Computer Networking: A Top Down

Approach
Seventh Edition

Chapter 3
Transport Layer

Slides in this presentation contain

hyperlinks. JAWS users should be able to
get a list of links by using INSERT+F7

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Transport Layer
our goals:
• understand principles behind transport layer services:
– multiplexing, demultiplexing
– reliable data transfer
– flow control
– congestion control
• learn about Internet transport layer protocols:
– UDP: connectionless transport
– TCP: connection-oriented reliable transport
– TCP congestion control
Learning Objectives (1 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

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

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
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 in Ann’s house sending letters to 12 kids in Bill’s house:

• hosts = houses

• processes = kids

• app messages = letters in envelopes

• transport protocol = Ann and Bill who demux to in-house siblings

• network-layer protocol = postal service

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
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:
– delay guarantees
– bandwidth guarantees

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (2 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Multiplexing/Demultiplexing

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
How Demultiplexing Works
• host receives IP datagrams
– each datagram has source IP
address, destination IP
address
– each datagram carries one
transport-layer segment
– each segment has source,
destination port number
• host uses IP addresses & port
numbers to direct segment to
appropriate socket

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Connectionless Demultiplexing
• recall: created socket has host- • recall: when creating datagram to
local port #: send into UDP socket, must
specify
– destination IP address
– destination port #

• when host receives UDP IP datagrams with same dest.

segment: port #, but different source IP
– checks destination port # in addresses and/or source port
segment numbers will be directed to
– directs UDP segment to same socket at dest
socket with that port #

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Connectionless Demux: Example

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Connection-Oriented Demux
• TCP socket identified by 4-tuple:
– source IP address
– source port number
– dest IP address
– dest port number
• demux: receiver uses all four values to direct segment to appropriate
socket
• server host may support many simultaneous TCP sockets:
– each socket identified by its own 4-tuple
• web servers have different sockets for each connecting client
– non-persistent HTTP will have different socket for each request

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Connection-Oriented Demux: Example (1 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Connection-Oriented Demux: Example (2 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (3 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
UDP: User Datagram Protocol [RFC 768]
• “no frills,” “bare bones” Internet • UDP use:
transport protocol – streaming multimedia apps
• “best effort” service, UDP (loss tolerant, rate
segments may be: sensitive)
– lost – DNS
– delivered out-of-order to – SNMP
app • reliable transfer over UDP:
• connectionless: – add reliability at
– no handshaking between application layer
UDP sender, receiver – application-specific error
– each UDP segment recovery!
handled independently of
others
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
UDP: Segment Header (1 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
UDP: Segment Header (2 of 2)

why is there a UDP?

• no connection establishment (which can add delay)
• simple: no connection state at sender, receiver
• small header size
• no congestion control: UDP can blast away as fast as
desired

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
UDP Checksum
Goal: detect “errors” (example, flipped bits) in transmitted segment

sender: receiver:
• treat segment contents, • compute checksum of received
including header fields, as segment
sequence of 16-bit integers
• check if computed checksum
• checksum: addition (one’s equals checksum field value:
complement sum) of – NO - error detected
segment contents
– YES - no error detected.
• sender puts checksum But maybe errors
value into UDP checksum nonetheless? More later
field ….

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Internet Checksum: Example

example: add two 16-bit integers

Note: when adding numbers, a carryout from the most

significant bit needs to be added to the result
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (4 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Principles of Reliable Data Transfer
• important in application, transport, link layers
– top-10 list of important networking topics!

• characteristics of unreliable channel will determine complexity of reliable

data transfer protocol (rdt)
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Reliable Data Transfer: Getting Started (1 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Reliable Data Transfer: Getting Started (2 of 2)
we’ll:
• incrementally develop sender, receiver sides of reliable data transfer
protocol (rdt)
• consider only unidirectional data transfer
– but control info will flow on both directions!
• use finite state machines (FSM) to specify sender, receiver

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt1.0: Reliable Transfer over a Reliable
Channel

• underlying channel perfectly reliable

– no bit errors
– no loss of packets
• separate FSMs for sender, receiver:
– sender sends data into underlying channel
– receiver reads data from underlying channel

sender receiver

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.0: Channel with Bit Errors (1 of 2)

• underlying channel may flip bits in packet

– checksum to detect bit errors
• the question: how to recover from errors:

How do humans recover from “errors”

during conversation?

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.0: Channel with Bit Errors (2 of 2)
• underlying channel may flip bits in packet
– checksum to detect bit errors
• the question: how to recover from errors:
– acknowledgements (ACKs): receiver explicitly tells sender that
pkt received OK
– negative acknowledgements (NAKs): receiver explicitly tells
sender that pkt had errors
– sender retransmits pkt on receipt of NAK
• new mechanisms in rdt2.0 (beyond rdt1.0):
– error detection
– feedback: control msgs (ACK, NAK) from receiver to sender

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.0: FSM Specification

sender receiver

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.0: Operation with No Errors

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.0: Error Scenario

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.0 Has a Fatal Flaw!
what happens if ACK / NAK handling duplicates:
corrupted?
• sender retransmits current
• sender doesn’t know what pkt if ACK / NAK corrupted
happened at receiver!
• sender adds sequence
• Can’t just retransmit: number to each pkt
possible duplicate
• receiver discards (doesn’t
deliver up) duplicate pkt

stop and wait

sender sends one packet, then waits for
receiver response
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.1: Sender, Handles Garbled ACK/NAKs

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.1: Receiver, Handles Garbled ACK/NAKs

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.1: Discussion
sender: receiver:
• seq # added to pkt • must check if received
packet is duplicate
• two Sequence #’s (0,1) will
suffice. Why? – state indicates whether
0 or 1 is expected pkt
• must check if received ACK/ seq #
NAK corrupted
• note: receiver can not
• twice as many states know if its last ACK/NAK
– state must “remember” received OK at sender
whether “expected” pkt
should have seq # of 0 or
1
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.2: A NAK-free Protocol
• same functionality as rdt2.1, using ACKs only
• instead of NAK, receiver sends ACK for last pkt
received OK
– receiver must explicitly include seq # of pkt being
ACKed
• duplicate ACK at sender results in same action as NAK:
retransmit current pkt

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt2.2: Sender, Receiver Fragments

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt3.0: Channels with Errors and Loss
new assumption: underlying channel can also lose packets (data,
ACKs)
– checksum, Sequence #, ACKs, retransmissions will be of
help … but not enough
approach: sender waits “reasonable” amount of time for ACK
• retransmits if no ACK received in this time
• if pkt (or ACK) just delayed (not lost):
– retransmission will be duplicate, but Sequence #’s already
handles this
– receiver must specify seq # of pkt being ACKed
• requires countdown timer

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt3.0 Sender

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt3.0 in Action (1 of 2)
(a) no loss (b) packet loss

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt3.0 in Action (2 of 2)
(c) ACK loss (d) premature timeout/ delayed ACK

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Performance of rdt3.0 (1 of 2)

• rdt3.0 is correct, but performance stinks

• example: 1 Gbps link, 15 ms prop. delay, 8000 bit packet:

L 8000 bits / packet

dtrans   9
 8 microseconds
R 10 bits / sec

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Performance of rdt3.0 (2 of 2)

– Usender: utilization – fraction of time sender busy

sending

L
R .008
Usender    0.00027
L 30.008
RTT +
R

– if RTT = 30 msec, 1KB pkt every 30 msec: 33kB/sec

thruput over 1 Gbps link
• network protocol limits use of physical resources!

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
rdt3.0: Stop-and-wait Operation

L
R .008
Usender    0.00027
L 30.008
RTT +
R
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Pipelined 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

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Pipelining: Increased Utilization

3-packet pipelining
increases utilization by
a factor of 3!

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Pipelined Protocols: Overview
Go-back-N: Selective Repeat:
• sender can have up to N • sender can have up to N
unacked packets in pipeline unack’ed packets in pipeline
• receiver only sends • rcvr sends individual ack for
cumulative ack each packet
– doesn’t ack packet if • sender maintains timer for each
there’s a gap unacked packet
• sender has timer for oldest – when timer expires,
unacked packet retransmit only that unacked
– when timer expires, packet
retransmit all unacked
packets

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
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 oldest in-flight pkt
• timeout(n): retransmit packet n and all higher seq # pkts in window

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
GBN: Sender Extended FSM

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
GBN: Receiver Extended FSM (1 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
GBN: Receiver Extended FSM (2 of 2)

• ACK-only: always send ACK for correctly-received pkt

with highest in-order seq #
– may generate duplicate ACKs
– need only remember expectedseqnum
• out-of-order pkt:
– discard (don’t buffer): no receiver buffering!
– re-ACK pkt with highest in-order seq #

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
GBN in Action

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Selective Repeat (1 of 3)

• 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
– limits seq #s of sent, unACKed pkts

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Selective Repeat: Sender, Receiver
Windows

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Selective Repeat (2 of 3)
sender
data from above:
• if next available seq # in window, send pkt
timeout(n):
• resend pkt n, restart timer
ACK(n) in [sendbase,sendbase+N]:
• mark pkt n as received
• if n smallest unACKed pkt, advance window base to next
unACKed seq #

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Selective Repeat (3 of 3)
receiver
pkt n in [rcvbase, rcvbase+N-1]
• send ACK(n)
• out-of-order: buffer
• in-order: deliver (also deliver buffered, in-order pkts), advance
window to next not-yet-received pkt
pkt n in [rcvbase-N,rcvbase-1]
• ACK(n)
otherwise:
• ignore
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Selective Repeat in Action

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Selective Repeat: Dilemma (1 of 2)
• example: (a) no problem

• seq #’s: 0, 1, 2, 3

• window size=3

• receiver sees no difference

in two scenarios!

• duplicate data accepted as

new in (b)

receiver can’t see sender side. receiver behavior identical in both cases!
something’s (very) wrong!

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Selective Repeat: Dilemma (2 of 2)
Q: what relationship between seq # size and window size to avoid problem in
(b)?
(b) oops!

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (5 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP: Overview RFCs: 793, 1122, 1323, 2018,
2581
• point-to-point: – MSS: maximum segment
– one sender, one receiver size

• reliable, in-order byte steam: • connection-oriented:

– no “message boundaries” – handshaking (exchange
of control msgs) inits
• pipelined: sender, receiver state
– TCP congestion and flow before data exchange
control set window size
• flow controlled:
• full duplex data: – sender will not
– bi-directional data flow in overwhelm receiver
same connection

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Segment Structure

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Sequence Numbers, ACKs (1 of 2)
sequence numbers:
– byte stream “number” of
first byte in segment’s
data
acknowledgements:
– seq # of next byte
expected from other side
– cumulative ACK
Q: how receiver handles out-of-
order segments
– A: TCP spec doesn’t say,
- up to implementor

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Sequence Numbers, ACKs (2 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Round Trip Time, Timeout (1 of 3)
Q: how to set TCP timeout Q: how to estimate RTT?
value?
• SampleRTT: measured time
• longer than RTT from segment transmission
– but RTT varies until ACK receipt
– ignore retransmissions
• too short: premature
timeout, unnecessary • SampleRTT will vary, want
retransmissions estimated RTT “smoother”
• too long: slow reaction to – average several recent
segment loss measurements, not just
current SampleRTT

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Round Trip Time, Timeout (2 of 3)
EstimatedRTT = (1-α)*EstimatedRTT + α *SampleRTT

• exponential weighted moving average

• influence of past sample decreases exponentially fast

• typical value:  = 0.125

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Round Trip Time, Timeout (3 of 3)
• timeout interval: EstimatedRTT plus “safety margin”
– large variation in EstimatedRTT  larger safety margin.
• estimate SampleRTT deviation from EstimatedRTT:
DevRTT = (1-β)* DevRTT +
β* SampleRTT - EstimatedRTT
(typically,β = 0.25)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (6 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Reliable Data Transfer
• TCP creates rdt service on top of IP’s unreliable service
– pipelined segments
– cumulative acks
– single retransmission timer
• retransmissions triggered by:
– timeout events
– duplicate acks
• let’s initially consider simplified TCP sender:
– ignore duplicate acks
– ignore flow control, congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Sender Events:
data rcvd from app: timeout:

• create segment with seq # • retransmit segment that caused

timeout
• seq # is byte-stream number
of first data byte in segment • restart timer
ack rcvd:
• start timer if not already
running • if ack acknowledges previously
– think of timer as for unacked segments
oldest unacked segment – update what is known to be
ACKed
– expiration interval:
TimeOutInterval – start timer if there are still
unacked segments

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Sender (Simplified)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP: Retransmission Scenarios (1 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP: Retransmission Scenarios (2 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP ACK Generation [RFC 1122, RFC 2581]

event at receiver TCP receiver action

arrival of in-order segment with delayed ACK. Wait up to 500ms
expected seq #. All data up to for next segment. If no next segment,
expected seq # already ACKed send ACK
arrival of in-order segment with immediately send single cumulative
expected seq #. One other ACK, ACKing both in-order segments
segment has ACK pending
arrival of out-of-order segment immediately send duplicate ACK,
higher-than-expect Sequence # . indicating Sequence # of next
Gap detected expected byte
arrival of segment that immediate send ACK, provided that
partially or completely fills gap segment starts at lower end of gap

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Fast Retransmit (1 of 2)
• time-out period often relatively long:
– long delay before resending lost packet
• detect lost segments via duplicate ACKs.
– sender often sends many segments back-to-back
– if segment is lost, there will likely be many duplicate ACKs.

TCP fast retransmit

if sender receives 3 ACKs for same data (“triple duplicate
ACKs”), resend unacked segment with smallest seq #
– likely that unacked segment lost, so don’t wait for
timeout

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Fast Retransmit (2 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (7 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Flow Control (1 of 2)

flow control
receiver controls
sender, so sender
won’t overflow
receiver’s buffer by
transmitting too
much, too fast

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Flow Control (2 of 2)
• receiver “advertises” free buffer space
by including rwnd value in TCP
header of receiver-to-sender
segments
– RcvBuffer size set via socket
options (typical default is 4096
bytes)
– many operating systems
autoadjust RcvBuffer

• sender limits amount of unacked (“in-

flight”) data to receiver’s rwnd value

• guarantees receive buffer will not

overflow

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (8 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Connection Management
before exchanging data, sender/receiver “handshake”:

• agree to establish connection (each knowing the other willing to establish

connection)

• agree on connection parameters

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Agreeing to Establish a Connection (1 of 2)

2-way handshake:
Q: will 2-way handshake
always work in network?
• variable delays
• retransmitted messages
(example. req_conn(x)) due
to message loss
• message reordering
• can’t “see” other side

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Agreeing to Establish a Connection (2 of 2)

2-way handshake failure scenarios:

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP3-Way Handshake

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP3-Way Handshake: FSM

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP: Closing a Connection (1 of 2)

• client, server each close their side of connection

– send TCP segment with FIN bit = 1
• respond to received FIN with ACK
– on receiving FIN, ACK can be combined with own FIN
• simultaneous FIN exchanges can be handled

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP: Closing a Connection (2 of 2)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (9 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
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!

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 1 (1 of 2)
• two senders, two receivers

• one router, infinite buffers

• output link capacity: R

• no retransmission

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 1 (2 of 2)
R
• maximum per-connection throughput:
2

• large delays as arrival rate, λin , approaches capacity

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 2 (1 of 6)

• one router, finite buffers

• sender retransmission of timed-out packet
– application-layer input = application-layer output: λin  λout
– transport-layer input includes retransmissions : λ'in  λin

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 2 (2 of 6)

idealization: perfect knowledge

• sender sends only when router
buffers available

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 2 (3 of 6)

• Idealization: known loss packets can be lost, dropped

at router due to full buffers
• sender only resends if packet known to be lost

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 2 (4 of 6)

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 2 (5 of 6)

Realistic: duplicates
• packets can be lost, dropped at router due to full buffers
• sender times out prematurely, sending two copies, both
of which are delivered

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 2 (6 of 6)
“costs” of congestion:

• more work (retrans) for given “goodput”

• unneeded retransmissions: link carries multiple copies of pkt

– decreasing goodput

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 3 (1 of 2)
• four senders Q: what happens as λin and λin ' increase ?

• multihop paths A: as red λin' increases, all arriving blue pkts

at upper queue are dropped, blue throughput → 0
• timeout/retransmit

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Causes/Costs of Congestion: Scenario 3 (2 of 2)

another “cost” of congestion:

• when packet dropped, any “upstream transmission
capacity used for that packet was wasted!

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Learning Objectives (10 of 10)
3.1 transport-layer services

3.2 multiplexing and demultiplexing

3.3 connectionless transport: UDP

3.4 principles of reliable data transfer

3.5 connection-oriented transport: TCP

– segment structure
– reliable data transfer
– flow control
– connection management

3.6 principles of congestion control

3.7 TCP congestion control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Congestion Control: Additive
Increase Multiplicative Decrease
• approach: sender increases transmission rate (window size),
probing for usable bandwidth, until loss occurs
– additive increase: increase cwnd by 1 MSS every RTT until
loss detected
– multiplicative decrease: cut cwnd in half after loss

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Congestion Control: Details

• cwnd is dynamic, function

of perceived network
congestion
TCP sending rate:
• roughly: send cwnd
bytes, wait RTT for ACKS,
then send more bytes

• sender limits transmission:

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Slow Start
• when connection begins,
increase rate exponentially
until first loss event:
– initially cwnd = 1 MSS
– double cwnd every RTT
– done by incrementing
cwnd for every ACK
received
• summary: initial rate is slow
but ramps up exponentially
fast

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP: Detecting, Reacting to Loss

• loss indicated by timeout:

– cwnd set to 1 MSS;
– window then grows exponentially (as in slow start) to
threshold, then grows linearly
• loss indicated by 3 duplicate ACKs: TCP RENO
– dup ACKs indicate network capable of delivering
some segments
– cwnd is cut in half window then grows linearly
• TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate
acks

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP: Switching from Slow Start to CA

Q: when should the

exponential increase
switch to linear?
1
A: when cwnd gets to
2
of its value before timeout.

Implementation:
• variable ssthresh
1
• on loss event, ssthresh is set to of cwnd just before loss event
2
* Check out the online interactive exercises for more examples:
http://gaia.cs.umass.edu/kurose_ross/interactive/
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Summary: TCP Congestion Control

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Throughput
• average TCP thruput as function of window size, RTT?
– ignore slow start, assume always data to send

• W: window size (measured in bytes) where loss occurs

– average window size (# in-flight bytes) is
3 3
– average thruput is W per RTT W
4 4

3 W
avg TCP thruput = bytes / sec
4 RTT

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Futures: TCP over “Long, Fat Pipes”
• example: 1500 byte segments, 100ms RTT, want 10 Gbps
throughput
• requires W = 83,333 in-flight segments
• throughput in terms of segment loss probability, L [Mathis
1997]:
1.22  MSS
TCP throughput 
RTT L

→ to achieve 10 Gbps throughput, need a loss rate of

L  2  10 10 – a very small loss rate!
• new versions of TCP for high-speed
Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
TCP Fairness

fairness goal: if K TCP sessions share same bottleneck

link of bandwidth R, each should have average rate of R
K

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Why is TCP Fair?
two competing sessions:
• additive increase gives slope of 1, as throughout increases
• multiplicative decrease decreases throughput proportionally

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Fairness (More)
Fairness and UDP Fairness, parallel TCP connections

• multimedia apps often do • application can open multiple parallel

not use TCP connections between two hosts
– do not want rate • web browsers do this
throttled by
congestion control • example, link of rate R with 9 existing
connections:
• instead use UDP: R
– new app asks for 1 TCP, gets rate
– send audio/video at 10
constant rate, – new app asks for 11 TCPs, gets R
tolerate packet loss 2

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Explicit Congestion Notification (ECN)
network-assisted congestion control:

• two bits in IP header (ToS field) marked by network router to indicate

congestion

• congestion indication carried to receiving host

• receiver (seeing congestion indication in IP datagram) ) sets ECE bit on

receiver-to-sender ACK segment to notify sender of congestion

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Chapter Summary
• principles behind transport next:
layer services:
• leaving the network “edge”
– multiplexing, (application, transport
demultiplexing layers)
– reliable data transfer
• into the network “core”
– flow control
– congestion control • two network layer chapters:
– data plane
• instantiation, implementation
in the Internet – control plane
– UDP
– TCP

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved
Copyright

Copyright © 2017, 2013, 2010 Pearson Education, Inc. All Rights Reserved