Global Information Systems and Software Technology (GISST)

Download Report

Transcript Global Information Systems and Software Technology (GISST)

University of British Columbia

CICS 515 (Part 1) Computer Networks

Lectures 3 – Transport Layer (TCP/UDP) Principle of Reliable Data Transfer

Instructor: Dr. Son T. Vuong Email: [email protected]

May 17 , 2012

The World Connected

Jump to first page

Lecture 3: Transport Layer

(Kurose Ch. 3)

Our goals:

understand principles behind transport layer services:  multiplexing/demultiplexing  reliable data transfer  flow control  congestion control learn about transport layer protocols in the Internet:  UDP: connectionless transport  TCP: connection oriented transport  TCP congestion control

CICS 515 – Summer 2012 © Dr. Son Vuong

Outline (Ch. 3)

3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport:

UDP

3.5 Connection-oriented transport:

TCP

 segment structure  reliable data transfer  flow control  connection management 3.4 Principles of reliable data transfer ( Sliding Window Protocol ) 3.6 Principles of congestion control 3.7 TCP congestion control

CICS 515 – Summer 2012 © Dr. Son Vuong

Internet Architecture

Defined by Internet Engineering Task Force (IETF) Hourglass Design Application vs Application Protocol (FTP, HTTP)

FTP HTTP TCP NV TFTP UDP IP

NET 1 NET 2 …

CICS 515 – Summer 2012 © Dr. Son Vuong

NET n Application TCP UDP IP Network 4

Example TCP/IP internet

H7 Network 1 (Ethernet) R3

H8 H1

H2 H3 Network 2 (Ethernet) R1 Network 4 (point-to-point) R2 H4 Network 3 (FDDI) H1 TCP IP IP R1 H5 IP R2 ETH ETH FDDI

CICS 515 – Summer 2012 © Dr. Son Vuong

FDDI PPP H6 PPP IP R3 ETH IP H8 TCP ETH 5

Transport services and protocols

provide

logical communication

between app processes running on different hosts transport protocols run in end systems  send side: breaks app messages into segments , passes to network layer  rcv side: reassembles segments into messages, passes to app layer more than one transport protocol available to apps  Internet: TCP and UDP

CICS 515 – Summer 2012 © Dr. Son Vuong

application transport network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical application transport network data link physical 6

Transport services and TCP Segments

TCP sends data in

Segments

(same concept as Packets, just a different name).

Data is written by the application program into the TCP buffers, and at some later time each Segment will be transmitted.

Application process Write bytes TCP Send buffer Application process Read bytes TCP Receive buffer Segment Segment … Segment

CICS 515 – Summer 2012 © Dr. Son Vuong

Transmit segments 7

Transport vs. network layer

network layer:

logical communication between hosts

transport layer:

logical communication between processes  relies on, enhances, network layer services Household analogy:

12 kids sending letters to 12 kids

processes = kids app messages = letters in envelopes hosts = houses transport protocol entities = Ann and Bill network-layer service = postal service

CICS 515 – Summer 2012 © Dr. Son Vuong

Internet transport-layer protocols

reliable, in-order delivery ( TCP )   congestion control flow control  connection setup unreliable, unordered delivery: UDP  no frills extension of “best effort” IP services not available: application transport network data link physical  delay guarantees  bandwidth guarantees network data link physical network data link physical network data link physical network data link physical network data link physical application transport network data link physical

CICS 515 – Summer 2012 © Dr. Son Vuong

Transport services and protocols

provide

logical communication

CICS 515 – Summer 2012 © Dr. Son Vuong

Multiplexing and demultiplexing

Ch. 3

3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport:

UDP

3.5 Connection-oriented transport:

TCP

 segment structure  reliable data transfer  flow control  connection management 3.4 Principles of reliable data transfer 3.6 Principles of congestion control 3.7 TCP congestion control

CICS 515 – Summer 2012 © Dr. Son Vuong

Multiplexing/demultiplexing

Demultiplexing at rcv host: delivering received segments to correct socket = socket = process Multiplexing at send host: gathering data from multiple sockets, enveloping data with header (later used for demultiplexing) application P3 transport network link physical host 1

CICS 515 – Summer 2012 © Dr. Son Vuong

application transport network link physical host 2 P2 P4 application transport network link physical host 3 12

Multiplexing/demultiplexing

Recall:

segment

- unit of data exchanged between transport layer entities  aka TPDU: transport protocol data unit P1 segment header segment application-layer data Hn Ht M segment P3 M application transport network Demultiplexing: delivering received segments to correct app layer processes Receiver P2 M M application transport network M P4 application transport network

CICS 515 – Summer 2012 © Dr. Son Vuong

Multiplexing/demultiplexing: examples

host A source port: x dest. port:

server B source port:

dest. port: x port use: simple

telnet

app Web client host C Source IP: C Dest IP: B source port: y dest. port: 80 Source IP: C Dest IP: B source port: x dest. port: 80 Web client host A

CICS 515 – Summer 2012 © Dr. Son Vuong

Source IP: A Dest IP: B source port: x dest. port: 80 Web server B port use: Web server 14

3.1 . Transport Mux Peer Instruction Question 3.1

Through a specific

port

(e.g. port 80), an application process (e.g. Web server) can communicate with multiple (the following number of) remote (e.g. client) processes:

Answer: (A)

Exactly 1

(B)

(C) (E)

None of the above 3

(D)

An arbitrary number

CICS 515 – Summer 2012 © Dr. Son Vuong

3.1b

. Transport Mux Peer Instruction Question 3.1b

Through a specific

socket

, an application process (e.g. Web server process or client process) can communicate with multiple (the following number of) remote (e.g. client or server) processes:

Answer: (A)

Exactly 1

(B)

(C) (E)

None of the above 3

(D)

An arbitrary number

CICS 515 – Summer 2012 © Dr. Son Vuong

Outline (Ch. 3)

3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport:

UDP

3.5 Connection-oriented transport:

TCP

 segment structure  reliable data transfer  flow control  connection management 3.4 Principles of reliable data transfer 3.6 Principles of congestion control 3.7 TCP congestion control

CICS 515 – Summer 2012 © Dr. Son Vuong

UDP: User Datagram Protocol [RFC 768]

“no frills,” “bare bones” Internet transport protocol “best effort” service, UDP segments may be:  lost Why is there a UDP?

no connection establishment (which can add delay)  delivered out of order to app

connectionless:

simple: no connection state at sender, receiver   no handshaking between UDP sender, receiver each UDP segment handled independently of others small segment header no congestion control: UDP can blast away as fast as desired

CICS 515 – Summer 2012 © Dr. Son Vuong

UDP: more

often used for streaming multimedia apps  loss tolerant  rate sensitive other UDP uses  DNS  SNMP reliable transfer over UDP: add reliability at application layer  application-specific error recovery!

CICS 515 – Summer 2012 © Dr. Son Vuong

UDP Header Format

0 16 SrcPort DstPort Checksum Length Data 31

Checksum

computed over: Pseudoheader + UDP header + Data Pseudoheader = src, dest IP addresses + Protocol no. (17) + UDP length

CICS 515 – Summer 2012 © Dr. Son Vuong

UDP Message Queue

Application process Application process Application process Ports Queues Packets demultiplexed UDP

CICS 515 – Summer 2012 © Dr. Son Vuong

Packets arrive 21

UDP checksum

Goal: detect “errors” (e.g., flipped bits) in transmitted segment Sender: treat segment contents as sequence of 16-bit integers checksum: addition (1’s complement sum) of segment contents sender puts checksum value into UDP checksum field

CICS 515 – Summer 2012 © Dr. Son Vuong

Receiver: compute checksum of received segment check if computed checksum equals checksum field value:  NO - error detected  YES - no error detected.

But maybe errors nonetheless?

….

More later 22

Internet Checksum Example

Note  When adding numbers, a carryout from the most significant bit needs to be added to the result Example: add two 16-bit integers

1 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1

wraparound

1 1 0 1 1 1 0 1 1 1 0 1 1 1 0 1 1

sum checksum

1 0 1 1 1 0 1 1 1 0 1 1 1 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0 0 0 1 1 Checksum at Rec 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 CICS 515 – Summer 2012 © Dr. Son Vuong

Outline (Ch. 3)

3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport:

UDP

3.5 Connection-oriented transport:

TCP

 segment structure  reliable data transfer  flow control  connection management 3.4 Principles of reliable data transfer 3.6 Principles of congestion control 3.7 TCP congestion control

CICS 515 – Summer 2012 © Dr. Son Vuong

Principles of Reliable data transfer

important in app., transport, link layers top-10 list of important networking topics!

characteristics of unreliable channel will determine complexity of reliable data transfer protocol (rdt)

CICS 515 – Summer 2012 © Dr. Son Vuong

Reliable data transfer: getting started

rdt_send():

called from above, (e.g., by app.). Passed data to deliver to receiver upper layer

deliver_data(): rdt

called by to deliver data to upper send side receive side

udt_send():

called by rdt, to transfer packet over unreliable channel to receiver

CICS 515 – Summer 2012 © Dr. Son Vuong rdt_rcv():

called when packet arrives on rcv-side of channel 26

Reliable Data Link Protocols

Functions of data link protocols:

 Flow control  Error control  Sequencing

Stop-and-wait (alternating-bit) protocol Sliding window protocol CICS 515 – Summer 2012 © Dr. Son Vuong

Protocols

Used for communications between entities in a distributed system Must speak the same language Key Elements:  data part (syntax): data format  control part (semantics): procedures Formal Description Techniques (FDTs) e.g. CFSM, Estelle, SDL, LOTOS

CICS 515 – Summer 2012 © Dr. Son Vuong

rdt1.0

: reliable transfer over a reliable channel

underlying channel perfectly reliable  no bit errors, no loss of packets  no flow control separate FSMs for sender, receiver:  sender sends data into underlying channel  receiver read data from underlying channel

Wait for call from above rdt_send(data) packet = make_pkt(data) udt_send(packet)

sender

Wait for call from below rdt_rcv(packet) extract (packet,data) deliver_data(data)

receiver

CICS 515 – Summer 2012 © Dr. Son Vuong

rdt2.0

: Protocol for unreliable channel

sender Stop-And-Wait Protocol

 Error control  Flow control

receiver rdt_send(data) sn d pkt = make_pkt(data, checksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) Wait for call from above Wait for ACK or NAK udt_send(sndpkt) rdt_rcv(rcvpkt) && corrupt(rcvpkt) udt_send(NAK) Wait for call from below rdt_rcv(rcvpkt) && isACK(rcvpkt)

rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK) CICS 515 – Summer 2012 © Dr. Son Vuong

rdt2.0: operation with no errors

rdt_send(data) sn

pkt = make_pkt(data, checksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) Wait for call from above Wait for ACK or NAK udt_send(sndpkt) rdt_rcv(rcvpkt) && corrupt(rcvpkt) udt_send(NAK) rdt_rcv(rcvpkt) && isACK(rcvpkt) L Wait for call from below rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK)

CICS 515 – Summer 2012 © Dr. Son Vuong

rdt2.0: error scenario

rdt_send(data) snkpkt = make_pkt(data, checksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) Wait for call from above Wait for ACK or NAK udt_send(sndpkt) rdt_rcv(rcvpkt) && isACK(rcvpkt) L

CICS 515 – Summer 2012 © Dr. Son Vuong

rdt_rcv(rcvpkt) && corrupt(rcvpkt) udt_send(NAK) Wait for call from below rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK) 32

Peer Instruction 3.2

Question

Is NAK necessary ( T or F ) and sufficient ( T or F ) ?

Answer:

(A): (B): (C): F, F F, T T, F (D): T, T (E): B, C and D CICS 515 – Summer 2012 © Dr. Son Vuong

Formal Specification of Protocols

Communicating Finite State Machines (CFSM)

- DATA SENDER 0 RECEIVER + DATA 0 (a) Eutopia +Error / -NAK 0 - DATA [TimeOut] / - DATA 1 + ACK +NAK / - DATA + DATA 0 1 - ACK (b) Stop-And-Wait Protocol CICS 515 – Summer 2012 © Dr. Son Vuong

Alternating Bit Protocol

SENDER RECEIVER

+ ACK 1

ACK 0 + DATA 1 + Error - ACK 1

0 3 0 3

- DATA 0

[TimeOut]

+ DATA 0 - DATA 1 + DATA 1

1 2 1 2 +

ACK 1 + DATA 0 + Error + ACK 0

(a) ABP with 4 states

- ACK 0 Init I=0

- S0 [T] S1 + [I==J] I=I+1

+ [I==J] Init J=1 + Error

S0 + [I!=J] J=J+1 t1

+ [I!=J]

(b) ABP with 2 states CICS 515 – Summer 2012 © Dr. Son Vuong S1 - Trans Label t1 From S0 To S1 When

Provided

I!= J

{

J=J+1

}

Alternating Bit Protocol (Correct Version?)

SENDER +

ACK 1 - DATA 0

ACK 1

RECEIVER

+ ACK 1

ACK 0 + Error - ACK 1

0 3 1 [TimeOut]

+ ACK 0 + DATA 0 - DATA 1

2 +

ACK 0 + DATA 0 + Error

(a) ABP with 4 states 0 1

+ DATA 1 - ACK 0

+ DATA 0

+ DATA 1 + Error + DATA 1 + Error Init J=0 Init I=0 + [I=J]

S0 - + [T] + [I==J] [I!=J] J=J+1 S1 I=I+1

+ Error * + [I=J]

(b) ABP with 2 states CICS 515 – Summer 2012 © Dr. Son Vuong

S0 * S1

+ Error

- * :

+ [I!=J] 36

Alternating Bit Protocol (Correct Version)

SENDER +

ACK 1 - DATA 0

ACK 1

RECEIVER

+ ACK 1

ACK 0 + Error - ACK 1

0 3 1 [T] + ACK 1 + ACK 0

+ ACK 0 + DATA 0 - DATA 1

2 +

ACK 0 + DATA 0 + Error

(a) ABP with 4 states 0 1

+ DATA 1 - ACK 0

+ DATA 0

+ DATA 1 + Error + DATA 1 + Error Init J=0 Init I=0 + [I=J]

S0 - + [T] + [I==J] [I!=J] J=J+1 S1 I=I+1

+ Error * + [I=J]

(b) ABP with 2 states CICS 515 – Summer 2012 © Dr. Son Vuong

S0 * S1

+ Error

- * :

+ [I!=J] 37

How to Recover from Errors?

If a bad frame is detected, what do we do?

 Forward Error Correction (FEC) - Using parity bits as a guide to fixing single bit errors.

 Respond to the sender by transmitting a message:  NAK = Negative Acknowledgement  ACK = Positive Acknowledgement (if no error)  Let the sender

timeout

and retransmit. Also called Automatic Repeat reQuest (ARQ)

CICS 515 – Summer 2012 © Dr. Son Vuong

Sender Receiv er Sender Receiv er Frame ACK Frame ACK Frame ACK Sender (a) (c )

Timeout for frame loss recovery

Receiv er Sender Receiv er Frame Frame ACK Frame Frame ACK ACK (b)

CICS 515 – Summer 2012 © Dr. Son Vuong

(d) 39

Questions:

What should we set the sender’s timeout value to be?

What is Piggybacking ? How does it save us from sending ACKs?

CICS 515 – Summer 2012 © Dr. Son Vuong

Efficiency: Stop-and-wait protocol

sender first packet bit transmitted, t = 0 last packet bit transmitted, t = L / R receiver RTT first packet bit arrives last packet bit arrives, send ACK ACK arrives, send next packet, t = RTT + L / R  Example : 1 Gbps link, 15 ms e-e prop. delay, 1KB packet: U sender = L / R RTT + L / R .

008 = 30.008

CICS 515 – Summer 2012 © Dr. Son Vuong

= 0.00027 microsec onds 41

Stop-and-Wait Protocol Efficiency

α = CICS 515 – Summer 2012 © Dr. Son Vuong

Efficiency of Stop and Wait Protocol

U = D/(D+H+A+2CI) =~ 1/(1 + 2*Tp/Td) U = 1/(1+2  ) U = channel utilization (efficiency) D = # bits in frame body H = # bits in frame header A = # bits in ACK C = channel capacity (bits/sec) I = Propagation Delay = Distance/Speed (300, 200, 230 m/  s for vacuum, fiber, copper)

CICS 515 – Summer 2012 © Dr. Son Vuong

Link Utilization (Efficiency) Example

CICS 515 – Summer 2012 © Dr. Son Vuong

Maximizing Link Efficiency

Example: 1.5Mbps link x 45ms RTT = 67.5Kbits (8KBytes) for the Delay x Bandwidth product.

Assuming frame size of 1KB, stop-and-wait uses about one-eighth of the link's capacity. Solution: Want the sender to be able to transmit up to 8+1 = 9 frames before having to wait for an ACK.

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

Pipelined (sliding window) protocols

Pipelining: sender allows multiple, “in-flight”, yet-to-be acknowledged pkts  range of sequence numbers must be increased  buffering at sender and/or receiver Two generic forms of pipelined protocols:

go-Back-N, selective repeat

Efficiency: Stop-and-wait protocol

008 = 30.008

= 0.00027 microsec onds 48

Pipelining: increased utilization

receiver first packet bit transmitted, t = 0 last bit transmitted, t = L / R sender RTT first packet bit arrives last packet bit arrives, send ACK last bit of 2 nd last bit of 3 rd packet arrives, send ACK packet arrives, send ACK ACK arrives, send next packet, t = RTT + L / R Increase utilization by a factor of 3!

U sender = 3 * L / R RTT + L / R .

024 = 30.008

= 0.0008 microsecon ds 49

Sliding Window Protocol Efficiency

Efficiency: Stop-and-wait protocol

008 = 30.008

= 0.00027 microsec onds 51

Peer Instruction 3.2: Questions

  Consider the previous example:  1 Gbps link, 15 ms e-e propagation delay, e.g. RTT = 30 ms 1KB packet (1) What would be the window size to achieve 100% efficiency?

(A) 15 (B) 30 (C) 300 (D) 3751 (E) 4000 (F) None

(2) How many bits at minimum are needed for the sequence number ?

Sliding Window Diagram

Sliding Window Protocol Example

Dealing with Errors

What happens if a frame is lost, or is corrupted?

Solution 1:

Go back N

 Receiver discards all subsequent frames  Sender must retransmit all frames, starting at the lost/corrupted frame.

 This implies that RWS = 1  No buffering is needed at the receiver  Ordering of frames is guaranteed  A lot of bandwidth is wasted if error rate is high.

Dealing with Errors cont.

Solution 2:

Selective Repeat

 Receiver sends a NAK, with a list of frames that are missing/corrupted.

 Receiver stores subsequent frames.

 Sender retransmits the bad frame(s) only.

 Need to buffer up to RWS frames at the receiver and order them correctly before sending to a higher protocol layer.

Sender

LAR LFS SWS 0 1 2 3 4 5 6 7 8 Sender side: SWS: send window size LAR: last ACK received LFS: last frame sent Sender: if more data to send (LFS-LAR < SWS) then send data, LFS++ if recv’ed ACK for LAR+1 then LAR++ if timer expires then send LAR+1 CICS 515 – Summer 2012 © Dr. Son Vuong

Receiver

Receiver side: RWS: receive window size LFR: last frame received LAF: largest acceptable frame LFR RWS LAF 0 1 2 3 4 5 6 7 8 Receiver: if recv’ed K <= LFR, K > LAF then discard else store if K == LFR+1 then LFR++, LAF++ (slide window) ACK, largest in-order received frame CICS 515 – Summer 2012 © Dr. Son Vuong

Selective Repeat Example - Animated

Selective Repeat Example

Selective Repeat Example - Animated

Go-Back-N

Sender: k-bit seq # in pkt header “window” of up to N, consecutive unack’ed pkts allowed ACK(n): ACKs all pkts up to, including seq # n “cumulative ACK”  may receive duplicate ACKs (see receiver) timer for each in-flight pkt

timeout(n):

retransmit pkt n and all higher seq # pkts in window

GBN in action

GBN - Scenario 1 of Confusion When W =< M-1 is violated

Window W = Modulus M = 8

What’s wrong ?

GBN - Scenario 1 of Confusion When W =< M-1 is violated

Confusion when Window W = Modulus M = 8

ACK,0 can mean Ack to request for retransmission of I,0; ACK, 0 can also mean Ack for receipt of all I frames 0 to 7, and to request for new packet 0.

Thus, upon receiving ACK, 0 Sender is confused whether Receiver has received all I packets 0 to 7, or none at all.

GBN - Scenario 2 of Confusion When W =< M-1 is violated

I, 0 I, 1 I, 2 lost I, 3 I, 4 ACK, 1 ACK, 2 ACK, 3 ACK, 4 I, 5 ACK, 5 I, 6 I, 7 ACK, 6 ACK, 7 I, 0 lost ACK, 0

Window W = Modulus M = 8

What’s wrong ?

GBN - Scenario 2 of Confusion When W =< M-1 is violated

I, 0 I, 1 I, 2 lost I, 3 I, 4 ACK, 1 ACK, 2 ACK, 3 ACK, 4 I, 5 ACK, 5 I, 6 I, 7 ACK, 6 ACK, 7 I, 0 lost ACK, 0 CICS 515 – Summer 2012 © Dr. Son Vuong If all ACKs got lost, Sender will timeout and retransmit duplicate I frame 0 I, 0 which will unfortunately be treated wrongly by the Receiver as the new fresh I Frame 0.

Selective Repeat

receiver

individually

acknowledges all correctly received pkts  buffers pkts, as needed, for eventual in-order delivery to upper layer sender only resends pkts for which ACK not received  sender timer for each unACKed pkt sender window  N consecutive seq #’s  again limits seq #s of sent, unACKed pkts

Selective repeat: sender, receiver windows

Selective repeat in action

Example: Go-back-N RWS = 1

SWS = 3 0 1 2 0 1 2 0 1 2 3 sequence numbers SWS = 3 RWS = 1 Case 1: All 3 packets get lost Case 2: All 3 packets received OK Ack 2, Expecting “0” 2 RWS = 1 Sender cannot tell which case (from ACK 2) !

Selective repeat: dilemma

Example:

seq #’s: 0, 1, 2, 3 window size=3 receiver sees no difference in two scenarios!

incorrectly passes duplicate data as new in (a) Q: what relationship between seq # size and window size?

Example: Selective Repeat RWS = SWS

Modulus = 4 SWS = 3 0 1 2 3 0 1 2 3 0 Receiver:

or delayed and I received a

new 0

Did 3 just get lost or did all my ACKs get lost and this is the

old 0

RWS = 3 0 1 2 3 0 1 2 Sequence space = Modulus >= SWS + RWS = 2SWS i.e. SWS =< Modulus/2 CICS 515 – Summer 2012 © Dr. Son Vuong

Example

SWS = 2 0 1 2 3 0 1 2 3 0 Receiver: No ambiguity!

If both of my ACKs got lost duplicate 0 would be received. Otherwise,

a new 2 or 3

would be received.

RWS = 2 0 1 2 3 0 1 2 Modulus = 4 and SWS = 2 => SWS =< Modulus/2 is satisfied CICS 515 – Summer 2012 © Dr. Son Vuong

TCP vs GobackN vs Selective Repeat

GobackN

RWS =1,

W =< M-1

Discard out-of-sequence segments Cumulative Ack e.g. ACK n Retransmit All outstanding segments e.g. segments n, n+1, …, N

Selective Repeat

• • RWS = SWS >1,

W =< M/2

Store out-of-sequence segments received • Individual Ack ( Selective Ack is an option) e.g. ACK n •Retransmit only 1 (oldest un-acked) segment, e.g. segment n only

TCP

is a hybrid compbined GobackN and Selective Repeat ( as shown in red )

Sliding Window Protocol

Summary

Used in dala link layer (node-to-node) Provide reliable transfer, flow/error control, and sequencing Efficiency depends on  =Tp/Td and window size