Data Communication and
Networks
Lecture 3
Data Link Control
(Part 1)
September 22, 2005

Data Link Layer
Source node
Application
Presentation
Session transport
Network
Data link
Physical
Packets
Frames
Intermediate node
Network
Bits
Data link
Physical
Signals
Destination node
Application
Presentation
Session transport
Network
Data link
Physical

OSI Layer 2 - Data Link Layer
 Primary function is to make Layer 1 into what appears to be a channel free of undetected errors
 Deals with data in chunks (typically 100s-1000s of bytes) generally called Frames
 This layer must create/recognize frame boundaries
 Physical layer does not care about frames, only bits
 Requires special bit patterns to signal boundaries
 May have to deal with possibility of pattern appearing in data
 Implements data link control protocols, such High Level
Data Link Control (HDLC)

Encoding Information Frames
Typical Fields in a Frame
Start
Frame
Delimiter
Destination
Address
Source
Address
Frame
Control
Data
Check sum

DLL Operation
Sender
NL correct and ordered
DLL
Retransmit if timeout
CRC
Frame
PL
ACK
CRC
Receiver
ACK if correct

Data Link Layer Services
 Flow Control :
 pacing between senders and receivers
 Error Detection :
 errors are caused by signal attenuation and noise.
 Receiver detects presence of errors:
 it signals the sender for retransmission or just drops the corrupted frame
 Error Correction :
 mechanism for the receiver to locate and correct the error without resorting to retransmission

Flow Control
 Ensuring the sending entity does not overwhelm the receiving entity
 Preventing buffer overflow
 Transmission time
 Time taken to emit all bits into medium
 Propagation time
 Time for a bit to traverse the link

Model of Frame Transmission

Stop and Wait
 Source transmits frame
 Destination receives frame and replies with acknowledgement
 Source waits for ACK before sending next frame
 Destination can stop flow by not send ACK
 Works well for a few large frames

Fragmentation
 Large block of data may be split into small frames
 Limited buffer size
 Errors detected sooner (when whole frame received)
 On error, retransmission of smaller frames is needed
 Prevents one station occupying medium for long periods
 Stop and wait becomes inadequate when data is fragmented into several small frames.

Stop and Wait Link Utilization

Sliding Windows Flow Control
 Allow multiple frames to be in transit
 Receiver has buffer W long
 Transmitter can send up to W frames without
ACK
 Each frame is numbered
 ACK includes number of next frame expected
 Sequence number bounded by size of field (k)
 Frames are numbered modulo 2 k

Example Sliding Window

Sliding Window Example
An example of sliding window in action …..

Sliding Window Enhancements
 Receiver can acknowledge frames without permitting further transmission (Receive Not
Ready)
 Must send a normal acknowledge to resume
 If duplex, use piggybacking
 If no data to send, use acknowledgement frame
 If data but no acknowledgement to send, send last acknowledgement number again, or have ACK valid flag (TCP)

Error Detection
 Additional bits added by transmitter for error detection code
 Parity
 Value of parity bit is such that character has even
(even parity) or odd (odd parity) number of ones
 Even number of bit errors goes undetected
 Checksum can be a simple XOR operation of bits to be checked
 DL protocols use more sophisticated methods, like Cyclic Redundancy Check (CRC)

Cyclic Redundancy Check
 For a block of k sequence (FCS) bits transmitter generates n bit
 Transmit k+n some number bits which is exactly divisible by
 Receive divides frame by that number
 If no remainder, assume no error
 For math, see Stallings chapter 6

Cyclic Redundancy Check
 Example: 16-Bit CRC for HDLC
 Performed on Address, Control, Data Fields
 Detects a Variety of Error Conditions
 Data Message is Represented as a Polynomial, M polynomial of order i = i
(x) = a
0
+ a
1 x + a
2 x 2 + a
3 x 3 + . . . + a i
 Data Message = a i a i-1
… a
3 a
2 a
1 a
0 x i where a i
= 0, 1
Start
Frame
Delimiter
Destination
Address
Source
Address
Frame
Control
Data
CRC

CRC Generation
 Given a Specific Polynomial G(x) called the generator, and a Data Message M(x), Calculate a Frame Check Sequence (FCS), or Checksum
 Append the FCS (Checksum) to the Data
Message for Transmission
 At the Receiver, Recalculate the Checksum, and
Check with the Transmitted Value

CRC Algorithm
 Multiply x r M(x), where r = degree of G(x)
 Add r Zeros to M(x), or Shift M(x) Left
 Divide x r M(x)/G(x) = R(x)
 R(x) = Reminder of Divide Operation
 Transmit x r M(x) + R(x) = T(x)
 XOR the Shifted Message and R(x)
 At the Receiver, Recalculate the R(x) and Check
Equal to the Transmitted FCS
 Errors Occurred in Transmission if Not equal

CRC Error Detection
 All Single Bit Errors
 All Double Bit Errors
 Any Odd Number of Errors
 Any Burst Error for Which the Length of the
Burst Error is Less Than the Length of the
Divisor Polynomial, G(x)
 Most Larger Burst Errors
 Several G(x) Polynomials Have Been Used

Example of G(x) Polynomials
 CRC-12
 X 12 + X 11 + X 3 + X 2 + X + 1
 CRC-16
 X 16 + X 15 + X 2 + 1
 CRC-CCITT
 X 16 + X 15 + X 5 + 1
 CRC-32
 X 32 + X 26 + X 23 + X 22 + X 16 + X 12 + X 11 + X 10
+ X 8 + X 7 + X 5 + X 4 + X 2 + X + 1
 CRC’s Are Implemented in Shift registers

Error Control
 Detection and correction of errors
 Lost frames
 Damaged frames
 Automatic repeat request
 Error detection
 Positive acknowledgment
 Retransmission after timeout
 Negative acknowledgement and retransmission

Automatic Repeat Request
(ARQ)
 Stop and wait
 Go back N
 Selective reject (selective retransmission)

Stop and Wait
 Source transmits single frame
 Wait for ACK
 If received frame damaged, discard it
 Transmitter has timeout
 If no ACK within timeout, retransmit
 If ACK damaged,transmitter will not recognize it
 Transmitter will retransmit
 Receive gets two copies of frame
 Use ACK0 and ACK1

Stop and Wait -
Diagram

Stop and Wait - Pros and Cons
 Simple
 Inefficient

Go Back N (1)
 Based on sliding window
 If no error, ACK as usual with next frame expected
 Use window to control number of outstanding frames
 If error, reply with rejection
 Discard that frame and all future frames until error frame received correctly
 Transmitter must go back and retransmit that frame and all subsequent frames

Go Back N - Damaged Frame
 Receiver detects error in frame i
 Receiver sends rejectioni
 Transmitter gets rejectioni
 Transmitter retransmits frame subsequent i and all

Go Back N - Lost Frame (1)
 Frame i lost
 Transmitter sends i+1
 Receiver gets frame i+1 out of sequence
 Receiver send reject i
 Transmitter goes back to frame retransmits i and

Go Back N - Lost Frame (2)
 Frame i lost and no additional frame sent
 Receiver gets nothing and returns neither acknowledgement nor rejection
 Transmitter times out and sends acknowledgement frame with P bit set to 1
 Receiver interprets this as command which it acknowledges with the number of the next frame it expects (frame i )
 Transmitter then retransmits frame i

Go Back N - Damaged
Acknowledgement
 Receiver gets frame acknowledgement ( i and send i+1 ) which is lost
 Acknowledgements are cumulative, so next acknowledgement ( i+n ) may arrive before transmitter times out on frame i
 If transmitter times out, it sends acknowledgement with P bit set as before
 This can be repeated a number of times before a reset procedure is initiated

Go Back N - Damaged Rejection
 As for lost frame (2)

Go Back N -
Diagram

Selective Reject
 Also called selective retransmission
 Only rejected frames are retransmitted
 Subsequent frames are accepted by the receiver and buffered
 Minimizes retransmission
 Receiver must maintain large enough buffer
 More complex logic in transmitter

Selective Reject -
Diagram