1. CSE 331: Introduction to Networks and Security Fall 2000 Instructor: Carl A. Gunter Slide Set 2

2. Issues l Information movement implies:  An information source  An information destination  A path from source to destination l Addresses are used to locate source and destination l Path can be static or dynamic

3. Paths are made of links l Links are interconnected by zero or more network elements, e.g., switches, routers, hubs, bridges, etc. l Links have delay and throughput l Path delay is sum of link delays plus switching delays l Path throughput = bottleneck link t’put

4. Some link types: l Multiple-access (e.g., Ethernet LAN)  10, 100 Mbps, 1km, 100m diameters l SONET Fiber (up to 9.6 Gbps) l CATV (usually 1-6 Mbps, asymmetric) l ISDN (64 Kbps*n with bonding) l POTS (56 Kbps w/luck & clever coding)

5. Making a network l Rules for interconnecting links l Rules for interpreting addresses l Link properties versus end-to-end properties  e.g., ATM guarantees in-order arrival  e.g., TCP makes reliable stream from datagrams

6. Together, rules make a protocol l Protocols are logically layered l Peer-to-peer, corresponding layers l Higher layers constructed using lower layer protocols l Examples: OSI and IP

7. Requirements l Connectivity l Cost-effective resource sharing l Functionality l Performance

8. Direct Links Point to Point Multiple Access

9. Switched Networks Network “Cloud” Hosts Network Elements Links

10. Internetworks Router (Gateway)

11. Further Distinctions l Addressing  Unicast  Broadcast  Multicast  Anycast l Size  Local  Wide  Metropolitan  Desk

12. Desk Area Network

13. Sharing Resources l How can hosts share the network if they want to use it at the same time?  How can they share links?  How can they share switches?

14. Multiplexing strategies l Two approaches  Synchronous Time-Division Multiplexing (STDM).  Statistical Multiplexing. l Tradeoffs  STDM can ensure service for a complete communication, but may waste resources.  Statistical multiplexing ensures less, but can make better use of resources.

15. Packets and Congestion

16. Functionality The aim of the network is to support the communication needs of applications. Reliable unicast channel Unreliable multicast channel with in-order delivery A B C F E D

17. Dealing with failure l Congestion l Bit or burst errors l Link or node outages

18. Performance l Bandwidth (throughput)  The number of bits that can be transmitted over the network in a certain period of time. l Latency (delay)  How long it takes a single bit to propagate from one end of the network to the other.  Round Trip Time (RTT): how long it takes for a bit to get from one end of the network to the other and back.

19. Key equations (Perceived) Latency = Propagation + Transmit + Queue Propagation = Distance / SpeedOfLight Transmit = Size / Bandwidth

20. Some Units l Mbps = 10**6 bits/sec l byte = 8 bits l KB = 2**10 bytes (= 8,192 bits) l MB = 2**20 bytes (= 8,388,608 bits) l ms = 10**-3 seconds l  s = 10**-6 seconds

21. Bandwidth vs. Latency l Which is the better deal:  Improve your bandwidth from 1 Mbps to 100 Mbps, or  Improve your RTT from 100 ms to 1 ms? l The answer depends on what you need to send.

22. Latency Bound l Send a 1 byte message Perceived Latency 100 ms1 ms 1 Mbps100.008 ms1.008 ms 100 Mbps100.00008 ms1.00008 ms Transmit Time 1 Mbps8  s 100 Mbps.08  s 99%.008%.8%

23. Bandwidth Bound l Send a 25 MB message Transmit Time 1 Mbps3.5 min 100 Mbps21 sec Perceived Latency 100 ms1 ms 1 Mbps210.1 sec210.001 sec 100 Mbps 21.1 sec 21.001 sec.05%.5%90%

24. Perceived latency

25. Other measures l Bit width (seconds): 1 bit / bandwidth. l Delay x Bandwidth (bits). l Instructions per mile: number of instructions a machine can execute in the time it takes to send a bit for a mile.

26. Cycles per mile example l RTT from Penn to Stanford: 120ms l Best in principle is 48ms l On 400 MHz workstation, 48 million cycles elapse in that time. l 6000 mile round trip means 8000 cycles per mile.

27. Bit width

28. Architecture l Many requirements introduce complexity. l Complexity can be controlled by abstraction. l Fundamental idea is to create a separation of concerns, so that each module can focus primarily on its own objectives.

29. Protocol “Stack” idea l Realization of layered model l Optimizations designed to improve performance  e.g., reducing copying  e.g., removing functional redundancy l Applications at top, cable at bottom

30. Applications on hosts Madison Chicago Saul Eniac Application

31. Basic stack Process-to-Process Channels Host-to-Host Connectivity Hardware Application Programs Request / Reply ChannelMessage Stream Channel

32. Protocol Stacks illustrated: l Service and peer interfaces Protocol High-level Object High-level Object Peer-to-peer interface Service interfaceService Interface Host #1 Host #2

33. Protocol graph

34. Encapsulation

35. Example: TCP/IP over Ethernet ETH Src ETH Dst IP Header TCP Packet (Sequence #, Checksum & Data)

36. Encapsulation Challenges l Copying Data between layers l Size differences  ATM: 53 bytes, 48 payload  Ethernet: 1536 bytes, 1500 payload  IP: variable to 65,536  Solved with fragmentation and reassembly

37. OSI network stack

38. Internet protocol graph FTPHTTPNVTFTP TCP UDF IP EthernetATMFDDI