ITEC350 Networks I Lecture 2 Hwajung Lee Elements of a Network Applications Stations Frames Switches Elements of a Network Applications Application Application Frame Networks connect applications on different stations Copyright 2003 Prentice-Hall Elements of a Network Stations Frame Server Station Client Station Mobile Client Station Stations are computers and other devices such as cellphones and PDAs Server Station Copyright 2003 Prentice-Hall Elements of a Network Frames Frame Server Station Client Station Mobile Client Station Stations communicate by sending messages called Frames Server Station Copyright 2003 Prentice-Hall Elements of a Network Switches Frame Switch Switch Switch Switch Frames may pass through multiple switches and routers; Each switch or router reads the frame And passes it on Copyright 2003 Prentice-Hall What is Computer Networking? – Vocabulary and Acronyms Computer Networking-- The discipline of study which examines how bits, in the form of energy, travel across some medium from a source to a destination. Network -- A communication system for connecting computers using a single transmission technology Internetwork -- A communication system for connecting networks together using routers. Router (Packet Switch) -- A computer that attaches two or more networks , forwarding packets from one network to another (building block of an internet). Ethernet – A physical transmission and link layer standard suite Bit – a single indivisible unit of information Byte – 8 bits (an octet; usually represents one character) Packets & Frames – A sequence of bytes sent as a single entity from source to destination Other Network Terminology A Network in the sense of Graph Theory is A set of Nodes N={A,B,C…} of cardinality n (n=number of Nodes in the set N) A set of Links L={AB,AC,…} Minimal .vs. Maximal connectivity of a given network Minimal connectivity: connecting the nodes with the least number of links (AKA spanning tree) • Many ways – Choose n-1 links out of n*(n-1)/2 Maximal connectivity: connecting every node with every other node • Exactly one way – n*(n-1)/2 links Topics of Networking Data Transmission Describes the transfer of info through electrical signals Packet Transmission Need for packets, embedding packets in other packets Addressing How to give a “name” to each source and destination Routable .vs. Unroutable address spaces Public .vs. Private address ranges Domain names .vs. IP addresses .vs. HW addresses Internetworking Protocols Universal service, end-to-end reliable transport, routing Applications Client-server paradigm, socket interface, Domain Name System (DNS), email, and ftp Network Operating Systems Quality of Service (QoS) Speed Bits per second (bps) Multiples of 1,000 (not 1,024) Kilobits per second (kbps) Megabits per second (Mbps) Gigabits per second (Gbps) Terabits per second (Tbps) Petabits per second (Pbps) Latency Latency or Delay delay measured in milliseconds (ms) Especially bad for some services such as voice communication or highly interactive applications Reliability Availability Percent of time the network is available to users for transmission and reception Telephone network: Five 9s (99.999%) Error Rate Percent of lost or damaged messages or bits Client/Server Architecture Usually, Two Types of Stations Clients and Servers Server Client PC Service Network Clients Receive Services Servers Provide Services Copyright 2003 Prentice-Hall Servers PC Server Standard PC Microprocessor Moderate speed and cost Microsoft Windows Server Novell Network LINUX (Version of UNIX) Fast (and Expensive) Custom-Built Microprocessor (e.g., Workstation SUN microsystems) UNIX Operating System Server Highly reliable Fastest Servers UltraReliable Specialized Mainframe Operating Server System Copyright 2003 Prentice-Hall Network, Server, & Client Oper. Systems Cisco IOS, JUNOS (Juniper), etc. Novell NetWare and IntranetWare UNIX, Linux Windows NT Server, Win2000 Server(W2K), Windows Server 2003(W2K3) Windows NT Workstation, Win2000 Prof. Windows 95 Windows-for-Workgroups Switching Decision Ethernet Switch 1 2 3 4 5 6 Switch Sends Signal out a Single Port Station A Transmits to Station C Station A Switch receives a frame, sends it back out -- learns where to send --by looking at address Station B Station C Station D Copyright 2003 Prentice-Hall Packet Switching (Routing) (1) Packet Original Message Switch (Router) Computer X A Packet B 1. Break message into Smaller packets (carried inside of frames) C Switching Decision D 2. Route packets individually; Packet switches along the way Make decisions about the packet E Computer Y F Copyright 2003 Prentice-Hall Packet Switching (2) Packet Switching Reduces Trunk Line Costs (Shares Hardware) Packets from several conversations are multiplexed on trunk lines Conversations do not need the full capacity of the shared trunk lines. This reduces trunk line costs. Trunk Link Packet from A to B Packet from C to D Copyright 2003 Prentice-Hall Packet Switching (3) Packet Trailer Data Field Header Other Header Field Address Fields Packet Structure Copyright 2003 Prentice-Hall Frames and Packets (1) Within a single network, the packet is carried in the data field of that network’s frame, probably across multiple switches Packet Frame Network 1 Copyright 2003 Prentice-Hall Frames and Packets (2) The first router removes packet from first network’s frame, sends back out in a frame with the second network’s frame format Same Packet Frame Network 1 Second Router First Router Network 2 Frame With Network 2’s Frame Format Copyright 2003 Prentice-Hall Frames and Packets (3) The second router removes packet from second network’s frame, sends back out in a frame with the third network’s frame Same format Packet Network 3 Network 2 Frame With Network 2’s Frame Format Frame With Network 3/s Second Frame Format Router Copyright 2003 Prentice-Hall Frames and Packets (4) Like passing a shipment (the packet) from a truck (frame) to an airplane (frame) at an airport. Receiver Shipper Same Shipment Truck Airport Airport Truck Airplane Copyright 2003 Prentice-Hall An Internet Multiple Networks Connected by Routers Path of a Packet is its Route Single Network Routers Packet Single Network Route Copyright 2003 Prentice-Hall An Internet Single Networks Have Switches Switches Connect Station-to-Router or Router-to-Router Network Y Network X Switches Routers Network Z Switches Copyright 2003 Prentice-Hall The Internet Browser The global Internet Has thousands of networks Webserver Software Network Packet Router Packet Route Router Router Packet User PC (Host) IP Address=128.150.50.9 Webserver (Host) IP Address=128.171.17.13 Host name=voyager.cba.hawaii.edu Copyright 2003 Prentice-Hall Addressing Addressing (Read pp 195-211, Minasi) Domain names: “radford.edu” IP Addresses: iii.jjj.kkk.lll, dotted decimal Example: Radford University has a computer (somewhere) with IP address 137.45.192.36 MAC (Hardware) Address Hexadecimal digits separated by colons or dash. See figure 4-10 in Panko if you are new to “HEX” Example: 00-06-6B-FF-0A-B4 Specific .vs. Broadcast (FF-FF-FF-FF-FF-FF) Addresses IP Addresses (do HW RU01) An IP Packet can be sent to A single workstation (unicast) • Efficient for data between pairs of addresses A specific list of workstations (multicast) • Efficient for specific groups, but must specify all individual workstations IP addresses All stations on a network (broadcast) • Efficient for large (unknown) group – use special broadcast IP address. IP addresses have a special broadcast address Class .vs. Classless Addressing. Special IP Addresses THIS computer - all 0’s--both prefix and suffix 0.0.0.0 THIS network broadcast - all 1’s prefix and suffix 255.255.255.255 remote net broadcast - net prefix all 1’s suffix Ex: 137.45.192.255 Network address - net prefix all 0’s suffix 137.45.192.0 loopback - 127.x.x.x but usually 127.0.0.1 Everything else is a Host IP Address like 137.45.192.96 IP Address Ranges, Or “Classes” From: To: Description 1.x.x.x 126.x.x.x Class A license 127.x.x.x 127.x.x.x Loop back 191.x.x.x Class B license (172.16 thru 31. 0. 0 reserved for private addresses) 192.x.x.x 223.x.x.x Class C license (192. 168. x. 0 reserved for private addresses) 224.0.0.0 224.0.0.255 Multicast: Reserved Link Local Addresses 224.0.1.0 238.255.255.255 Multicast: Globally Scoped Addresses 239.0.0.0 239.255.255.255 Multicast: Limited Scope Addresses 240.x.x.x 255.255.255.254 Experimental 128.x.x.x 255.255.255.255 Broadcast IP Format 137.45.104.172 Dotted Decimal vs Binary 137.45.104.172 10001001001011010110100010101100 Conversion Between Decimal & Binary 128 64 32 16 8 4 2 1 X X X X X X X X 1 0 0 0 1 0 0 1 = = = = = = = = 128 0 0 0 8 0 0 1 137 Conversion Between Decimal & Binary 128 64 32 16 8 4 2 1 1 128 0 0 0 0 0 0 1 8 0 0 0 0 1 1 137 0 0 1 0 1 1 0 1 0 0 32 0 8 4 0 1 45 0 1 1 0 1 0 0 0 0 64 32 0 8 0 0 0 1 128 0 0 1 32 0 0 1 8 1 4 0 0 0 0 104 172 Subnetting Subnet Mask for Class C 137.45.104.172 255.255.255.0 “Anding” a Binary Subnet Mask 10001001001011010110100010101100 11111111111111111111111100000000 10001001001011010110100000000000 subnet ID = (137.45.104.0) Why Subnets? In class A, B, or C networks, there are too many IP addresses to fit on one segment. Thus, need routers and subnets to isolate parts. Subnets: A new interpretation IP Addresses had a new subnet field inserted between network & local fields IP address := <network-number><subnet-number><host-number> Ex: A Class A Network with 8-bit subnet field 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0| NETWORK | SUBNET | Host number | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Class C subnet example Read Minasi, pp. 206 – 211 Also, see www.minasi.com -- newsletters, etc. Look at IP Subnetting Tutorial http://www.ralphb.net/IPSubnet/index.html WAN Link Router Router Network address: 192.168.1.64 Los Angeles Mask: 255.255.255.192 Ethernet Sw. Host addresses 192.168.1.65-126 PC3 PC1 PC2 New York Ethernet Sw. Network address: 192.168.1.128 Mask: 255.255.255.192 Host addresses 192.168.1.129-190 PC4 PC5 PC6 PC7 Subnet example 192.168.1.0 = Basic Class C Network ID 255.255.255.0 = Class C Mask Old Class C Boundary Between Network and Local 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 0| (Sub)NETWORK | Local Addr| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Must Use 2 extra bits for the First feasible sub-division of Class C into two subnets New Class C Subnet Boundary Between 192.168.1.64, 192.168.1.128 New sub-Network IDs Network and Local 255.255.255.192 = New Subnet Mask SubNetwork IDs, Host Ranges & Broadcast Addresses Using extra two bits in Network ID 00 – Can’t use because this is the part of the original Class C’s Network ID 01 – Available 01000000 = 64 10 – Available 10000000 = 128 11 – Can’t use because this is part of the original Class C’s broadcast address Hence 192.168.1.64 is the first sub-Network ID 192.168.1.128 is the second Binary for the subnetwork IDs Byte boundaries shown by dashed lines Subnet IDs = Local address field of all zeroes (6 bits) 01 or 10 to get bottom byte (8 bits) Result = 64 or 128 when translated to decimal 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 0| (Sub)NETWORK 0 1|0 0 0 0 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 192 . 168 . 1 . 64 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 0| (Sub)NETWORK 1 0|0 0 0 0 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 192 . 168 . 1 . 128 Binary for Masks (Old .vs. New) A Mask is a device for indicating how long the (sub)network field is All 1’s covering the entire network id portion 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 1|1 1 1 1 1 1… OLD NETWORK MASK …1 1 1 1|0 0 0 0 0 0 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 255 . 255 . 255 . 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 1|1 1 1 1 1 1… (Sub)NETWORK MASK …1 1 1 1 1 1|0 0 0 0 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 255 . 255 . 255 . 192 Host Ranges Network Mask is 255.255.255.192 192.168.1.64 has 62 host addresses First available host address = 192.168.1.65 Last available host address = 192.168.1.126 Broadcast address = 192.168.1.127 192.168.1.128 has 62 host addresses First available host address = 192.168.1.129 Last available host address = 192.168.1.190 Broadcast address = 192.168.1.191 Minasi p207 Binary for Broadcast addresses Broadcast addresses have all 1’s in the host field Remember, we always translate 8 bit octets to decimal! DO HOMEWORK RU02 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 0| (Sub)NETWORK 0 1|1 1 1 1 1 1| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 192 . 168 . 1 . 127 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 0| (Sub)NETWORK 1 0|1 1 1 1 1 1| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 192 . 168 . 1 . 191 Recap Network Classes IANA (Internet Assigned Numbers Authority) Class A IP address := <8bits>.<24bits> 16 Million hosts in a class A network domain Class B IP address = <16bits>.<16bits> 65534 hosts in a class B network domain Class C IP address = <24bits>.<8bits> 256 hosts in a class C network domain Routable and Nonroutable Addresses Nonroutable Address [RFC 1918] Internet Router ignore the following addresses. • 10.0.0.0 – 10.255.255.255 • 172.16.0.0 – 172.31.255.255 • 192.168.0.0 – 192.168.255.255 Millions of networks can exist with the same nonroutable address. “Intranet” : Internal Internet Side benefit : “Security” NAT (Network Address Translation) router Note on Classful vs. Classless Note that, in classful subnetting, we lose quite a few blocks of addresses. RFC 1519 (Classless Inter-Domain Routing = CIDR) was introduced in 1993 to deal with rapid depletion of IP address space due to “Classful Fragmentation” Problem: Given the entire internet was “classful” in 1993, how to transition to classless methods? What exactly is the impact to internet protocols (in all the millions of devices and hosts) of such a change? Impact of CIDR We needed new routing protocols (haven’t introduced those yet) We need new ways of handling masks We will revisit classless addressing and routing in a few weeks, but the bottom line is: There is a way to use all those un-used addresses (all zeroes, all ones) that we discarded in classful subnetting. Minasi p. 209~210 shows a classless subnetting of a class C network. HW (e.g., Ethernet) Addresses A Hardware (HW) address of all 1’s signifies the broadcast address at the link layer of Ethernet Ethernet NICs can also be configured (through software) with several Multicast addresses All Ethernet NICs will accept a packet with either Individual HW address of NIC The broadcast address Any of the configured multicast addresses Finally, Ethernet NICs can be put into promiscuous mode – accept all packets regardless of HW address Useful for monitoring, “sniffing”, debugging