned with the mean G per frame time. That is G ≥ N
· At low load: N is approximately =0, there will be few collisions. Hence few retransmissions that
is G=N
· At high load: N >>1, many retransmissions and hence G>N.
· Under all loads: throughput S is just the offered load G times the probability of successful
transmission P0
S = G*P0
The probability that k frames are generated during a given frame time is given by Poisson
distribution
P[k]= Gke-G / K!
So the probability of zero frames is just e-G. The basic throughput calculation follows a Poisson
distribution with an average number of arrivals of 2G arrivals per two frame time. Therefore, the
lambda parameter in the Poisson distribution becomes 2G.
Hence P0 = e-2G
Hence the throughput S = GP0 = Ge-2G
We get for G = 0.5 resulting in a maximum throughput of 0.184, i.e. 18.4%.
Pure Aloha had a maximum throughput of about 18.4%. This means that about 81.6% of the total
available bandwidth was essentially wasted due to losses from packet collisions.
Slotted or Impure ALOHA
An improvement to the original Aloha protocol was Slotted Aloha. It is in 1972 Roberts published
a method to double the throughput of a pure ALOHA by using discrete time-slots. His proposal
was to divide the time into discrete slots corresponding to one frame time. This approach requires
the users to agree to the frame boundaries. To achieve synchronization one special station emits a
pip at the start of each interval similar to a clock. Thus the capacity of slotted ALOHA increased
to the maximum throughput of 36.8%.
The throughput for pure and slotted ALOHA system is shown in figure 3.3. A station can send
only at the beginning of a timeslot and thus collisions are reduced. In this case, the average
number of aggregate arrivals is G arrivals per 2X seconds. This leverages the lambda parameter to
be G. The maximum throughput is reached for G = 1.
Shashank Agnihotri
Computer Networks – Page 54
Figure 3.3: Throughput versus offered load traffic
With Slotted Aloha, a centralized clock sends out small clock tick packets to the outlying stations.
Outlying stations are allowed to send their packets immediately after receiving a clock tick. If
there is only one station with a packet to send, this guarantees that there will never be a collision
for that packet. On the other hand if there are two stations with packets to send, this algorithm
guarantees that there will be a collision, and the whole of the slot period up to the next clock tick is
wasted. With some mathematics, it is possible to demonstrate that this protocol does improve the
overall channel utilization, by reducing the probability of collisions by a half.
It should be noted that Aloha’s characteristics are still not much different from those experienced
today by Wi - Fi, and similar contention-based systems that have no carrier sense capability. There
is a certain amount of inherent inefficiency in these systems. It is typical to see these types of
networks’ throughput break down significantly as the number of users and message burstiness
increase. For these reasons, applications which need highly deterministic load behavior often use
token-passing schemes (such as token ring) instead of contention systems.
For instance ARCNET is very popular in embedded applications. Nonetheless, contention based
systems also have significant advantages, including ease of management and speed in initial
communication. Slotted Aloha is used on low bandwidth tacticalSatellite
communications networks by the US Military, subscriber based Satellite communications
networks, and contact lessRFID technologies.
3.3 LAN Protocols
With slotted ALOHA, the best channel utilization that can be achieved is 1 / e. This is hardly
surprising since with stations transmitting at will, without paying attention to what other stations
are doing, there are bound to be many collisions. In LANs, it is possible to detect what other
stations are doing, and adapt their behavior accordingly. These networks can achieve a better
utilization than 1 / e.
CSMA Protocols:
Protocols in which stations listen for a carrier (a transmission) and act accordingly are
called Carrier Sense Protocols."Multiple Access" describes the fact that multiple nodes send
and receive on the medium. Transmissions by one node are generally received by all other nodes
using the medium. Carrier Sense Multiple Access (CSMA) is a probabilistic Media Access
Shashank Agnihotri
Computer Networks – Page 55
Control (MAC) protocol in which a node verifies the absence of other traffic before transmitting
on a shared physical medium, such as an electrical bus, or a band of electromagnetic spectrum.
The following three protocols discuss the various implementations of the above discussed
concepts:
i) Protocol 1. 1-persistent CSMA:
When a station has data to send, it first listens to the channel to see if any one else is transmitting.
If the channel is busy, the station waits until it becomes idle. When the station detects an idle
channel, it transmits a frame. If a collision occurs, the station waits a random amount of time and
starts retransmission.
The protocol is so called because the station transmits with a probability of a whenever it finds the
channel idle.
ii) Protocol 2. Non-persistent CSMA:
In this protocol, a conscious attempt is made to be less greedy than in the
1-persistent CSMA protocol. Before sending a station senses the channel. If no one else is sending,
the station begins doing so itself. However, if the channel is already in use, the station does not
continuously sense the channel for the purpose of seizing it immediately upon detecting the end of
previous transmission. Instead, it waits for a random period of time and then repeats the algorithm.
Intuitively, this algorithm should lead to better channel utilization and longer delays than 1persistent CSMA.
iii) Protocol 3. p - persistent CSMA
It applies to slotted channels and the working of this protocol is given below:
When a station becomes ready to send, it senses the channel. If it is idle, it transmits with a
probability p. With a probability of q = 1 – p, it defers until the next slot. If that slot is also idle, it
either transmits or defers again, with probabilities p and q. This process is repeated until either the
frame has been transmitted or another station has begun transmitting. In the latter case, it acts as if
there had been a collision. If the station initially senses the channel busy, it waits until the next slot
and applies the above algorithm.
CSMA/CD Protocol
In computer networking, Carrier Sense Multiple Access with Collision Detection (CSMA/CD) is a
network control protocol in which a carrier sensing scheme is used. A transmitting data station that
detects another signal while transmitting a frame, stops transmitting that frame, transmits a jam
signal, and then waits for a random time interval. The random time interval also known as
"backoff delay" is determined using the truncated binary exponential backoff algorithm. This
delay is used before trying to send that frame again. CSMA/CD is a modification of pure Carrier
Shashank Agnihotri
Computer Networks – Page 56
Sense Multiple Access (CSMA).
Collision detection is used to improve CSMA performance by terminating transmission as soon as
a collision is detected, and reducing the probability of a second collision on retry. Methods for
collision detection are media dependent, but on an electrical bus such as Ethernet, collisions can be
detected by comparing transmitted data with received data. If they differ, another transmitter is
overlaying the first transmitter’s signal (a collision), and transmission terminates immediately.
Here the collision recovery algorithm is nothing but an binary exponential algorithm that
determines the waiting time for retransmission. If the number of collisions for the frame hits 16
then the frame is considered as not recoverable.
CSMA/CD can be in anyone of the following three states as shown in figure 3.4.
1. Contention period
2. transmission period
3. Idle period
Figure 3.4: States of CSMA / CD: Contention, Transmission, or Idle
A jam signal is sent which will cause all transmitters to back off by random intervals, reducing the
probability of a collision when the first retry is attempted. CSMA/CD is a layer 2 protocol in
the OSI model. Ethernet is the classic CSMA/CD protocol.
Collision Free Protocols
Although collisions do not occur with CSMA/CD once a station has unambiguously seized the
channel, they can still occur during the contention period. These collisions adversely affect the
system performance especially when the cable is long and the frames are short. And also
CSMA/CD is not universally applicable. In this section, we examine some protocols that resolve
the contention for the channel without any collisions at all, not even during the contention period.
In the protocols to be described, we assume that there exists exactly N stations, each with a unique
address from 0 to N-1 “wired” into it. We assume that the propagation delay is negligible.
Shashank Agnihotri
Computer Networks – Page 57
i) A Bit Map Protocol
In this method, each contention period consists of exactly N slots. If station 0 has a frame to send,
it transmits a 1 bit during the zeroth slot. No other station is allowed to transmit during this slot.
Regardless of what station 0 is doing, station 1 gets the opportunity to transmit a 1 during slot 1,
but only if it has a frame queued. In general, station j may announce that it has a frame to send by
inserting a 1 bit into slot j. After all N stations have passed by, each station has complete
knowledge of which stations wish to transmit. At that point, they begin transmitting in a numerical
order.
Since everyone agrees on who goes next, there will never be any collisions. After the last ready
station has transmitted its frame, an event all stations can monitor, another N bit contention period
is begun. If a station becomes ready just after its bit slot has passed by, it is out of luck and must
remain silent until every station has had a chance and the bit map has come around again.
Protocols like this in which the desire to transmit is broadcast before the actual transmission are
called Reservation Protocols.
ii) Binary Countdown
A problem with basic bit map protocol is that the overhead is 1 bit per station, so it odes not scale
well to networks with thousands of stations. We can do better by using binary station address.
A station wanting to use the channel now broadcasts its address as a binary bit string, starting with
the high-order bit. All addresses are assumed to be of the same length. The bits in each address
position from different stations are Boolean ORed together. We call this protocol as Binary
countdown, which was used in Datakit. It implicitly assumes that the transmission delays are
negligible so that all stations see asserted bits essentially simultaneously.
To avoid conflicts, an arbitration rule must be applied: as soon as a station sees that a high-order
bit position that is 0 in its address has been overwritten with a 1, it gives up.
Example: If stations 0010, 0100, 1001, and 1010 are all trying to get the channel for transmission,
these are ORed together to form a 1. Stations 0010 and 0100 see the 1 and know that a higher
numbered station is competing for the channel, so they give up for the current round. Stations 1001
and 1010 continue.
The next bit is 0, and both stations continue. The next bit is 1, so station 1001 gives up. The
winner is station 1010 because t has the highest address. After winning the bidding, it may now
transmit a frame, after which another bidding cycle starts.
This protocol has the property that higher numbered stations have a higher priority than lower
numbered stations, which may be either good or bad depending on the context.
iii) Limited Contention Protocols
Until now we have considered two basic strategies for channel acquisition in a cable network:
Contention as in CSMA, and collision – free methods. Each strategy can be rated as to how well it
Shashank Agnihotri
Computer Networks – Page 58
does with respect to the two important performance measures, delay at low load, and channel
efficiency at high load.
Under conditions of light load, contention (i.e. pure or slotted ALOHA) is preferable due to its low
delay. As the load increases, contention becomes increasingly less attractive, because the overhead
associated with channel arbitration becomes greater. Just the reverse is true for collision free
protocols. At low load, they have high delay, but as the load increases, the channel efficiency
improves.
It would be more beneficial if we could combine the best features of contention and collision free
protocols and arrive at a protocol that uses contention at low load to provide low delay, but uses a
collision free technique at high load to provide good channel efficiency. Such protocols can be
called Limited Contention protocols.
iv) Adaptive Tree Walk Protocol
A simple way of performing the necessary channel assignment is to use the algorithm devised by
US army for testing soldiers for syphilis during World War II. The Army took a blood sample
from N soldiers. A portion of each sample was poured into a single test tube. This mixed sample
was then tested for antibodies. If none were found, all the soldiers in the group were declared
healthy. If antibodies were present, two new mixed samples were prepared, one from soldiers 1
through N/2 and one from the rest. The process was repeated recursively until the infected soldiers
were detected.
For the computerized version of this algorithm, let us assume that stations are arranged as the
leaves of a binary tree as shown in figure 3.4 below:
Figure 3.5: A tree for four stations
In the first contention slot following a successful frame transmission, slot 0, all stations are
permitted to acquire the channel. If one of them does, so fine. If there is a collision, then during
slot 1 only stations falling under node 2 in the tree may compete. If one of them acquires the
channel, the slot following the frame is reserved for those stations under node 3. If on the other
hand, two or more stations under node 2 want to transmit, there will be a collisions during slot 1,
Shashank Agnihotri
Computer Networks – Page 59
in which case it is node 4’s turn during slot 2.
In essence, if a collision occurs during slot 0, the entire tree is searched, depth first to locate all
ready stations. Each bit slot is associated with some particular node in a tree. If a collision occurs,
the search continues recursively with the node’s left and right children. If a bit slot is idle or if only
one station transmits in it, the searching of its node can stop because all ready stations have been
located.
When the load on the system is heavy, it is hardly worth the effort to dedicate slot 0 to node 1,
because that makes sense only in the unlikely event that precisely one station has a frame to send.
At what level in the tree should the search begin? Clearly, the heavier the load, the farther down
the tree the search should begin.
3.4 IEEE 802 standards for LANs
IEEE has standardized a number of LAN’s and MAN’s under the name of IEEE 802. Few of the
standards are listed in figure 3.6. The most important of the survivor’s are 802.3 (Ethernet) and
802.11 (wireless LAN). Both these two standards have different physical layers and different
MAC sub layers but converge on the same logical link control sub layer so they have same
interface to the network layer.
IEEE No
802.3
802.4
802.5
802.6
802.11
802.15.1
802.15.4
802.16
Name
Ethernet
WiFi
Bluetooth
ZigBee
WiMa
Title
CSMA/CD Networks (Ethernet)
Token Bus Networks
Token Ring Networks
Metropolitan Area Networks
Wireless Local Area Networks
Wireless Personal Area Networks
Wireless Sensor Networks
Wireless Metropolitan Area Networks
Figure 3.6: List of IEEE 802 Standards for LAN and MAN
Ethernets
Ethernet was originally based on the idea of computers communicating over a shared coaxial cable
acting as a broadcast transmission medium. The methods used show some similarities to radio
systems, although there are major differences, such as the fact that it is much easier to detect
collisions in a cable broadcast system than a radio broadcast. The common cable providing the
communication channel was likened to the ether and it was from this reference the name
"Ethernet" was derived.
From this early and comparatively simple concept, Ethernet evolved into the complex networking
technology that today powers the vast majority of local computer networks. The coaxial cable was
Shashank Agnihotri
Computer Networks – Page 60
later replaced with point-to-point links connected together by hubs and/or switches in order to
reduce installation costs, increase reliability, and enable point-to-point management and
troubleshooting. Star LAN was the first step in the evolution of Ethernet from a coaxial cable bus
to a hub-managed, twisted-pair network.
Above the physical layer, Ethernet stations communicate by sending each other data packets, small
blocks of data that are individually sent and delivered. As with other IEEE 802 LANs, each
Ethernet station is given a single 48-bit MAC address, which is used both to specify the
destination and the source of each data packet. Network interface cards (NICs) or chips normally
do not accept packets addressed to other Ethernet stations. Adapters generally come programmed
with a globally unique address, but this can be overridden, either to avoid an address change when
an adapter is replaced, or to use locally administered addresses.
The most kinds of Ethernets used were with the data rate of 10Mbps. The table 3.1 gives the
details of the medium used, number of nodes per segment and distance it supported, along with the
application.
Table 3.1 Different 10Mbps Ethernets used
Name
10Base5
10Base2
10Base-T
10Base-F
Cable
Type
Max
Segment
Length
Thick coax 500 m
Nodes per Advantages
Segment
Thin coax
Twisted
Pair
Fiber
Optics
185 m
100 m
30
1024
2000 m
1024
100
Original Cable;
Now Obsolete
No hub needed
Cheapest system
Best between
buildings
Fast Ethernet
Fast Ethernet is a collective term for a number of Ethernet standards that carry traffic at the
nominal rate of 100 Mbit/s, against the original Ethernet speed of 10 Mbit/s. Of the 100 megabit
Ethernet standards 100baseTX is by far the most common and is supported by the vast majority of
Ethernet hardware currently produced. Full duplex fast Ethernet is sometimes referred to as "200
Mbit/s" though this is somewhat misleading as that level of improvement will only be achieved if
traffic patterns are symmetrical. Fast Ethernet was introduced in 1995 and remained the fastest
version of Ethernet for three years before being superseded by gigabit Ethernet.
A fast Ethernet adaptor can be logically divided into a medium access controller (MAC) which
deals with the higher level issues of medium availability and a physical layer interface (PHY). The
MAC may be linked to the PHY by a 4 bit 25 MHz synchronous parallel interface known as MII.
Repeaters (hubs) are also allowed and connect to multiple PHYs for their different interfaces.
Shashank Agnihotri
Computer Networks – Page 61
· 100BASE-T is any of several Fast Ethernet standards for twisted pair cables.
· 100BASE-TX (100 Mbit/s over two-pair Cat5 or better cable),
· 100BASE-T4 (100 Mbit/s over four-pair Cat3 or better cable, defunct),
· 100BASE-T2 (100 Mbit/s over two-pair Cat3 or better cable, also defunct).
The segment length for a 100BASE-T cable is limited to 100 meters. Most networks had to be
rewired for 100-megabit speed whether or not they had supposedly been CAT3 or CAT5 cable
plants. The vast majority of common implementations or installations of 100BASE-T are done
with 100BASE-TX.
100BASE-TX is the predominant form of Fast Ethernet, and runs over two pairs of category 5 or
above cable. A typical category 5 cable contains 4 pairs and can therefore support two 100BASETX links. Each network segment can have a maximum distance of 100 metres. In its typical
configuration, 100BASE-TX uses one pair of twisted wires in each direction, providing 100 Mbit/s
of throughput in each direction (full-duplex).
The configuration of 100BASE-TX networks is very similar to 10BASE-T. When used to build a
local area network, the devices on the network are typically connected to a hub or switch, creating
a star network. Alternatively it is possible to connect two devices directly using a crossover cable.
In 100BASE-T2, the data is transmitted over two copper pairs, 4 bits per symbol. First, a 4 bit
symbol is expanded into two 3-bit symbols through a non-trivial scrambling procedure based on a
linear feedback shift register.
100BASE-FX is a version of Fast Ethernet over optical fiber. It uses two strands of multi-mode
optical fiber for receive (RX) and transmit (TX). Maximum length is 400 metres for halfduplex connections or 2 kilometers for full-duplex.
100BASE-SX is a version of Fast Ethernet over optical fiber. It uses two strands of multi-mode
optical fiber for receive and transmit. It is a lower cost alternative to using 100BASE-FX, because
it uses short wavelength optics which are significantly less expensive than the long wavelength
optics used in 100BASE-FX. 100BASE-SX can operate at distances up to 300 meters.
100BASE-BX is a version of Fast Ethernet over a single strand of optical fiber (unlike 100BASEFX, which uses a pair of fibers). Single-mode fiber is used, along with a special multiplexer which
splits the signal into transmit and receive wavelengths.
Gigabit Ethernet
Gigabit Ethernet (GbE or 1 GigE) is a term describing various technologies for
transmitting Ethernet packets at a rate of a gigabit per second, as defined by the IEEE 802.32005 standard. Half duplex gigabit links connected through hubs are allowed by the specification
but in the marketplace full duplex with switches is the norm.
Gigabit Ethernet was the next iteration, increasing the speed to 1000 Mbit/s. The initial standard
Shashank Agnihotri
Computer Networks – Page 62
for gigabit Ethernet was standardized by the IEEE in June 1998 as IEEE 802.3z. 802.3z is
commonly referred to as 1000BASE-X (where -X refers to either -CX, -SX, -LX, or -ZX).
IEEE 802.3ab, ratified in 1999, defines gigabit Ethernet transmission over unshielded twisted pair
(UTP) category 5, 5e, or 6cabling and became known as 1000BASE-T. With the ratification of
802.3ab, gigabit Ethernet became a desktop technology as organizations could utilize their existing
copper cabling infrastructure.
Initially, gigabit Ethernet was deployed in high-capacity backbone network links (for instance, on
a high-capacity campus network). Fiber gigabit Ethernet has recently been overtaken by 10 gigabit
Ethernet which was ratified by the IEEE in 2002 and provided data rates 10 times that of gigabit
Ethernet. Work on copper 10 gigabit Ethernet over twisted pair has been completed, but as of July
2006, the only currently available adapters for 10 gigabit Ethernet over copper requires specialized
cabling.
InfiniBand connectors and is limited to 15 m. However, the 10GBASE-T standard specifies use of
the traditional RJ-45connectors and longer maximum cable length. Different gigabits Ethernet are
listed in table 3.2.
Table 3.2 Different Gigabit Ethernets
Name
1000BASE-T
1000BASE-SX
1000BASE-LX
1000BASE-CX
1000BASE-ZX
medium
unshielded twisted pair
multi-mode fiber
single-mode fiber
balanced copper cabling
single-mode fiber
IEEE 802.3 Frame format
Preamble SOF Destination Source
Address
Address
Length Data Pad Checksum
Figure 3.7: Frame format of IEEE 802.3
· Preamble field
Each frame starts with a preamble of 8 bytes, each containing bit patterns “10101010”. Preamble
is encoded using Manchester encoding. Thus the bit patterns produce a 10MHz square wave for
6.4 micro sec to allow the receiver’s clock to synchronize with the sender’s clock.
· Address field
The frame contains two addresses, one for the destination and another for the sender. The length of
address field is 6 bytes. The MSB of destination address is ‘0’ for ordinary addresses and ‘1’ for
group addresses. Group addresses allow multiple stations to listen to a single address. When a
Shashank Agnihotri
Computer Networks – Page 63
frame is sent to a group of users, all stations in that group receive it. This type of transmission is
referred to as multicasting. The address consisting of all ‘1’ bits is reserved for broadcasting.
· SOF: This field is 1 byte long and is used to indicate the start of the frame.
· Length:
This field is of 2 bytes long. It is used to specify the length of the data in terms of bytes that is
present in the frame. Thus the combination of the SOF and the length field is used to mark the end
of the frame.
· Data :
The length of this field ranges from zero to a maximum of 1500 bytes. This is the place where the
actual message bits are to be placed.
· Pad:
When a transceiver detects a collision, it truncates the current frame, which means the stray bits
and pieces of frames appear on the cable all the time. To make it easier to distinguish valid frames
from garbage, Ethernet specifies that valid frame must be at least 64 bytes long, from the
destination address to the checksum, including both. That means the data field come must be of 46
bytes. But if there is no data to be transmitted and only some acknowledgement is to be
transmitted then the length of the frame is less than what is specified for the valid frame. Hence
these pad fields are provided, i.e. if the data field is less than 46 bytes then the pad field comes into
picture such that the total data and pad field must be equal to 46 bytes minimum. If the data field is
greater than 46 bytes then pad field is not used.
· Checksum:
It is 4 byte long. It uses a 32-bit hash code of the data. If some data bits are in error, then the
checksum will be wrong and the error will be detected. It uses CRC method and it is used only for
error detection and not for forward error correction.
IEEE 802.4 Standard - Token Bus
This standard was proposed by Dirvin and Miller in 1986.
In this standard, physically the token bus is a linear or tree-shaped cable onto which the stations
are attached. Logically, the stations are organized into a ring, with each station knowing the
address of the station to its “left” or “right”. When the logical ring is initialized, the highest
numbered station may send the first frame. After it is done, it passes permission to its immediate
neighbor by sending the neighbor a special control frame called a token. The token propagates
around the logical ring, with only the token holder being permitted to transmit frames. Since only
one station at a time holds the token, collisions do not occur.
Shashank Agnihotri
Computer Networks – Page 64
Note: The physical order in which the stations are connected to the cable is not important.
Since the cable is inherently a broadcast medium, each station receives each frame, discarding
those not addressed to it. When a station passes the token, it sends a token frame specifically
addressed to its logical neighbor in the ring, irrespective of where the station is physically located
on the cable.
Figure 3.8: Token Passing
IEEE 802.5 Standard - Token Ring
A ring is really not a broadcast medium, but a collection of individual point-to-point links that
happen to form a circle. Ring engineering is almost entirely digital. A ring is also fair and has a
known upper bound on channel access.
sec, each bit occupies 200/R meters on the ring. This means, for example, that a 1-Mbps ring
whose circumference is 1000 meters can contain only 5 bits on it at once. sec. With a typical
propagation speed of about 200 m/A major issue in the design and analysis of any ring network is
the “physical length” of a bit. If the data rate of the ring is R Mbps, a bit is emitted every 1/R
A ring really consists of a collection of ring interfaces connected by point-to-point lines. Each bit
arriving at an interface is copied into a 1-bit buffer and then copied out onto the ring again. While
in the buffer, the bit can be inspected and possibly modified before being written out. This copying
step introduces a 1-bit delay at each interface.
In a token ring a special bit pattern, called the token, circulates around the ring whenever all
stations are idle. When a station wants to transmit a frame, it is required to seize the token and
remove it from the ring before transmitting. Since there is only one token, only one station can
transmit at a given instant, thus solving the channel access problem the same way the token bus
Shashank Agnihotri
Computer Networks – Page 65
solves it.
3.5 Fiber Optic Networks
Fiber optics is becoming increasingly important, not only for wide area point-to-point links, but
also for MANs and LANs. Fiber has high bandwidth, is thin and lightweight, is not affected by
electromagnetic interference from heavy machinery, power surges or lightning, and has excellent
security because it is nearly impossible to wiretap without detection.
FDDI (Fiber Distributed Data Interface)
It is a high performance fiber optic token ring LAN running at 100 Mbps over distances up to 200
km with up to 1000 stations connected. It can be used in the same way as any of the 802 LANs,
but with its high bandwidth, another common use is as a backbone to connect copper LANs.
FDDI – II is a successor of FDDI modified to handle synchronous circuit switched PCM data for
voice or ISDN traffic, in addition to ordinary data.
FDDI uses multimode fibers. It also uses LEDs rather than lasers because FDDI may sometimes
be used to connect directly to workstations.
The FDDI cabling consists of two fiber rings, one transmitting clockwise and the other
transmitting counter clockwise. If any one breaks, the other can be used as a backup.
FDDI defines two classes of stations A and B. Class A stations connect to both rings. The cheaper
class B stations only connect to one of the rings. Depending on how important fault tolerance is,
an installation can choose class A or class B stations, or some of each.
S/NET
It is another kind of fiber optic network with an active star for switching. It was designed and
implemented at Bell laboratories. The goal of S/NET is very fast switching.
Each computer in the network has two 20-Mbps fibers running to the switch, one for input and one
for output. The fibers terminate in a BIB (Bus Interface Board). The CPUs each have an I/O
device register that acts like a one-word window into BIB memory. When a word is written to that
device register, the interface board in the CPU transmits the bits serially over the fiber to the BIB,
where they are reassembled as a word in BIB memory. When the whole frame to be transmitted
has been copied to BIB memory, the CPU writes a command to another I/O device register to
cause the switch to copy the frame to the memory of the destination BIB and interrupt the
destination CPU.
Access to this bus is done by a priority algorithm. Each BIB has a unique priority. When a BIB
wants access to the bus it asserts a signal on the bus corresponding to its priority. The requests are
recorded and granted in priority order, with one word transferred (16 bits in parallel) at a time.
When all requests have been granted, another round of bidding is started and BIBs can again
request the bus. No bus cycles are lost to contention, so switching speed is 16 bits every 200 nsec,
Shashank Agnihotri
Computer Networks – Page 66
or 80 Mbps.
3.6 Summary
This unit discusses the Medium Access Sublayer. It discusses in detail about LANs and WANs. It
discusses the basic LAN protocols called ALOHA protocols. It describes the IEEE 802 standards
for LANS. It discusses the importance of Fiber Optic Networks and cabling used as backbone for
LAN connectivity.
3.7 Self Assessment Questions
1. The Data Link Layer of the ISO OSI model is divided into ______ sublayers
a) 1 b) 4 c) 3 d) 2
2. The ______ layer is responsible for resolving access to the shared media or resources.
a) physical b) MAC sublayer c) Network d) Transport
3. A WAN typically spans a set of countries that have data rates less than _______ Mbps
a) 2 b) 1 c) 4 d) 100
4. The ________ model consists of N users or independent stations.
5. The Aloha protocol is an OSI _______ protocol for LAN networks with broadcast topology
6. In ______ method, each contention period consists of exactly N slots
3.8 Terminal Questions
1. Discuss ALOHA protocols
2. Discuss various LAN protocols
3. Discuss IEEE 802 standards for LANs
3.9 Answers to Self Assessment Questions
1. d
2. b
3. b
4. Station
Shashank Agnihotri
Computer Networks – Page 67
5. layer 2
6. A Bit Map Protocol
3.10 Answers to Terminal Questions
1. Refer to section 3.2
2. Refer to section 3.3
3. Refer to section 3.4
ATM Protocol Structure
Figure 33 shows the ATM layered architecture as described in ITU-T recommendation I.321
(1992). This is the basis on which the B-ISDN Protocol Reference Model has been defined.
Figure 33: ATM Protocol Architecture

ATM Physical Layer
The physical layer accepts or delivers payload cells at its point of access to the ATM layer. It
provides for cell delineation which enables the receiver to recover cell boundaries. It generates
and verifies the HEC field. If the HEC cannot be verified or corrected, then the physical layer will
discard the errored cell. Idle cells are inserted in the transmit direction and removed in the
receiving direction.
For the physical transmission of bits, 5 types of transmission frame adaptations are specified (by
the ITU and the ATM Forum). Each one of them has its own lower bound or upper bound for the
amount of bits it can carry (from 12.5 Mbps to 10 Gbps so far).
1.
Synchronous Digital Hierarchy (SDH)
155 Mbps;
2.
Plesiochronous Digital Hierarchy (PDH)
3.
4.
Cell Based
155 Mbps;
Fibre Distributed Data Interface (FDDI) = 100 Mbps;
5.
Synchronous Optical Network (SONET)
34 Mbps;
51 Mbps.
Shashank Agnihotri
Computer Networks – Page 68
The actual physical link could be either optical or coaxial with the possibility of Unshielded
Twisted Pair (UTP Category 3/5) and Shielded Twisted Pair (STP Category 5) in the mid range
(12.5 to 51 Mbps).

ATM Layer
ATM layer mainly performs switching, routing and multiplexing. The characteristic features of the
ATM layer are independent of the physical medium. Four functions of this layer have been
identified.
1.
cell multiplexing (in the transmit direction)
2.
cell demultiplexing (at the receiving end)
3.
VPI/VCI translation
4.
cell header generation/extraction.
This layer accepts or delivers cell payloads. It adds appropriate ATM cell headers when
transmitting and removes cell headers in the receiving direction so that only the cell information
field is delivered to the ATM Adaptation Layer.
At the ATM switching/cross connect nodes VPI and VCI translation occurs. At a VC switch new
values of VPI and VCI are obtained whereas at a VP switch only new values for the VPI field are
obtained (see Figure 34). Depending on the direction, either the individual VPs and VCs are
multiplexed into a single cell or the single cell is demultiplexed to get the individual VPs and
VCs.
Figure 34: VC/VP Switching in ATM

ATM Adaptation Layer (AAL)
The ATM Adaptation Layer (AAL) is between ATM layer and the higher layers. Its basic function
is the enhanced adaptation of services provided by the ATM layer to the requirements of the
higher layers.
This layer accepts and delivers data streams that are structured for use with user's own
communication protocol. It changes these protocol data structures into ATM cell payloads when
receiving and does the reverse when transmitting. It inserts timing information required by users
into cell payloads or extracts from them. This is done in accordance with five AAL service
classes defined as follows.
1.
AAL1 - Adaptation for Constant Bit Rate (CBR) services (connection oriented, 47 byte
payload);
2.
AAL2 - Adaptation for Variable Bit Rate (VBR) services (connection oriented, 45 byte
payload);
3.
AAL3 - Adaptation for Variable Bit Rate data services (connection oriented, 44 byte
payload);
4.
AAL4 - Adaptation for Variable Bit Rate data services (connection less, 44 byte payload);
5.
AAL5 - Adaptation for signalling and data services (48 byte payload).
Shashank Agnihotri
Computer Networks – Page 69
In the case of transfer of information in real time, AAL1 and AAL2 which support connection
oriented services are important. AAL4 which supports a connection less service was originally
meant for data which is sensitive to loss but not to delay. However, the introduction of AAL5
which uses a 48 byte payload with no overheads, has made AAL3/4 redundant. Frame Relay
and MPEG -2 (Moving Pictures Expert Group) video are two services which will specifically use
AAL5.
ATM Services

CBR Service
This supports the transfer of information between the source and destination at a constant bit
rate. CBR service uses AAL1. A typical example is the transfer of voice at 64 Kbps over ATM.
Another usage is for the transport of fixed rate video.
This type of service over an ATM network is sometimes called circuit emulation (similar to a
voice circuit on a telephone network).

VBR Service
This service is useful for sources with variable bit rates. Typical examples are variable bit rate
audio and video.

ABR and UBR Services
The definition of CBR and VBR has resulted in two other service types called Available Bit Rate
(ABR) services and Unspecified Bit Rate (UBR) services.
ABR services use the instantaneous bandwidth available after allocating bandwidths for CBR
and VBR services. This makes the bandwidth of the ABR service to be variable. Although there
is no guaranteed time of delivery for the data transported using ABR services, the integrity of
data is guaranteed. This is ideal to carry time insensitive (but loss sensitive) data such as in
LAN-LAN interconnect and IP over ATM.
UBR service, as the name implies, has an unspecified bit rate which the network can use to
transport information relating to network management, monitoring, etc.
EXAMPLE NETWORKS – connection-oriented networks: X.25, Frame Relay and ATM
by admin under Computer Networks
2 connection-oriented networks: X.25, Frame Relay and ATM
Since the beginning of connectivity arose a war between those who support subnets no
connection-oriented (ie, datagrams) and supporters of the subnets oriented the connection.
The main supporters of the subnets are connection oriented in community ARPANET / Internet.
Remember that the original desire of the DoD to establish and build ARPANET was to have a
network that could function even after multiple impacts of weapons destroy nuclear numerous
routers and transmission lines.
Therefore, tolerance errors was on his list of priorities, not so much they could charge
customers.
This approach led to a design not connection oriented where each packet is routed
independently of any other package.
Therefore, if some routers fall during a session, no harm because the system can reconfigure
itself dynamically for subsequent packets to find a route to your destination, even if it is different
from the one used by the previous packages.
Shashank Agnihotri
Computer Networks – Page 70
The connection-oriented field comes from the world of telephone companies.
In the telephone system, the caller must dial the number of the party to call and wait connection
before you can talk or send data.
This establishes a connection setup route through the phone system that is maintained until the
call is terminated.
All words or packets follow the same route.
If a line or switch goes down along the way, the call is canceled. This property is precisely what
the DoD did not like.
So why do you like the phone companies? For two reasons:
1. Quality service.
2. Billing.
By first establishing a connection, the subnet can reserve resources such as space buffering and
processing power (CPU) in routers.
Attempting to set a call and available resources are insufficient, the call is rejected and the caller
receives a busy signal.
once a connection is established, it gives good service.
With no network connection-oriented, if too many packets arrive at the same router to same
time, the router is saturated and may lose some packets.Perhaps the issuer notice this and send
it back, but the quality of service is uneven and inadequate to audio or video unless the network
has low.
Needless to say, to provide adequate audio quality is something the phone companies take
great care, hence the preference for the connections.
The second reason that phone companies prefer the connection-oriented service is that they are
accustomed to charging for connection time.
When a call long distance (domestic or international) is charged per minute.
When they arrived networks were drawn precisely to a model in which the charge per minute
would be easy to do.
If you have to establish a connection before sending the data, that is when billing clock starts
running. If no connection, no charge.
Ironically, maintaining billing records is very expensive.
If a telephone company adopt a flat monthly rate with unlimited calls and no billing or
maintenance of a record, probably save a lot of money, despite the increase in calls this policy
would generate.
Shashank Agnihotri
Computer Networks – Page 71
However, there are policies, regulations and other factors that weigh in against doing this.
Interestingly, the flat rate service exists in other sectors. For example, cable TV is billed at a flat
monthly rate regardless of how many programs display.
It could have been designed with pay per view as a basic concept, but it was not, in part by the
high cost of turnover (and given the quality of most television programs, shame can not be
discounted entirely).
Also, many parks charge an admission fee per day with unlimited access to games, in contrast
to carnivals that charge per game.
That said, we should not be surprising that all the networks designed by the telephone industry
have been connection-oriented subnets.
What is surprising is that the Internet is also inclined in that direction in order to provide better
audio and video service.
For now examine some connection-oriented networks.
X.25 and Frame Relay
Our first example of connection-oriented network is X.25, which was the first network public
data.
Deployed in the 1970′s, when telephone service was a monopoly everywhere and the telephone
company in each country expected to have a data network country itself.
To use X.25, a computer, first established a connection to the remote computer, that is, made a
phone call.
This connection was a connection number for use in the transfer of data packets (since it could
open many connections at the same time).
Data sets were very simple, consisted of a 3-byte header and up to 128 bytes of data.
The header consists of a number of 12-bit connection, a packet sequence number, a receipt
confirmation number and some number of bits.
X.25 networks operated for almost ten years with mixed results.
In the 1980′s, X.25 networks were replaced largely by a new type of network called Frame
Relay.
This is a connection-oriented network without error control or flow.
As might be connection-oriented, packets are delivered in order (if to surrender all).
Shashank Agnihotri
Computer Networks – Page 72
The properties of order delivery, no error control or flow made the Frame Relay LAN-like wide
area.
Its most important application is the interconnection LANs in multiple locations of a company.
Frame Relay enjoyed modest success and even is still used in some parts.
Asynchronous Transfer Mode
Another type of network connection-oriented, perhaps more importantly, ATM (Asynchronous
Transfer Mode).
The reason for this strange name is because the telephone system in most of the transmission is
synchronous (the closest thing to a clock) and ATM do not.
ATM was designed in the early 1990s and launched in the midst of an incredible exaggeration
(Ginsburg, 1996; Goralski, 1995; Ibe, 1997; kimnaras et al., 1994, and Stallings, 2000).
ATM would solve all the problems of merging telecommunications connectivity and voice,
data, cable, telex, telegraph, carrier pigeons, cans connected by string, drums, smoke signals
and everything else in a single integrated system that could provide all services for all
needs. That did not happen.
In large part, the problems were similar to those described in the issue of OSI, ie an unwelcome
appearance, along with technology, implementation and misguided policies.
Having knocked out telephone companies in the first assault, much of the Internet community
saw ATM as when the Internet was the opponent of the telcos: the second part.
But not so in reality and this time even in datagrams compromisers fans realized that the quality
of Internet service left much to be desired.
To make a long story, was much more successful ATM OSI and currently use deep within the
telephone system, often in the transport of IP packets.
As businesses today is used primarily to carry its internal transport, users are unaware of its
existence, but definitely alive and has health.
ATM virtual circuits
Since ATM networks are connection oriented, data transmission requires that you first send a
package to connect.
Establishment as the message continues its path through the subnet, all switches are in the path
created an entry in its internal tables noting the existence of the connection and reserving
whatever resources they need the connection.
Shashank Agnihotri
Computer Networks – Page 73
Often the connections are called virtual circuits, in analogy with the physical circuits used in the
telephone system.
Most ATM networks also support permanent virtual circuits, the connections standing between
two hosts (distant). They are similar to leased lines in the world phone.
Each connection, temporarily or permanently, has a single connection handle.
Once connected, each side can begin transmitting data.
The basic idea that is based ATM is to transmit all the information into small packets of fixed
size, called cells.
The cells have a size of 53 bytes, five of which are header and 48 payload, so the sender and
receiver hosts and all intermediate switches can know which cells belong to which connections.
This information allows each switch knows how to send each incoming cell.
The switching of cells is done in hardware at high speed.
In fact, the main argument for having fixed-size cells is that it is easy build hardware switches to
handle small cells of fixed length.
Packages variable length IP must be routed through software, which is a slower process.
Another advantage of ATM is that the hardware can be configured to send an incoming cell to
multiple output lines, a property necessary for the management of a television program to be
broadcast to multiple receivers.
Finally, the small cells do not block any long line, making it easier to guarantee quality of service.
All cells follow the same route to the destination.
The delivery of cells is not guaranteed, but the order itself. If the cells 1 and 2 are sent in that
order, then they should arrive in the same order, never first 2 then 1.
However, one or both of these may be lost in the way.
At higher levels of their proper protocols cell recover losses.
Note that while this warranty is not perfect, is better than the Internet.
There, the packets only lost, but also delivered in disarray.
ATM, in contrast, ensures that cells were never delivered disorder.
ATM networks are organized as traditional WAN, with lines and switches (routers).
Shashank Agnihotri
Computer Networks – Page 74
The most common speeds for ATM networks are 155 and 622 Mbps, although higher speeds
are supported.Speed was chosen because it is 155 Mbps which required to transmit highdefinition television.
The exact choice of 155.52 Mbps was made for compatibility with the SONET transmission
system from AT & T.
The physical layer is concerned with the physical environment: voltages, bit timing and other
aspects.
ATM does not prescribe a particular set of rules, only specifies that the cells ATM can be sent as
is cable or fiber, but can be packed into the payload of other transport systems.
In other words, ATM is designed to be independent of the transmission medium.
The ATM layer is responsible for the cells and transport.
Defines the layout of a cell and indicates the mean fields of the header.
It also deals with the establishment and the release of the virtual circuits. Congestion control is
also located here.
Since most applications do not need to work directly with the cells (although some may do so),
has defined a top layer to the ATM layer to users send packets bigger than a cell.
ATM interface segments these packets, transmitted from individual cells and reassembles at the
other end.
This layer is AAL (ATM Adaptation Layer).
Unlike the first two-dimensional reference models, the ATM model is defined as if three
dimensional.
The user plane
deals with data transport, flow control, error correction and other user functions.
In contrast, the control plane deals with connection management.
The management functions of the layer plane and are related to resource management
and coordination between coats.
Each of the physical layer and AAL are divided into two subnetworks, one in the bottom
that does the job and the convergence sublayer on top that provides the interface own
immediate upper layer.
The PMD sublayer (Physical Medium Dependent) interacts with the real cable.
Shashank Agnihotri
Computer Networks – Page 75
Move bit in and out and handles the timing bits, ie the time between each bit to
transmission.
This layer will be different for different carriers and cables.
The other sublayer of the physical layer is the sublayer TC (Transmission Convergence).
When cells are transmitted, the TC layer sends a string of bits to the PMD layer.
This is simple. At the other extreme, the TC sublayer receives a series of input bits of the PMD
sublayer.
Its job is to convert this bit stream into a stream of cells to the ATM layer.
Manages all aspects related to the indications of where the cells begin and end the flow of bits.
In the ATM model, this feature occurs in the physical layer.
In the OSI model and large of the other networks, the work of framing, ie convert a number of
bits in the rough a sequence of frames or cells, is the task of the data link layer.
As mentioned earlier, the ATM layer handles cells, including their generation and transportation.
The most interesting aspect of ATM is located here. It is a combination of the data link layer and
network OSI model, there is a split in sublayers.
The AAL layer is divided into a sublayer SAR (Segmentation and Reassembly) and CS
(Convergence Sublayer).
The lower sublayer packet fragmented cells in the transmit side and rejoins at the destination.
The upper sublayer allows ATM systems offer different types of services to different applications
(eg, file transfer and video on demand have different requirements concerning error handling,
timing, etc.).
However, since there is a substantial installed base, it is likely that he was still in use for
some years.
Shashank Agnihotri
Computer Networks – Page 76
Permanent and switched virtual circuits in ATM, frame relay,
and X.25
Switched virtual circuits (SVCs) are generally set up on a per-call basis and are disconnected
when the call is terminated; however, apermanent virtual circuit (PVC) can be established as
an option to provide a dedicated circuit link between two facilities. PVC configuration is usually
preconfigured by the service provider. Unlike SVCs, PVC are usually very seldom
broken/disconnected.
A switched virtual circuit (SVC) is a virtual circuit that is dynamically established on demand and
is torn down when transmission is complete, for example after a phone call or a file download.
SVCs are used in situations where data transmission is sporadic and/or not always between the
same data terminal equipment (DTE) endpoints.
A permanent virtual circuit (PVC) is a virtual circuit established for repeated/continuous use
between the same DTE. In a PVC, the long-term association is identical to the
data transfer phase of a virtual call. Permanent virtual circuits eliminate the need for repeated
call set-up andclearing.

Frame relay is typically used to provide PVCs.

ATM provides both switched virtual connections and permanent virtual connections,
as they are called in ATM terminology.

X.25 provides both virtual calls and PVCs, although not all X.25 service providers or
DTE implementations support PVCs as their use was much less common than SVCs
X.25
X.25 is packet-switched network based WAN protocol for WAN communications. It
delineates data exchange and control of information within a user appliance, Data
Terminal Equipment (DTE) and a network node, Data Circuit Terminating Equipment
(DCE). X.25 comprises physical links such as packet-switching exchange (PSE)
nodes for networking hardware, leased lines, and telephone or ISDN connections.
Its unique functionality is its capacity to work effectively on any type of system that
is connected to the network. X.25, although replaced by superior technology,
continues to be in use. It utilizes a connection-oriented service that enables data
packets to be transmitted in an orderly manner.
Shashank Agnihotri
Computer Networks – Page 77
Network congestion
In data networking and queueing theory, network congestion occurs when a link or node is
carrying so much data that its quality of servicedeteriorates. Typical effects include queueing
delay, packet loss or the blocking of new connections. A consequence of these latter two is that
incremental increases in offered load lead either only to small increases in network throughput,
or to an actual reduction in network throughput.
Network protocols which use aggressive retransmissions to compensate for packet loss tend to
keep systems in a state of network congestion even after the initial load has been reduced to a
level which would not normally have induced network congestion. Thus, networks using these
protocols can exhibit two stable states under the same level of load. The stable state with low
throughput is known as congestive collapse.
Modern networks use congestion control and network congestion avoidance techniques to try to
avoid congestion collapse. These include: exponential backoff in protocols such
as 802.11's CSMA/CA and the original Ethernet, window reduction in TCP, and fair queueing in
devices such as routers. Another method to avoid the negative effects of network congestion is
implementing priority schemes, so that some packets are transmitted with higher priority than
others. Priority schemes do not solve network congestion by themselves, but they help to
alleviate the effects of congestion for some services. An example of this is 802.1p. A third
method to avoid network congestion is the explicit allocation of network resources to specific
flows. One example of this is the use of Contention-Free Transmission Opportunities
(CFTXOPs) in the ITU-TG.hn standard, which provides high-speed (up to 1 Gbit/s) Local area
networking over existing home wires (power lines, phone lines and coaxial cables).
RFC 2914 addresses the subject of congestion control in detail.
Network capacity
The fundamental problem is that all network resources are limited, including router processing
time and link throughput.
For example:

Today's (2006) Wireless LAN effective bandwidth throughput (15-100Mbit/s) is easily
filled by a single personal computer.

Even on fast computer networks (e.g. 1 Gbit), the backbone can easily be congested by a
few servers and client PCs.
Shashank Agnihotri
Computer Networks – Page 78

Because P2P scales very well, file transmissions by P2P have no problem filling and will
fill an uplink or some other network bottleneck, particularly when nearby peers are preferred over
distant peers.

Denial of service attacks by botnets are capable of filling even the largest Internet
backbone network links (40 Gbit/s as of 2007), generating large-scale network congestion
Congestive collapse
Congestive collapse (or congestion collapse) is a condition which a packet
switched computer network can reach, when little or no useful communication is happening due
to congestion. Congestion collapse generally occurs at choke points in the network, where the
total incoming traffic to a node exceeds the outgoing bandwidth. Connection points between
a local area network and a wide area network are the most likely choke points.
When a network is in such a condition, it has settled (under overload) into a stable state where
traffic demand is high but little usefulthroughput is available, and there are high levels
of packet delay and loss (caused by routers discarding packets because their output queuesare
too full) and general quality of service is extremely poor.
History
Congestion collapse was identified as a possible problem as far back as 1984 (RFC 896, dated
6 January). It was first observed on the early Internet in October 1986, when the NSFnet phase-I
backbone dropped three orders of magnitude from its capacity of 32 kbit/s to 40 bit/s, and this
continued to occur until end nodes started implementing Van Jacobson's congestion
control between 1987 and 1988.
Cause
When more packets were sent than could be handled by intermediate routers, the intermediate
routers discarded many packets, expecting the end points of the network to retransmit the
information. However, early TCP implementations had very bad retransmission behavior. When
this packet loss occurred, the end points sent extra packets that repeated the information lost;
doubling the data rate sent, exactly the opposite of what should be done during congestion. This
pushed the entire network into a 'congestion collapse' where most packets were lost and the
resultant throughput was negligible.
Congestion control
Shashank Agnihotri
Computer Networks – Page 79
Congestion control concerns controlling traffic entry into a telecommunications network, so as
to avoid congestive collapse by attempting to avoid oversubscription of any of the processing
or link capabilities of the intermediate nodes and networks and taking resource reducing steps,
such as reducing the rate of sending packets. It should not be confused with flow control, which
prevents the sender from overwhelming the receiver.
Theory of congestion control
The modern theory of congestion control was pioneered by Frank Kelly, who
applied microeconomic theory and convex optimization theory to describe how individuals
controlling their own rates can interact to achieve an "optimal" network-wide rate allocation.
Examples of "optimal" rate allocation are max-min fair allocation and Kelly's suggestion
of proportional fair allocation, although many others are possible.
The mathematical expression for optimal rate allocation is as follows. Let
be the rate of flow
,
be the capacity of link , and
be 1 if flow uses link and 0 otherwise. Let , and be
the corresponding vectors and matrix. Let
be an increasing, strictly convexfunction, called
the utility, which measures how much benefit a user obtains by transmitting at rate . The
optimal rate allocation then satisfies
such that
The Lagrange dual of this problem decouples, so that each flow sets its own rate, based only on
a "price" signalled by the network. Each link capacity imposes a constraint, which gives rise to a
Lagrange multiplier,
. The sum of these Lagrange multipliers,
is the price to
which the flow responds.
Congestion control then becomes a distributed optimisation algorithm for solving the above
problem. Many current congestion control algorithms can be modelled in this framework, with
being either the loss probability or the queueing delay at link .
A major weakness of this model is that it assumes all flows observe the same price, while sliding
window flow control causes "burstiness" which causes different flows to observe different loss or
delay at a given link.
Classification of congestion control algorithms
Main article: Taxonomy of congestion control
Shashank Agnihotri
Computer Networks – Page 80
There are many ways to classify congestion control algorithms:

By the type and amount of feedback received from the network: Loss; delay; single-bit or
multi-bit explicit signals

By incremental deployability on the current Internet: Only sender needs modification;
sender and receiver need modification; only router needs modification; sender, receiver and
routers need modification.

By the aspect of performance it aims to improve: high bandwidth-delay product networks;
lossy links; fairness; advantage to short flows; variable-rate links

By the fairness criterion it uses: max-min, proportional, "minimum potential delay"
Avoidance
The prevention of network congestion and collapse requires two major components:
1.
A mechanism in routers to reorder or drop packets under overload,
2.
End-to-end flow control mechanisms designed into the end points which respond to
congestion and behave appropriately.
The correct end point behaviour is usually still to repeat dropped information, but progressively
slow the rate that information is repeated. Provided all end points do this, the congestion lifts and
good use of the network occurs, and the end points all get a fair share of the available
bandwidth. Other strategies such as slow-start ensure that new connections don't overwhelm the
router before the congestion detection can kick in.
The most common router mechanisms used to prevent congestive collapses are fair
queueing and other scheduling algorithms, and random early detection, or RED, where packets
are randomly dropped proactively triggering the end points to slow transmission before
congestion collapse actually occurs. Fair queueing is most useful in routers at choke points with
a small number of connections passing through them. Larger routers must rely on RED.
Some end-to-end protocols are better behaved under congested conditions than others. TCP is
perhaps the best behaved. The first TCP implementations to handle congestion well were
developed in 1984[citation needed], but it was not until Van Jacobson's inclusion of an open source
solution in the Berkeley Standard Distribution UNIX ("BSD") in 1988 that good TCP
implementations became widespread.
UDP does not, in itself, have any congestion control mechanism. Protocols built atop UDP must
handle congestion in their own way. Protocols atop UDP which transmit at a fixed rate,
independent of congestion, can be troublesome. Real-time streaming protocols, including
Shashank Agnihotri
Computer Networks – Page 81
many Voice over IP protocols, have this property. Thus, special measures, such as quality-ofservice routing, must be taken to keep packets from being dropped from streams.
In general, congestion in pure datagram networks must be kept out at the periphery of the
network, where the mechanisms described above can handle it. Congestion in the Internet
backbone is very difficult to deal with. Fortunately, cheap fiber-optic lines have reduced costs in
the Internet backbone. The backbone can thus be provisioned with enough bandwidth to keep
congestion at the periphery.[citation needed]
Practical network congestion avoidance
Implementations of connection-oriented protocols, such as the widely-used TCP protocol,
generally watch for packet errors, losses, or delays (see Quality of Service) in order to adjust the
transmit speed. There are many different network congestion avoidance processes, since there
are a number of different trade-offs available. [1]
TCP/IP congestion avoidance
Main article: TCP congestion avoidance algorithm
The TCP congestion avoidance algorithm is the primary basis for congestion control in the
Internet. [2] [3] [4] [5] [6]
Problems occur when many concurrent TCP flows are experiencing port queue buffer tail-drops.
Then TCP's automatic congestion avoidance is not enough. All flows that experience port queue
buffer tail-drop will begin a TCP retrain at the same moment - this is called TCP global
synchronization.
Active Queue Management (AQM)
Main article: Active Queue Management
Purpose
Random early detection
Main article: Random early detection
Main article: Weighted random early detection
One solution is to use random early detection (RED) on network equipments port queue
buffer. [7] [8] On network equipment ports with more than one queue buffer, weighted random
early detection (WRED) could be used if available.
Shashank Agnihotri
Computer Networks – Page 82
RED indirectly signals to sender and receiver by deleting some packets, e.g. when the average
queue buffer lengths are more than e.g. 50% (lower threshold) filled and deletes linearly more or
(better according to paper) cubical more packets, [9] up to e.g. 100% (higher threshold). The
average queue buffer lengths are computed over 1 second at a time.
Robust random early detection (RRED)
Main article: Robust random early detection
Robust Random Early Detection (RRED) algorithm was proposed to improve the TCP
throughput against Denial-of-Service (DoS) attacks, particularly Low-rate Deinal-of-Service
(LDoS) attacks. Experiments have confirmed that the existing RED-like algorithms are notably
vulnerable under Low-rate Denial-of-Service (LDoS) attacks due to the oscillating TCP queue
size caused by the attacks [10]. RRED algorithm can significantly improve the performance of
TCP under Low-rate Denial-of-Service attacks [10].
Recent Publications in low-rate Denial-of-Service (DoS) attacks
Flowbased-RED/WRED
Some network equipment are equipped with ports that can follow and measure each flow
(flowbased-RED/WRED) and are hereby able to signal to a too big bandwidth flow according to
some QoS policy. A policy could divide the bandwidth among all flows by some criteria.
IP ECN
Main article: Explicit Congestion Notification
Another approach is to use IP ECN.[11] ECN is only used when the two hosts signal that they
want to use it. With this method, an ECN bit is used to signal that there is explicit congestion.
This is better than the indirect packet delete congestion notification performed by the
RED/WRED algorithms, but it requires explicit support by both hosts to be effective. [12] Some
outdated or buggy network equipment drops packets with the ECN bit set, rather than ignoring
the bit. More information on the status of ECN including the version required for Cisco IOS,
by Sally Floyd,[7] one of the authors of ECN.
When a router receives a packet marked as ECN capable and anticipates (using RED)
congestion, it will set an ECN-flag notifying the sender of congestion. The sender then ought to
decrease its transmission bandwidth; e.g. by decreasing the tcp window size (sending rate) or by
other means.
Cisco AQM: Dynamic buffer limiting (DBL)
Shashank Agnihotri
Computer Networks – Page 83
Cisco has taken a step further in their Catalyst 4000 series with engine IV and V. Engine IV and
V has the possibility to classify all flows in "aggressive" (bad) and "adaptive" (good). It ensures
that no flows fill the port queues for a long time. DBL can utilize IP ECN instead of packetdelete-signalling. [13] [14]
TCP Window Shaping
Congestion avoidance can also efficiently be achieved by reducing the amount of traffic flowing
into a network. When an application requests a large file, graphic or web page, it usually
advertises a "window" of between 32K and 64K. This results in the server sending a full window
of data (assuming the file is larger than the window). When there are many applications
simultaneously requesting downloads, this data creates a congestion point at an upstream
provider by flooding the queue much faster than it can be emptied. By using a device to reduce
the window advertisement, the remote servers will send less data, thus reducing the congestion
and allowing traffic to flow more freely. This technique can reduce congestion in a network by a
factor of 40.[citation needed]
Side effects of congestive collapse avoidance
Radio links
The protocols that avoid congestive collapse are often based on the idea that data loss on the
Internet is caused by congestion. This is true in nearly all cases; errors during transmission are
rare on today's fiber based Internet. However, this causes WiFi, 3G or other networks with a
radio layer to have poor throughput in some cases since wireless networks are susceptible to
data loss due to interference. The TCP connections running over a radio based physical
layer see the data loss and tend to believe that congestion is occurring when it isn't and
erroneously reduce the data rate sent.
Short-lived connections
The slow-start protocol performs badly for short-lived connections. Older web browsers would
create many consecutive short-lived connections to the web server, and would open and close
the connection for each file requested. This kept most connections in the slow start mode, which
resulted in poor response time.
To avoid this problem, modern browsers either open multiple connections simultaneously
or reuse one connection for all files requested from a particular web server. However, the initial
performance can be poor, and many connections never get out of the slow-start regime,
significantly increasing latency.