Packet Transmission (Part I)

1. Packet Transmission (Part I) Packets, Frames, and Error Detection

2. The Problem • Cannot afford individual network connection per pair of computers • Reasons • Installing wires consumes time and money • Maintaining wires consumes money (especially long-distance connections)

3. Solution • Network has • shared central core • many attached stations • only one source can transmit data at a time

4. The Problem With Sharing • Demand high • Some applications have large transfer • Some applications cannot wait • Need mechanism for fairness

5. Packet Switching Principle • Solution for fairness • Divide data into small units called packets • Allow each station opportunity to send a packet before any station sends another • Form of statistical time-division multiplexing

6. Illustration Of Packet Switching • Packets from different sources are interleaved • Send one packet, allow other stations opportunity to send before sending again • Efficient use of resources (since they are used on a demand). Nobody reserves a lane on a freeway • Can accommodate bursty traffic (as opposed to circuit-switching where transmission is at constant rate in telephone network).

7. Features Of Packet-Switching • Store-and-forward: intermediate nodes (e.g., routers) store (buffer) incoming packets, process them and forward them to the appropriate outgoing link. • Allows for flexibility and robustness. Packets can travel through alternative paths (adaptive routing). • Undesired situations such congestion, long delays may occur.

8. Packet Details • Depend on underlying network • minimum/maximum size • format • Hardware packet called a frame

9. Example Frame Format Used With RS-232 • RS-232 is character-oriented • Special characters • start of header (soh) • end of text (eot)

10. When Data Contains Special Characters • Translate to alternative form • Called byte stuffing • Example

11. Illustration Of Frame With Byte Stuffing • Stuffed frame longer than original

12. Bit-Oriented Frames • Delineate frame with a special bit pattern: 01111110 • If bit pattern occurs at the data, use bit-stuffing • sender: insert bit 0 after five consecutive 1’s • receiver: if receive five consecutive 1’s, check next bit • if it is 0, remove it • if they are 10, mark the end of the frame • if they are 11, error

13. Transmission Errors • Data can be corrupted during transmission • Bit loss • Bit values changed • Frame includes additional information to detect/correct error • Set by sender • Checked by receiver • Statistical guarantee

14. Common Approaches • Forward error control • each transmitted frame contains information to detect and correct errors • high overhead, used only if retransmission is impractical, such as simplex transmission, long delay, high error rate • Backward error control • transmitted frame only contains information to detect errors • retransmission used to recover from errors • low error rate in data communication

15. Common Approaches (cont’d) • No error control • real time traffic • not requires a 100% error rate • example: real time voice and video

16. Error Detection And Recovery • Parity bit • One additional bit per character • Can use • Even parity • Odd parity • Can not handle error that changes two bits • Block sum check • one parity bit for each row and each column of a block of bits • Can detect any single, double and triple bit errors, but may not detect some combination of errors

17. Error Detection And Recovery • Checksum • Treat data as a sequence of integers • Compute and send arithmetic sum • Handles multiple bit errors • Cannot handle all errors

18. Error Detection And Recovery • Cyclic Redundancy Check (CRC) • Mathematical function for data • More complex to compute • Handles more errors – can detect 99.98% errors

19. Example Checksum Computation • Checksum computed over data • Checksum appended to frame • See example in class

20. Illustration Of Errors A Checksum Fails To Detect • Second bit reversed in each item • Checksum is the same

21. Cyclic Redundancy Check (CRC) • Idea: given a k-bit frame, the transmitter generates an n-bit sequence known as the Cyclic Redundancy Check (CRC), so that the resulting (k+n)-bit frame is exactly divisible (modulo-2) by some predetermined number called a generator. • Example in class • Commonly used polynomial generators: • CRC-12: • CRC-16: • CRC-CCITT:

22. CRC (cont’d) • Assume the degree of the generator is n • n zero bits appended to the end of the frame • Modulo-2 operations used to determine the remainder CRC • Sender sends the data plus CRC • Receiver performs the modulo-2 operations on (k+n)-bit frame (original data plus CRC) using the same generator • If the remainder is zeros, assuming NO error. Otherwise, discard the frame.

23. Building Blocks For CRC • Exclusive or • Shift register • a shows status before shift • b shows status after shift • output same as top bit

24. Example Of CRC Hardware • Computes 16-bit CRC • Registers initialized to zero • Bits of message shifted in • CRC found in registers

25. Illustration Of Frame Using CRC • CRC cover data only

26. Summary • Packet technology • Invented to provide fair access in shared network • Sender divides data into small packets • Hardware packets called frames • Can use packet-switching with RS-232 • Special character delimit beginning and end of frame • Bye-stuffing needed when special characters appear in data

27. Summary (cont’d) • To detect data corruption • Sender adds information to packet • Receiver checks • Techniques • Parity bit • Block sum check • Checksum • Cyclic Redundancy Check (CRC) • Provide statistical guarantees

28. Reading Materials • Chapter 7: All Sections