Segment Format

7/31/2019 Segment Format

1/36

CS 640 1

Data structure and inputprocessing


2/36

CS 640 2

Transport Control Protocol

Outline

TCP objectives revisited

TCP basics

New algorithms for RTOcalculation


3/36

CS 640 3

TCP Overview

TCP is the most widely used Internet protocol Web, Peer-to-peer, FTP, telnet,

A two way, reliable, byte stream oriented end-to-endprotocol

Includes flow and congestion control

Closely tied to the Internet Protocol (IP)

A focus of intense study for many years

Our goal is to understand the RENO version of TCP RENO is most widely used TCP today RFC 2001 (now expired)

RENO mainly specifies mechanisms for dealing with congestion


4/36

CS 640 4

TCP Features Connection-oriented

Byte-stream app writes bytes

TCP sendssegments

app reads bytes

Reliable data transfer

Application process

Writebytes

TCP

Send buffer

Segment Segment Segment

Transmit segments

Application process

Readbytes

TCP

Receive buffer

Full duplex

Flow control: keep sender fromoverrunning receiver

Congestion control: keepsender from overrunning

network


5/36

CS 640 5

12.11 TCP Package

A TCP package involving a table called

Transmission Control Blocks, a set of timers, and

three software modules: main module, inputprocessing module, output processing module.


6/36

CS 640 6

TCP Package (Contd)

Transmission Control Block (TCBs)

To control the connection, TCP uses a structure to hold

information about each connection TCP keeps an array of TCBs in the form of a table


7/36

CS 640 7

TCP Package (Contd)

State : defining the state of the connection according to thestate transition diagram

Process : defining the process using this connection at this

machine as a client or a server Local IP address : defining the IP address of the local

machine used by this connection

Local port number : defining the local port number used by

this connection Remote IP address

Remote port address

Interface : defining the local interface

Local window : holding information about the window atthe local TCP


8/36

CS 640 8

TCP Package (Contd)

Sending sequence number

Receiving sequence number

Sending ACK number

Time-out values : retransmission time-out,

persistence time-out, keepalive time-out, and so on

Buffer size : defining the size of the buffer at thelocal TCP

Buffer pointer : pointer to buffer where the

receiving data is kept until is read by the

application


9/36

CS 640 9

TCP Package (Contd) Main Module : The main module is invoked by an arrived

TCP segment, a time-out, or a message from an application

program


10/36

CS 640 10

TCP Package (Contd)

Main Module (contd)


11/36

CS 640 11

TCP Package (Contd)

Main Module (contd)


12/36

CS 640 12

TCP Package (Contd) Main Module (contd)


13/36

CS 640 13

TCP D Package (Contd) Main Module (contd)


14/36

CS 640 14

TCP Package (Contd)

Main Module (contd)


15/36

CS 640 15

TCP Package (Contd)

Main Module (contd)


16/36

CS 640 16

TCP Package (Contd)

Main Module (contd)


17/36

CS 640 17

TCP Package (Contd)

Main Module (contd)


18/36

CS 640 18

TCP Package (Contd)

Input processing module

handles all the details needed to process data or

acknowledgment received when TCP is in theESTABLISHED state

sends an ACK if needed, takes care of the window size,

does error checking, and so on

Output processing module handles all the details needed to send out data received

from application program when TCP is in the

ESTABLISHED state

handles retransmission time-outs, persistent time-outs,


19/36

CS 640 19

Segment Format


20/36

CS 640 20

Segment Format

Options (variable)

Data

Checksum

SrcPort DstPort

HdrLen 0 Flags

UrgPtr

AdvertisedWindow

SequenceNum

Acknowledgment

0 4 10 16 31


21/36

CS 640 21

Segment Format (cont) Each connection identified with 4-tuple:

(SrcPort, SrcIPAddr, DsrPort, DstIPAddr)

Sliding window + flow control acknowledgment, SequenceNum, AdvertisedWinow

Flags SYN, FIN, RESET, PUSH, URG, ACK

Checksum is the same as UDP pseudo header + TCP header + data

Sender

Data(SequenceNum)

Acknowledgment +AdvertisedWindow

Receiver


22/36

CS 640 22

Sequence Numbers

32 bit sequence numbers Wrap around supported

TCP breaks byte stream from application into packets

(limited by Max. Segment Size) Each byte in the data stream is considered

Each packet has a sequence number

Initial number selected at connection time Subsequent numbers indicate first data byte number in packet

ACKs indicate next byte expected


23/36

CS 640 23

Comparison


24/36

CS 640 24

Sequence Number Wrap Around

Bandwidth Time Until Wrap Around

T1 (1.5 Mbps) 6.4 hours

Ethernet (10 Mbps) 57 minutesT3 (45 Mbps) 13 minutes

FDDI (100 Mbps) 6 minutes

STS-3 (155 Mbps) 4 minutes

STS-12 (622 Mbps) 55 seconds

STS-24 (1.2 Gbps) 28 seconds

Protect against this by adding a 32-bit timestamp to TCP header


25/36

CS 640 25

Connection Establishment

Active participant(client)

Passive participant(server)

SYN,SequenceNum=x

SYN+AC

K,Sequen

ceNum=y

,

ACK,Acknowledgment=y+1

Acknowled

gment=x+

1


26/36

CS 640 26

Connection TerminationActive participant

(server)Passive participant

(client)

FIN,SequenceNum= x

Acknowledgment=y+1

Acknowled

gment=x+

1

FIN,Sequ

enceNum=

y


27/36

CS 640 27

Finite state machineimplementation


28/36

CS 640 28

State Transition DiagramCLOSED

LISTEN

SYN_RCVD SYN_SENT

ESTABLISHED

CLOSE_WAIT

LAST_ACKCLOSING

TIME_WAIT

FIN_WAIT_2

FIN_WAIT_1

Passive open Close

Send/SYN

SYN/SYN + ACK

SYN + ACK/ACK

SYN/SYN + ACK

ACK

Close/FIN

FIN/ACKClose/FIN

FIN/ACKACK+FIN

/ACK

Timeout after twosegment lifetimes

FIN/ACK

ACK

ACK

ACK

Close/FIN

Close

CLOSED

Active open/SYN


29/36

CS 640 29

Reliability in TCP

Checksum used to detect bit level errors

Sequence numbers used to detect sequencing errors

Duplicates are ignored Reordered packets are reordered (or dropped)

Lost packets are retransmitted

Timeouts used to detect lost packets

Requires RTO calculation

Requires sender to maintain data until it is ACKed


30/36

CS 640 30

Sliding Window Revisited

Sending side LastByteAcked< =LastByteSent

LastByteSent < =

LastByteWritten buffer bytes between

LastByteAckedandLastByteWritten

Sending application

LastByteWritten

TCP

LastByteSentLastByteAcked

Receiving application

LastByteRead

TCP

LastByteRcvdNextByteExpected

Receiving side LastByteReadMaxSenderBuffer

Always send ACK in response to arriving data segment

Persist sending one byte seg. whenAdvertisedWindow= 0


32/36

CS 640 32

Keeping the Pipe Full

16-bitAdvertisedWindow controls amount of pipelining

Assume RTT of 100ms

Add scaling factor extension to header to enable larger windows

Bandwidth Delay x Bandwidth Product

T1 (1.5 Mbps) 18KB

Ethernet (10 Mbps) 122KB

T3 (45 Mbps) 549KB

FDDI (100 Mbps) 1.2MBOC-3 (155 Mbps) 1.8MB

OC-12 (622 Mbps) 7.4MB

OC-24 (1.2 Gbps) 14.8MB


33/36

CS 640 33

Making TCP More Efficient

Delayed acknowledgements

Delay for about 200ms

Try to piggyback ACKs with data Acknowledge every other packet

Many instances in transmission sequence which require

an ACK

Dont forget Nagles algorithm

Can be switched off


34/36

CS 640 34

Karn/Partridge Algorithm for RTO

Two degenerate cases with timeouts and RTT measurements Solution: Do not sample RTT when retransmitting

After each retransmission, set next RTO to be double the valueof the last Exponential backoff is well known control theory method

Loss is most likely caused by congestion so be careful

Sender Receiver

Originaltransmission

ACK

SampleRTT

Retransmission

Sender Receiver

Originaltransmission

ACK

SampleRTT

Retransmission


35/36

CS 640 35

Jacobson/ Karels Algorithm

In late 80s, Internet was suffering from congestion collapse New Calculations for average RTT Jacobson 88

Variance is not considered when setting timeout value If variance is small, we could set RTO = EstRTT

If variance is large, we may need to set RTO > 2 x EstRTT

New algorithm calculates both variance and mean for RTT

Diff = sampleRTT - EstRTT

EstRTT = EstRTT + (dx Diff) Dev = Dev + d ( |Diff| - Dev)

Initially settings forEstRTT and Dev will be given to you where d is a factor between 0 and 1

typical value is 0.125


36/36

CS 640 36

Jacobson/ Karels contd.

TimeOut = xEstRTT + xDev where = 1 and = 4

When variance is small, TimeOut is close to EstRTT

When variance is large Dev dominates the calculation

Another benefit of this mechanism is that it is very efficient toimplement in code (does not require floating point)

Notes algorithm only as good as granularity of clock (500ms on Unix)

accurate timeout mechanism important to congestion control (later)

These issues have been studied and dealt with in new RFCsfor RTO calculation.

TCP RENO uses Jacobson/Karels

Segment Format

Documents

Transcript of Segment Format