Chapter Eight

The Transport Layer

Objectives• Take a look at some other forms of logical

addressing.• Examine the process of encapsulation.• Take a closer look at flow control.• Examine error correction/detection in detail.• See how the Transport layer controls

congestion on the network.

The Transport Layer

• The responsibilities of the Transport Layer are:– Handle end-to-end addressing– Repackage long message into smaller segments for

transmission– At the receiving end, rebuild packets into the original

message– Monitor flow control of data– Handle end-to-end error detection and recovery– Handle congestion control on the network

The Transport Layer

• Why do we need transport layer?– Network layer is focused on

the routers. It provides logical communication between hosts.

– Transport layer runs on end-user devices. It provides logical communication between processes

Household analogy:

12 kids sending letters to 12 kids

processes = kids

app messages = letters in envelopes

hosts = houses

transport protocol = John and Bill

network-layer protocol = postal service

Addressing in the Transport Layer

• Ports and sockets can tell the OS what data is intended for what applications.– Ports are 16-bit numbers that identify applications

or processes.– Sockets are a logical address consisting of a

combination of a port and an IP address.

Ports• Well-known ports– Assigned by Internet Assigned Number Authority

(IANA)– Occupy ports 0 through 1023

• Ephemeral ports– Used by the client software to establish a link

between applications– Generally assigned by the application when it

launches

Some Commonly Used Ports

Port Protocol20 FTP, File Transfer Protocol, data

21 FTP, File Transfer Protocol, control

23 Telnet

25 SMTP, Simple Mail Transfer Protocol

80 HTTP, HyperText Transfer Protocol

109 POP, Post Office Protocol, version 2

110 POP, Post Office Protocol, version 3

666 Doom, ID software

Transport Layer Connections

• Connectionless connections– No virtual connection is created.– Data is basically thrown out onto the wire and the

transmitting workstation assumes it will arrive safely.– The UDP is an example of a connectionless service

• Connection-oriented connections– A virtual connection is created.– For every packet transmitted, either an ACK or a NACK must

be returned.– The TCP is an example of connection-oriented service

UDP•often used for

streaming multimedia apps

▸loss tolerant▸rate sensitive

•other UDP uses▸DNS

source port # dest port #

32 bits

data

UDP segment format

length checksumLength, in

bytes of UDPsegment,including

header

The Real-Time Transport Protocol

•RFC 1889•Basic function of RTP is to Multiplex several

real-time data streams onto a single UDP stream•(a) The position of RTP in the protocol stack.

(b) Packet nesting.

TCP segment structure

source port # dest port #

32 bits

applicationdata

(variable length)

sequence number

acknowledgement number

Receive window

Urg data pnterchecksum

FSRPAUheadlen

notused

Options (variable length)

source port # dest port #

32 bits

applicationdata

(variable length)

sequence number

acknowledgement number

Receive window

Urg data pnterchecksum

FSRPAUheadlen

notused

Options (variable length)

URG: urgent data (generally not used)

ACK: ACK #valid

PSH: push data now(generally not used)

RST, SYN, FIN:Connection’ estab

(setup, teardowncommands)

# bytes rcvr willing

to accept

countingby bytes

of data(not segments!)

Internetchecksum(as in UDP)

TCP segment structure

– A 32-bit sequence number keeps packets in order.– A 32-bit acknowledgement number is used to verify the

packet.– 4-bit Header Length – Indicate the size of the entire TCP

header the receiver– URG – 0 or 1. When set to 1, this bit indicate the urgent

pointer field is valid and should be considered.– ACK – 0 or 1. When set to 1, this bit indicates that

acknowledgement number field is valid and being used

TCP segment structure• A window sized field dictates how many packets will be

sent before waiting for ACKS.– PSH – 0 or 1. When set to 1, this bit tells the receiver to

pass all data received at the point to the receiving application immediately.

– RST – 0 or 1. This bit indicates an error condition has been detected and notify the receiver to reset the connection

TCP segment structure

– SYN – 0 or 1. This bit synchronizes the sequence numbers in order to establish a connection

– 16 bit TCP checksum – ensure that the TCP header has not been modified in transmit

– 16-bit Urgent Pointer – This pointer is added to the sequence number field to yield the sequence number of urgent data.

Flow Control

• Buffer overflow– Memory fills; transmission stops

• Stop and wait– Send a frame and wait for the reply

• Neither methods very useful for busy networks

• Rarely used

socketdoor

TCPsend buffer

TCPreceive buffer

socketdoor

segment

applicationwrites data

applicationreads data

Advanced Flow Control• Static window– A fixed number of frames are transmitted.– The transmitting station waits for the replies.– No adjustments in transmission speed can be made.

• Sliding window– It starts with a higher number of frames.– As failures occur, the number of frames transmitted drops.– If a frame is dropped, that frame and all frames following

it will get retransmitted.

MORE Flow Control• Selectively repeat– A number of

frames are transmitted.

– If a failure occurs, only the bad packets need to be transmitted.

MORE Flow Control• Go Back N– It is similar to sliding

window except that a single ACK is sent for all frames in a window.

– If a failure occurs, the protocol counts back the correct number of frames and retransmits all.

Error Control in Transport

• The error correction in Data Link was bit-level error correction.– If user data was corrupted, the error was detected

and, if possible, fixed.

• Transport layer error correction is end-to-end.– There may have error during encapsulation– If a packet is lost or corrupted, the error is fixed.

Error Control in Transport

• Packet level errors can include packet loss, packet corruption, and packet duplication. The network uses– three-way handshake– sequence number – time-out for each packet

TCP Connection Establishment

• Recall: TCP sender, receiver establish connection before exchanging data segments

• initialize TCP variables:▸seq. #s▸buffers, flow control info

(e.g. RcvWindow)

Three way handshake:

Step 1: client host sends TCP SYN segment to server

specifies initial seq#

no data

Step 2: server host receives SYN, replies with SYNACK

segment

server allocates buffers

specifies server initial seq. #

Step 3: client receives SYNACK, replies with ACK segment,

which may contain data

TCP Connection Establishment

• (a) TCP connection establishment in the normal case.

• (b) Call collision. – only one connection is established

TCP Connection Close• Closing a connection:• Step 1: client end system

sends TCP FIN control segment to server

• Step 2: server receives FIN, replies with ACK. Closes connection, sends FIN.

• Step 3: client receives FIN, replies with ACK.

• Step 4: server, receives ACK. Connection closed.

client

FIN

server

ACK

ACK

FIN

close

close

closed

timed

wait

client

FIN

server

ACK

ACK

FIN

close

close

closed

timed

wait

Principles of Congestion Control• Congestion:–informally: too many sources sending too much data

too fast for network to handle –different from flow control!

• manifestations:▸lost packets (buffer overflow at routers)▸long delays (queueing in router buffers)

• a top-10 problem!

Congestion Control

• (a) A fast network feeding a low capacity receiver.

• (b) A slow network feeding a high-capacity receiver.

Congestion Control

• No single device can control overall network congestion.

• Therefore, Transport does what it can to make sure THIS DEVICE does not contribute to congestion.

• Connections requiring excessive retransmission of data are dropped.

Approaches towards congestion control

• End-end congestion control:▸no explicit feedback from network ▸congestion inferred from end-system observed loss,

delay▸approach taken by TCP

• Network-assisted congestion control:▸routers provide feedback to end systems▸single bit indicating congestion (SNA, DECbit, TCP/IP

ECN, ATM)▸explicit rate sender should send at

Two broad approaches towards congestion control:

TCP Congestion Control• Slow start (Jacobson 1998)– Start with the maximum segment size– If this is acknowledge then double the window size– Send two maximum segemnt size– Repeat

• When the CongWin = threshold, increase linearly.– Threshold = 1/2 of CongWin value before timeout.

– Initially 64KB in addition to receiver flow control and congestion control window

• When timeout occur – reduce threshold to half of the congestion window– Congestion window is reset to 1 segment

Slow Start• When connection

begins, increase rate exponentially until threshold:

• double CongWin every RTT

• done by incrementing CongWin for every ACK received

• Summary: initial rate is slow but ramps up exponentially fast

Host A

one segment

RTT

Host B

timetime

two segments

four segments

TCP Congestion Control (2)

• An example of the Internet congestion algorithm.

Summary: TCP Congestion Control• When CongWin is below Threshold, sender in

slow-start phase, window grows exponentially.• When CongWin is above Threshold, sender is in

congestion-avoidance phase, window grows linearly.

• When timeout occurs, Threshold set to CongWin/2 and CongWin is set to 1 MSS.

TCP Round Trip Time and Timeout•Q: how to set TCP timeout value?

▸longer than RTT▸but RTT varies▸too short: premature timeout–unnecessary retransmissions▸too long: slow reaction to segment loss

▸Q: how to estimate RTT?▸SampleRTT: measured time from segment transmission until

ACK receipt▸ignore retransmissions▸SampleRTT will vary, want estimated RTT smoother ▸average several recent measurements, not just current

SampleRTT

TCP Timer Management• Variable Retransmission based on RTT• Timeout based on Round Trip Time (RTT)– RTT = αRTT + (1-α)M– M is the time the ack received– α is smoothing factor typically 7/8

• A better estimate– Timeout = RTT + 4xD– D = αD + (1-α)|RTT-M|

• Karn’s Algorithm– Do not use RTT if retransmission happens– Time out is doubled on every failure

Example RTT estimation:RTT: gaia.cs.umass.edu to fantasia.eurecom.fr

100

150

200

250

300

350

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

time (seconnds)

RTT (m

illiseconds)

SampleRTT Estimated RTT