1. CSE331: Introduction to Networks and Security Lecture 2 Fall 2002

2. CSE331 Fall 20022 Announcements HW1 is due Friday –Homework and lectures are available only inside upenn.edu domain CSE331 newsgroup: upenn.cis.cse331

3. CSE331 Fall 20023 Recap General network concepts –Nodes and links Design goals –Connectivity –Resource sharing –Functionality –Performance Direct link networks –Bandwidth (bits/sec) –Latency = prop. delay + transmit + queue

4. CSE331 Fall 20024 Bandwidth vs. Latency Which is the better deal: –Improve your bandwidth from 1 Mbps to 100 Mbps, or –Improve your RTT from 100 ms to 1 ms? The answer depends on what you need to send.

5. CSE331 Fall 20025 Latency Bound Send 1 byte 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%

6. CSE331 Fall 20026 Bandwidth Bound Send 25 MB 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%

7. CSE331 Fall 20027 Tradeoffs RTT from Penn to Stanford is approx. 100ms 800 MHz workstation –80 million cycles elapsed in that time Data compression –Trades machine cycles for bandwidth (Question: Why is RTT important?)

8. CSE331 Fall 20028 Network Architecture Need a way to deal with the complex requirements of a network Use abstraction to separate concerns –Decompose problem –Modular changes Protocol Layers

9. CSE331 Fall 20029 Protocols A protocol is a specification of an interface between modules (often on different machines) Sometimes “protocol” is used to mean the implementation of the specification. Examples?

10. CSE331 Fall 200210 Interprocess communication Zeta Central Saul Eniac Application

11. CSE331 Fall 200211 Example Protocol Stack Process-to-Process Channels Host-to-Host Connectivity Hardware Application Programs Request / Reply ChannelMessage Stream Channel

12. CSE331 Fall 200212 Protocol Interfaces Service Interfaces –Communicate up and down the stack Peer Interfaces –Communicate to counterpart on another host Protocol High-level Object High-level Object Peer-to-peer interface Service interface Service Interface Host #1 Host #2

13. CSE331 Fall 200213 Example Protocol Graph File App RRP HHP Host 1Host 2 Video App MSP File App RRP HHP Video App MSP

14. CSE331 Fall 200214 Encapsulation File App RRP HHP Host 1Host 2 Video App MSP File App RRP HHP Video App MSP DATA RRPDATARRP HHP DATARRP DATA

15. CSE331 Fall 200215 Open Systems Interconnection (OSI) Application Presentation Session Transport Network Data Link Physical End Host Reference model – not actual implementation. Transmits messages (e.g. FTP or HTTP) Data format issues (e.g. big- vs. little-endian) Manages multiple streams of data Process to process protocols Routes packets among nodes in network Packages bit streams into frames Transmits raw bits over link

16. CSE331 Fall 200216 Open Systems Interconnection (OSI) Application Presentation Session Transport Network Data Link Physical Network Data Link Physical Network Data Link Physical Application Presentation Session Transport Network Data Link Physical Nodes in network End Host

17. CSE331 Fall 200217 Internet Protocol Graph FTPHTTPNVTFTP TCP UDP IP EthernetATMFDDI

18. CSE331 Fall 200218 Problem: Physical connection Transmitting signals Encoding & decoding bits Error detection and correction Reliable transmission

19. CSE331 Fall 200219 Signaling Components Network Adapter Network Adapter Host signal Signaling Components Network adapters encode streams of bits into signals. Simplification: Assume two discrete signals—high and low. Practice: Two different voltages on copper link.

20. CSE331 Fall 200220 Bit Encodings Bits: 0010111101000010 NRZ Non-return to zero 1 = high, 0 = low Problem: Can lead to long sequences of high or low signals.

21. CSE331 Fall 200221 NRZ Problem 1: baseline wander Receivers keeps average signal seen so far –Significantly higher than average = 1 –Significantly lower than average = 0 Long periods of high/low changes average –Makes it harder to detect significant changes Bits: 0010111101000010 NRZ

22. CSE331 Fall 200222 NRZ Problem 2: clock recovery Sender & Receiver are driven by clocks –Must keep clocks synchronized to recover bits –When signal changes, can resynchronize clocks Long periods of high/low lead to clock drift Bits: 0010111101000010 NRZ

23. CSE331 Fall 200223 NRZI: Non-return Transition from current signal to encode “1” Stay at current signal to encode “0” Does not help for consecutive 0’s Bits: 0010111101000010 NRZ NRZI

24. CSE331 Fall 200224 Manchester Encoding Exclusive-OR of clock and NRZ encoding Doubles rate at which signal transitions are made – less time for receiver to detect changes Bits: 0010111101000010 NRZ Manchester Clock

25. CSE331 Fall 200225 Efficiency Bit rate vs. Baud rate NRZ and NRZI: bit rate = baud rate = 100% –Suffer from baseline wander and clock drift Manchester: bit rate = ½ baud rate = 50% –Prevents baseline wander and clock drift Can you do better? (Question: can bit rate exceed baud rate?)

26. CSE331 Fall 200226 4B/5B Encode 4 bits of data using 5 bits Choose codes to avoid long sequences of 0’s Send 5 bit sequences using NRZI 80% efficiency Encoding –No more than one leading 0 –No more than two trailing 0’s 000011110 000101001 001010100 001110101 010001010 010101011 011001110 011101111 ……