ethernet/íðnet
Nobody is heard correctly.
NET0183 Networks and Communications
Lectures 5 and 6
Ethernet
“Ethernet is named for the historical luminiferous ether through
which electromagnetic radiations were once alleged to propagate.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
1
Historical Case Study
http://portal.acm.org/beta/citation.cfm?id=360253
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
2
1. Background
• “One can characterize distributed computing as a spectrum
of activities varying in their degree of decentralization, with
one extreme being remote computer networking and the
other extreme being multiprocessing.”
• “Remote computer networking is the loose interconnection
of previously isolated, widely separated, and rather large
computing systems.”
– e.g. ALOHANET of 1970...
• “Multiprocessing is the construction of previously monolithic
and serial computing systems from increasingly numerous
and smaller pieces computing in parallel.”
– e.g. dual or multi-core processors in laptop computers of 2010...
• “Near the middle of this spectrum is local networking…”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
3
The distributed computing spectrum in 1976.
Mbps here is mega bit per second.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
4
http://en.wikipedia.org/wiki/Data_rate_units
3
10
6
10
MBps here is mega byte per second.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
5
http://en.wikipedia.org/wiki/Data_rate_units
9
10
12
10
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
6
http://www.ntt.co.jp/news/news06e/0609/060929a.html
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
7
Local Area Network
http://en.wikipedia.org/wiki/Local_area_network
A local area network (LAN) is a computer network covering a small
physical area, like a home, office, or small group of buildings, such
as a school, or an airport. The defining characteristics of LANs, in
contrast to wide-area networks (WANs), include their usually higher
data-transfer rates, smaller geographic area, and lack of a need for
leased telecommunication lines.
ARCNET, Token Ring and many other technologies have been used
in the past, and G.hn may be used in the future, but Ethernet over
twisted pair cabling, and Wi-Fi are the two most common
technologies currently in use.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
8
“Just as computer networks have grown across
continents and oceans to interconnect major computing
facilities around the world, they are now growing down
corridors and between buildings to interconnect minicomputers in offices and laboratories .”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
9
3. Design Principles
• The network should be able to grow smoothly to
accommodate several buildings.
• The communication system must be inexpensive.
• Control should be distributed to avoid the
reliability problems of a central controller.
– If a central controller goes down, the whole network
is down.
• The initial design of the Ethernet was based on
ALOHANET. Ethernet is a broadcast packet
switching network, and not a store and forward
network like ARPANET.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
10
coaxial cable/samása strengur
fibre optics/ljósþráðatækni
3. Design Principles
• Like ALOHANET, Ethernet was expected to carry bursty
traffic so that “conventional synchronous time-division
multiplexing (STDM) would be inefficient”.
– Allocating each node a time slot would be wasteful. Many time
slots would go unused.
• “The Ether is a logically passive medium for the
propagation of digital signals and can be constructed
using any number of media including coaxial cables,
twisted pairs, and optical fibers.”
• Several design innovations were also necessary, including
the idea of detecting interference with a packet as it is
being transmitted.
– When we start to speak, if we hear many other voices speaking,
we stop speaking because we know we will not be heard
properly. We wait for a quieter moment.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
11
topology/grannfræði
3.1 Topology
• A topology was chosen to make it easy to meet the
changing needs of office and laboratory buildings.
– “Any station wishing to join an Ethernet taps into the Ether
at the nearest convenient point.”
• The topology is that of an unrooted tree.
– There is only one path between any source and
destination, so avoiding multipath interference.
– Being unrooted, the network can be extended from any
point in any direction.
– The tree is a logically passive medium. A failure of an
active element should affect communications at a single
node only.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
12
“Unlike an Aloha Network, which is a star network with an
outgoing broadcast channel and an incoming multi-access
channel, an Ethernet supports many-to- many communication
with a single broadcast multi-access channel.”
Some high-level network topologies.
http://www.11h11.com/hugobox/chaku/NetworkTopologies.png
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
13
Terminator
transceiver/sendiviðtæki
Figure 1. A two-segment Ethernet. ©ACM
Tap
Segment #1 might be in one corridor with
the network cable run under the floor or in
the ceiling. Segment #2 might represent
the cable being run down another corridor.
A transceiver is a device that both
transmits and receives signals.
“A tap is a device for physically connecting
to the Ether while disturbing its
transmission characteristics as little as
possible.”
#1
#2
8/25/2009
Andy comments: This figure illustrates that
a high-level topological map of a network
can disguise the relationship between the
nodes and the network cable. Ethernet is
an unrooted tree regarding the cable.
Stations tap into this unrooted tree.
NET0183 Networks and Communications
by Dr Andy Brooks
14
3.2 Control
• Stations might attempt to transmit a packet at roughly
the same time resulting in packets which overlap and
which are unrecognisable to receivers. Packets are said
to collide.
• A station abandons a transmission when a collision has
been detected.
• Stations then attempt to retransmit packets after a
random period of time.
• Ethernet controllers in each station adjust the mean
retransmission interval in proportion to the count of
collisions.
– As collisions increase, the random time intervals before
retransmission increase.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
15
3.2 Control
• “A station could usurp the Ether by not adjusting its
retransmission interval with increasing traffic or by
sending very large packets. Both practices are now
prohibited by low-level software in each station.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
16
3.3 Addressing
• Each packet has a source and destination specified
in the packet´s header.
• Any station can read a packet into its memory but
normally only the destination station would do this.
• A zero destination address is a wildcard which
matches all addresses.
– “A packet with a destination of zero is called a broadcast
packet”.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
17
reliability/áreiðanleiki
3.4 Reliability
• Packets may be lost because of noise or a downed
receiver at a packet´s intended destination.
– The Ethernet does not guarantee error-free
communication.
• “Protocols used to communicate through an Ethernet
must assume that packets will be received correctly at
intended destinations only with high probability.”
– Applications have to deal with error recovery.
• “This policy becomes more important as Ethernets
are interconnected in a hierarchy of networks
through which packets must travel farther and suffer
greater risks.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
18
3.5.1 Carrier detection
• A packet´s bits are phase encoded which means there is at
least one transition each bit time.
• Receivers can listen to the passing of a packet.
• “... we speak of the presence of a carrier as a packet passes a
transceiver.”
• “With carrier detection we are able to implement
deference: no station will start transmitting while hearing
carrier.”
• “With carrier detection, collisions should occur only when
two or more stations find the Ether silent and begin
transmitting simultaneously: within an Ether end-to-end
propagation time.”
– It takes time for signals to travel the length of the network cable.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
19
3.5.2 Interference detection
• “Interference is indicated when the transceiver notices a
difference between the value of the bit it is receiving from
the Ether and the value of the bit it is attempting to
transmit.”
• A packet can be scheduled for retransmission as soon as it
is found to have been damaged.
• Only a fraction of packet time is wasted since damaged
packets are truncated, in contrast to ALOHANET where
colliding packets ran to completion.
• The frequency of collisions is used to adjust retransmission
intervals to help make better use of the communication
channel.
– As collisions increase, the random time intervals before
retransmission increase.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
20
checksum/prófsumma
3.5.3 Packet error detection
• During transmission a checksum is computed and
added to a packet. When the packet is read from the
Ether, the checksum is recomputed.
– hardware checksum
• Packets are discarded if the checksums are
inconsistent.
• “In this way transmission errors, impulse noise errors,
and errors due to undetected interference are caught
at a packet's destination.”
Andy comments: A simple checksum breaks a message down into fixed-sized words
and then adds the numeric value of these words together. Zero bits may have to be
appended at the end of the message to have a properly-sized word.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
21
3.5.4 Truncated packet filtering
• To reduce the processing load at listening
stations, truncated packets (only a few bits) are
filtered out in hardware.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
22
3.5.5 Collision consensus enforcement
• “When a station determines that its
transmission is experiencing interference, it
momentarily jams the Ether to insure that all
other participants in the collision will detect
interference and, because of deference, will be
forced to abort.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
23
4. Implementation
• Parameters of the experimental Etherenet were: 1
kilometre, 3 megabits per second, and 256 stations.
– 28 = 256
• The first byte of a packet was the destination address field.
• The second byte of a packet was the source address field.
• “256 is a number small enough to allow each station to
get an adequate share of the available bandwidth and
approaches the limit of what we can achieve with current
techniques for tapping cables.”
• Higher level protocols “can accommodate extended
address spaces with additional fields inside the packet and
software to interpret them”.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
24
4.1 Ether
• The experimental ethernet was implemented using lowloss coaxial cable with off-the-shelf CATV taps and
connectors.
– CATV Community Antenna Television
• A smaller-diameter coax was used within station clusters
and a larger-diameter coax for low-loss runs between
clusters.
– Larger-diameter coax suffers less signal attenuation.
• “The cost of coaxial cable Ether is insignificant relative to
the cost of the distributed computing systems supported
by Ethernet.”
coaxial
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
25
4.2 Transceivers
• The transceivers can tolerate everyday electrical noise from
typewriters and electric drills.
• “When unpowered, the transceievr disconnects itself
electrically from the Ether.”
• A broken transceiver should not bring down the whole
Ethernet.
– A watchdog circuit shuts down the transceiver if the output
stage “acts suspiciously”.
• “Even though our experimental transceivers are very
simple and can tolerate only limited signal attenuation,
they have proven quite adequate and reliable. A more
sophisticated transceiver design might permit passive
branching of the Ether and wider station separation.”
– passive branching : without repeaters
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
26
4.3 Interface
• The interface serializes and deserializes the parallel data
used by its station.
– bit bit bit...<->16-bit word, 16-bit word, 16-bit word
• Interfaces have hardware to compute a 16-bit cyclic
redundancy checksum (CRC) on serial data as it is
transmitted and received.
– CRCs provide protection against errors in the Ether and not
against errors in the parallel portions of the interface hardware
or station.
• “Higher-level software checksums are recommended for
applications in which a higher degree of reliability is
required.”
Andy comments: More and more checks to ensure that data has been transmitted
correctly and that data has not be compromised by malware, means that more and
more of the bandwidth is used just for checking.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
27
4.3 Interface
• “When carrier goes away, the interface checks that an
integral number of 16-bit words has been received and that
the CRC is correct. The last word received is assumed to
be the CRC and is not copied into the packet buffer.”
– The packet buffer is in the station´s main memory.
• “These interfaces ordinarily include hardware for
accepting only those packets with appropriate addresses in
their headers.”
– Hardware address filtering reduces packet processing in
software.
• Interfaces had to be built for several different stations on
the experimental ethernet.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
28
Fig. 2. Ethernet packet layout.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
29
4.4 Controller
• A controller is station-specific firmware or software which
controls getting packets on or off the Ether.
• Retransmission intervals are multiples of a slot, “the
maximum time between starting a transmission and
detecting a collision.”
– end-to-end round trip delay
• “Each time a transmission attempt ends in collision, the
controller delays for an interval of random length with a
mean twice that of the previous interval,…”
– “This heuristic approximates an algorithm we have called Binary
Exponential Backoff.”
• “When the network is unloaded and collisions are rare,
the mean seldom departs from one and
retransmissions are prompt.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
30
5. Growth
5.1 Signal Cover
• To support growth of the network, a simple unbuffered
packet repeater can be used to extend signal cover.
• Using repeaters, however, makes the Ether active
rather than passive.
– When a transceiver goes down, only its station ceases
communication. When a repeater goes down,
communications are severed with a whole section of the
network.
• “We operate an experimental two-segment packet
repeater, but hope to avoid relying on them.”
– More sophisticated transceivers can be used instead of
relying on repeaters to extend the signal cover.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
31
5. Growth
5.2 Traffic Cover
• Adding more stations will eventually put too
much load on the network.
• One solution is to use packet filters.
– An unbuffered traffic-filtering repeater.
– A packet filter also extends the signal cover.
• Packet filters only pass packets from one Ether
segment to the other if the destination station is
located on the new segment.
Andy comments: Security firewalls can use packet filtering. Packets are inspected according
to various rules and are passed through the firewall only if they do not break any rule.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
32
5. Growth
5.3 Address Cover
• Experimental Ethernet has only an 8-bit address.
Continued growth of the network will mean the
number of stations will exceed 256.
• Addresses can be extended down into a station by
adding fields to identify ports or processes within a
station.
• Addresses can be extended up into the internetwork by
adding fields to identify destination stations on remote
networks.
• Address cover can be extended using packet gateways.
“A gateway repeats packets addressed to itself as an
intermediary”.
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
33
5. Growth
5.3 Address Cover
• A gateway also extends the signal and traffic cover.
• “Failure of the single repeater connecting two
segments partitions the network; failure of a gateway
need not partition the net if there are paths through
other gateways between the segments.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
34
protocol/samskiptareglur
7. Protocol
“There is more to the construction of a viable packet
communication system than simply providing the mechanisms
for packet transport. Methods for error correction, flow
control, process naming, security, and accounting must also be
provided through higher-level protocols implemented on top of
the Ether control protocol described in Sections 3 and 4 above.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
35
file transfer/skráaflutningur
7.2 EFTP
• EFTP (Ethernet File Transfer Protocol) is one
simple error-controlling packet protocol.
– The first 16 bits are the destination and source
addresses.
– The second 16 bits contain the packet type.
• data, ack, abort, end, endreply
– The next 16 bits are a sequence number.
– The next 16 bits are a length.
– After a number of optional 16-bit data words, there is
a 16-bit software checksum.
• “The Ethernet´s hardware checksum is present only on the
Ether and is not counted at this level of protocol.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
36
Fig. 4. EFTP packet layout ©ACM
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
37
7.2 EFTP
• The software checksum protects against failures in
parallel data paths within stations that are not checked
by the CRC.
• A file´s data packets are numbered from 0.
• “Each data packet is retransmitted periodically by
the sender until an ack packet with a matching
sequence number is returned from the receiver.”
• “When a packet arrives with a sequence number
one less than that expected, it is acknowledged
and discarded; the presumption is that its ack was
lost and needs retransmission.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
38
7.2 EFTP
• “When all the data has been transmitted, an end
packet is sent with the next consecutive sequence
number and than the sender waits for a matching
endreply.”
• “At any time during all of this, either side is free
to decide communication has failed and just give up,
it is considered polite to send an abort packet to
end the communication promptly in the event of,
say, a user-initiated abort or a file system error.”
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
39
7.2.3 EFTP shortcomings
• The protocol only provides for file transfer between
stations in a single network.
– The protocol does not deal with processes within stations
or with gateways between networks.
• Different users of a station are not distinguished.
• Files are unnamed.
• Data is transmitted in integral numbers of 16-bit words.
– What happens when the file is an odd number of bytes?
8/25/2009
NET0183 Networks and Communications
by Dr Andy Brooks
40