Chapter 3: Transport Layer

Applications
application
… built on ...
Reliable (or unreliable) transport transport
… built on ...
network
Best-effort global packet delivery
… built on ... link
Best-effort local packet delivery
… built on ... physical

Physical transfer of bits

The source PowerPoint slides are public available, provided by
Authors (JFK/KWR). They are revised for CS536@Purdue.
Roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
§ Principles of congestion control
§ TCP congestion control
§ Evolution of transport-layer
functionality
Transport Layer: 3-2
Transport services and protocols
application
transport

§ provide logical communication mobile network

network
data link
physical
between application processes national or global ISP

running on different hosts

lo g
ica
l en
§ transport protocols actions in end

d-e
systems:

nd
local or

tra
• sender: breaks application messages regional ISP

nsp
into segments, passes to network layer

ort
home network content
• receiver: reassembles segments into provider
network
messages, passes to application layer datacenter
applicationnetwork
transport
network

§ two transport protocols available to data link

physical

Internet applications enterprise

network
• TCP, UDP
Transport Layer: 3-3
Transport Layer Actions

Sender:
application § is passed an application- application
app. msg
layer message
transport § determines segment TThhtransport
app. msg
header fields values
network (IP) § creates segment network (IP)

link
§ passes segment to IP link

physical physical

Transport Layer: 3-4

Transport Layer Actions

Receiver:
application § receives segment from IP application
§ checks header values
app. msg
transport § extracts application-layer transport
message
network (IP) network (IP)
§ demultiplexes message up
link to application via socket link

physical physical
Th app. msg

Transport Layer: 3-5

Transport vs. network layer services and protocols
household analogy:
§network layer: logical
communication between 12 kids in Ann’s house sending
letters to 12 kids in Bill’s
hosts house:
§transport layer: logical § hosts = houses
communication between § processes = kids
§ app messages = letters in
processes envelopes
• relies on, enhances, network § transport protocol = Ann and Bill
layer services who demux to in-house siblings
§ network-layer protocol = postal
service

Transport Layer: 3-7

Two principal Internet transport
protocols application
transport
§ TCP: Transmission Control Protocol mobile network
network
data link
physical
• reliable, in-order delivery national or global ISP

lo g
• congestion control

ica
• flow control

l en
d-e
• connection setup

nd
local or
§ UDP: User Datagram Protocol

tra
regional ISP

nsp
• unreliable, unordered delivery

ort
home network content
provider
• no-frills extension of “best-effort” IP network datacenter
applicationnetwork

§ services not available: transport

network
data link

• delay guarantees physical

• bandwidth guarantees enterprise

network

Transport Layer: 3-8

Chapter 3: roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
§ Principles of congestion control
§ TCP congestion control
§ Evolution of transport-layer
functionality
Transport Layer: 3-9
Multiplexing/demultiplexing
multiplexing at sender: demultiplexing at receiver:
handle data from multiple use header info to deliver
sockets, add transport header received segments to correct
(later used for demultiplexing) socket

application

application P1 P2 application socket

P3 transport P4
process
transport network transport
network link network
link physical link
physical physical

Transport Layer: 3-15

How demultiplexing works
§ host receives IP datagrams 32 bits
• each datagram has source IP source port # dest port #
address, destination IP address
• each datagram carries one other header fields
transport-layer segment
• each segment has source, application
destination port number data
§ host uses IP addresses & port (payload)

numbers to direct segment to

appropriate socket TCP/UDP segment format

Transport Layer: 3-16

Connectionless demultiplexing
Recall: when receiving host receives
§ when creating socket, must UDP segment:
• checks destination port # in
specify host-local port #: segment
DatagramSocket mySocket1 • directs UDP segment to
= new DatagramSocket(12534);
socket with that port #
§ when creating datagram to
send into UDP socket, must
specify IP/UDP datagrams with same dest.
port #, but different source IP
• destination IP address addresses and/or source port
• destination port # numbers will be directed to same
socket at receiving host
Transport Layer: 3-17
 Connectionless demultiplexing: an example
DatagramSocket
serverSocket = new
DatagramSocket
DatagramSocket mySocket2 = DatagramSocket mySocket1 =
new DatagramSocket (6428); new DatagramSocket (5775);
(9157); application
application application
P1
P3 P4
transport
transport transport
network
network link network
link physical link
physical physical

source port: 6428 source port: 6428

dest port: 9157 dest port: 5775

source port: 9157 source port: 5775

dest port: 6428 dest port: 6428
Transport Layer: 3-18
Connection-oriented demultiplexing
§ TCP socket identified by § server may support many
4-tuple: simultaneous TCP sockets:
• source IP address • each socket identified by its
• source port number own 4-tuple
• dest IP address • each socket associated with
• dest port number a different connecting client
§ demux: receiver uses all
four values (4-tuple) to
direct segment to
appropriate socket
Transport Layer: 3-19
Connection-oriented demultiplexing: example
application
application P4 P5 P6 application
P1 P2 P3
transport
transport transport
network
network link network
link physical link
physical server: IP physical
address B

host: IP source IP,port: B,80 host: IP

address A dest IP,port: A,9157 source IP,port: C,5775 address C
dest IP,port: B,80
source IP,port: A,9157
dest IP, port: B,80
source IP,port: C,9157
dest IP,port: B,80
Three segments, all destined to IP address: B,
dest port: 80 are demultiplexed to different sockets
Transport Layer: 3-20
Summary
§ Multiplexing, demultiplexing: based on segment, datagram
header field values
§ UDP: demultiplexing using destination port number (only)
§ TCP: demultiplexing using 4-tuple: source and destination IP
addresses, and port numbers

Transport Layer: 3-21

Chapter 3: roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
§ Principles of congestion control
§ TCP congestion control
§ Evolution of transport-layer
functionality
Transport Layer: 3-22
UDP: User Datagram Protocol
Why is there a UDP?
§ “no frills,” “bare bones”
Internet transport protocol § no connection
establishment (which can
§ “best effort” service, UDP add RTT delay)
segments may be: § simple: no connection state
• lost at sender, receiver
• delivered out-of-order to app § small header size
§ connectionless: § no congestion control
• no handshaking between UDP § UDP can blast away as fast as
desired!
sender, receiver
§ can function in the face of
• each UDP segment handled congestion
independently of others
Transport Layer: 3-23
UDP: User Datagram Protocol
§ UDP use:
§ streaming multimedia apps (loss tolerant, rate sensitive)
§ DNS
§ SNMP
§ HTTP/3
§ if reliable transfer needed over UDP (e.g., HTTP/3):
§ add needed reliability at application layer
§ add congestion control at application layer

Transport Layer: 3-24

UDP: Transport Layer Actions

SNMP client SNMP server

application application

transport transport
(UDP) (UDP)

network (IP) network (IP)

link link

physical physical

Transport Layer: 3-26

UDP: Transport Layer Actions

SNMP client SNMP server

UDP sender actions:
application § got an application-layer application
SNMP msg
message
transport transport
§ determines UDP segment UDP
UDPhh SNMP msg
(UDP) header fields values (UDP)

network (IP) § creates UDP segment network (IP)

link
§ passes segment to IP link

physical physical

Transport Layer: 3-27

UDP: Transport Layer Actions

SNMP client SNMP server

UDP receiver actions:
application § receives segment from IP application
§ checks UDP checksum
transport transport
SNMP msg header value
(UDP) (UDP)
§ extracts application-layer
network
UDPh SNMP(IP)
msg message network (IP)
§ demultiplexes message up
link link
to application via socket
physical physical

Transport Layer: 3-28

UDP segment header
32 bits
source port # dest port #
length checksum

application length, in bytes of

data UDP segment,
(payload) including header

data to/from
UDP segment format application layer

Transport Layer: 3-29

UDP checksum
Goal: detect errors (i.e., flipped bits) in transmitted segment
1st number 2nd number sum

Transmitted: 5 6 11

Received: 4 6 11

receiver-computed
checksum
= sender-computed
checksum (as received)

Transport Layer: 3-30

UDP checksum
Goal: detect errors (i.e., flipped bits) in transmitted segment
sender: receiver:
§ treat contents of UDP § compute checksum of received
segment (including UDP header segment
fields and IP addresses) as
sequence of 16-bit integers § check if computed checksum equals
§ checksum: addition (one’s checksum field value:
complement sum) of segment • Not equal - error detected
content • Equal - no error detected. But maybe
§ checksum value put into errors nonetheless? More later ….
UDP checksum field
Transport Layer: 3-31
Internet checksum: an example
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 1 0 1 1 1 0 1 1 1 0 1 1 1 1 0 0
checksum 0 1 0 0 0 1 0 0 0 1 0 0 0 0 1 1

Note: when adding numbers, a carryout from the most significant bit needs to be
added to the result

* Check out the online interactive exercises for more examples: http://gaia.cs.umass.edu/kurose_ross/interactive/
Transport Layer: 3-32
Internet checksum: weak protection!
example: add two 16-bit integers
0 1
1 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 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 Even though
numbers have
sum 1 0 1 1 1 0 1 1 1 0 1 1 1 1 0 0 changed (bit
flips), no change
checksum 0 1 0 0 0 1 0 0 0 1 0 0 0 0 1 1 in checksum!

Transport Layer: 3-33

Summary: UDP
§ “no frills” protocol:
• segments may be lost, delivered out of order
• best effort service: “send and hope for the best”
§ UDP has its plusses:
• no setup/handshaking needed (no RTT incurred)
• can function when network service is compromised
• helps with reliability (checksum)
§ build additional functionality on top of UDP in application layer
(e.g., HTTP/3)
Transport Layer: 3-34
Chapter 3: roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
§ Principles of congestion control
§ TCP congestion control
§ Evolution of transport-layer
functionality
Transport Layer: 3-35
Principles of reliable data
transfer
vimportant @ application, transport, link layers
§ Reliable transport of packets
§ A single sender and a single receiver
§ Packet delivery imperfect
§ With bit errors, dropping packets, out-of-order delivery, duplicate copies, long
delay, ….

logical end-end reliable transport

sender receiver

X packets received
packets in queue/buffer errors loss
Packet delivery misbehaviors Transport Layer: 3-36
 Principles of reliable data transfer

sending receiving
process process
application data data
transport
reliable channel

reliable service abstraction

Transport Layer: 3-37

 Principles of reliable data transfer

sending receiving sending receiving

process process process process
application data data application data data
transport transport
reliable channel
sender-side of receiver-side
reliable service abstraction reliable data of reliable data
transfer protocol transfer protocol

transport
network
unreliable channel

reliable service implementation

Transport Layer: 3-38

Principles of reliable data transfer

sending receiving
process process
application data data
transport

sender-side of receiver-side
Complexity of reliable data reliable data
transfer protocol
of reliable data
transfer protocol
transfer protocol will depend
(strongly) on characteristics of transport
network
unreliable channel (lose, unreliable channel
corrupt, reorder data?)
reliable service implementation

Transport Layer: 3-39

Principles of reliable data transfer

sending receiving
process process
application data data
transport

sender-side of receiver-side
Sender, receiver do not know reliable data
transfer protocol
of reliable data
transfer protocol
the “state” of each other, e.g.,
was a message received? transport
network
§ unless communicated via a unreliable channel

message
reliable service implementation

Transport Layer: 3-40

 Reliable data transfer protocol (rdt): interfaces
rdt_send(): called from above, deliver_data(): called by rdt
(e.g., by app.). Passed data to to deliver data to upper layer
deliver to receiver upper layer
sending receiving
process process
rdt_send() data data
deliver_data()

sender-side data receiver-side

implementation of implementation of
rdt reliable data packet rdt reliable data
transfer protocol transfer protocol
udt_send() Header data Header data rdt_rcv()

unreliable channel
udt_send(): called by rdt rdt_rcv(): called when packet
to transfer packet over arrives on receiver side of
Bi-directional communication over
unreliable channel to receiver unreliable channel channel
Transport Layer: 3-41
Reliable data transfer: getting started
We will:
§ incrementally develop sender, receiver sides of reliable data transfer
protocol (rdt)
§ consider only unidirectional data transfer
• but control info will flow in both directions!
§ use finite state machines (FSM) to specify sender, receiver
event causing state transition
actions taken on state transition
state: when in this “state”
next state uniquely state state
determined by next 1 event
event 2
actions

Transport Layer: 3-42

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

Wait for rdt_send(data) rdt_rcv(packet)

Wait for
sender call from packet = make_pkt(data) receiver call from extract (packet,data)
above udt_send(packet) below deliver_data(data)

Transport Layer: 3-43

 “Stop and Wait” Scenario
§Simple setting: one packet at a time (stop and wait)
• One sender, one receiver
• sender has infinite number of packets to transfer to the receiver
• sender starts one-packet transmission at a time, and will not
proceed with the next new packet transmission until the current
packet has been successfully received & acknowledged by the
receiver.

sender receiver

One packet in transit packets received

packets in the buffer

Transport Layer: 3-44

 “Stop and Wait” Scenario
§We progressively consider more complex cases
• Bit errors
• Packet loss
• Duplicate copies of the same packet
• Long delay (thus also out of order)
• ….
§Designs: rdt2.0 (initial) à rdt3.0 (stop & wait)

sender receiver

X packets received
packets in the buffer errors loss

Packet delivery misbehaviors Transport Layer: 3-45

rdt2.0: channel with bit errors
§ underlying channel may flip bits in packet
• checksum (e.g., Internet checksum) to detect bit errors
§ the question: how to recover from errors?

How do humans recover from “errors” during conversation?

Transport Layer: 3-46

 rdt2.0: channel with bit errors
§ How to detect bit errors in packet?
• Internet checksum algorithm
§ 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 packet upon receiving NAK

§ new mechanisms in rdt2.0 (beyond rdt1.0):

• Error detection at receiver
• Feedback from receiver: control messages (ACK,NAK) from receiver to
sender
• Retransmission at the sender upon NAK feedback
Transport Layer: 3-47
rdt2.0: FSM specifications
rdt_send(data)
snkpkt = make_pkt(data, checksum)
udt_send(sndpkt)
rdt_rcv(rcvpkt) &&
Wait for Wait for isNAK(rcvpkt)
sender call from ACK or udt_send(sndpkt) rdt_rcv(rcvpkt) && corrupt(rcvpkt)
above NAK
udt_send(NAK)

rdt_rcv(rcvpkt) && isACK(rcvpkt)

Wait for
L call from receiver
below

rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)

extract(rcvpkt,data)
deliver_data(data)
udt_send(ACK)

Transport Layer: 3-48

rdt2.0: FSM specification
rdt_send(data)
snkpkt = make_pkt(data, checksum)
udt_send(sndpkt)
rdt_rcv(rcvpkt) &&
Wait for Wait for isNAK(rcvpkt)
sender call from ACK or udt_send(sndpkt) rdt_rcv(rcvpkt) && corrupt(rcvpkt)
above NAK
udt_send(NAK)

rdt_rcv(rcvpkt) && isACK(rcvpkt)

Wait for
L call from receiver
below

Note: “state” of receiver (did the receiver get my

message correctly?) isn’t known to sender unless rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)
extract(rcvpkt,data)
somehow communicated from receiver to sender deliver_data(data)
§ that’s why we need a protocol! udt_send(ACK)

Transport Layer: 3-49

rdt2.0: operation with no errors
rdt_send(data)
snkpkt = make_pkt(data, checksum)
udt_send(sndpkt)
rdt_rcv(rcvpkt) &&
Wait for Wait for isNAK(rcvpkt)
sender call from ACK or udt_send(sndpkt) rdt_rcv(rcvpkt) && corrupt(rcvpkt)
above NAK
udt_send(NAK)

rdt_rcv(rcvpkt) && isACK(rcvpkt)

Wait for
L call from receiver
below

rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)

extract(rcvpkt,data)
deliver_data(data)
udt_send(ACK)

Transport Layer: 3-50

rdt2.0: corrupted packet scenario
rdt_send(data)
snkpkt = make_pkt(data, checksum)
udt_send(sndpkt)
rdt_rcv(rcvpkt) &&
Wait for Wait for isNAK(rcvpkt)
sender call from ACK or udt_send(sndpkt) rdt_rcv(rcvpkt) && corrupt(rcvpkt)
above NAK
udt_send(NAK)

rdt_rcv(rcvpkt) && isACK(rcvpkt)

Wait for
L call from receiver
below

rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)

extract(rcvpkt,data)
deliver_data(data)
udt_send(ACK)

Transport Layer: 3-51

rdt2.0 in action
sender receiver sender receiver
send pkt0 pkt0 send pkt0 pkt0
rcv pkt0 rcv pkt0
ack send ack ack send ack
rcv ack rcv ack
send pkt1 pkt1
send pkt1 pkt1
error
rcv pkt1 rcv garbled pkt1,
ack send ack drop pkt1
rcv ack nack
send pkt2 pkt2 send NACK
rcv pkt2 rcv nack
ack send ack Resend pkt1 pkt1
rcv pkt1
ack send ack
rcv ack1
send pkt2 pkt2
(a) no error rcv pkt2
ack send ack

(b) packet with bit errors

3-52 Transport Layer: 3-52
rdt2.0 has a fatal flaw!
what happens if ACK/NAK handling duplicates:
corrupted? § sender retransmits current pkt
§ sender doesn’t know what if ACK/NAK corrupted
happened at receiver! § sender adds sequence number
§ can’t just retransmit: possible to each pkt
duplicate § receiver discards (doesn’t
deliver up) duplicate pkt

stop and wait

sender sends one packet, then
waits for receiver response
Transport Layer: 3-53
 rdt2.0’s flaw: garbled ACK/NACK
sender receiver sender receiver
send pkt0 pkt0 send pkt0 pkt0
rcv pkt0 rcv pkt0
ack send ack ack send ack
rcv ack rcv ack
send pkt1 pkt1 send pkt1
rcv pkt1 pkt1
ack send ack errors
send pkt2 rcv garbled pkt1
nack
how to know? Pkt2
resend pkt1 send NACK
how to know? Pkt1

(a) Corrupted ack (b) Corrupted NACK

Simply retransmitting upon corrupted ACK/NACK is not sufficient!

Sender cannot tell whether the corrupted message is ACK or NACK!
Receiver cannot tell whether the received message is a new packet or a retransmitted packet! Transport Layer: 3-54
rdt2.1: need seq #!
sender receiver sender receiver
send pkt0 pkt0 send pkt0 pkt0
rcv pkt0 rcv pkt0
ack send ack ack send ack
rcv ack rcv ack
send pkt1 pkt1 send pkt1
rcv pkt1 pkt1
ack send ack rcv garbled pkt1
rcv garbled drop pkt 1
resend pkt1 pkt1 rcv dup pkt1 nack
rcv garbled send NACK
drop dup pkt1 pkt1
ack send ack resend pkt1
rcv ack rcv pkt1
send pkt2 pkt2 ack send ack
rcv pkt2 rcv ack
ack send ack send pkt2 pkt2
rcv pkt2
ack send ack

(a) Corrupted ack (b) Corrupted NACK

Transport Layer: 3-55
rdt2.1: sender, handles garbled ACK/NAKs
rdt_send(data)
sndpkt = make_pkt(0, data, checksum)
udt_send(sndpkt) rdt_rcv(rcvpkt) &&
( corrupt(rcvpkt) ||
Wait for Wait for
ACK or
isNAK(rcvpkt) )
call 0 from
NAK 0 udt_send(sndpkt)
above
rdt_rcv(rcvpkt)
&& notcorrupt(rcvpkt) rdt_rcv(rcvpkt)
&& isACK(rcvpkt) && notcorrupt(rcvpkt)
&& isACK(rcvpkt)
L
L
Wait for Wait for
ACK or call 1 from
rdt_rcv(rcvpkt) && NAK 1 above
( corrupt(rcvpkt) ||
isNAK(rcvpkt) ) rdt_send(data)

udt_send(sndpkt) sndpkt = make_pkt(1, data, checksum)

udt_send(sndpkt)
rdt2.1: receiver, handles garbled ACK/NAKs
rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)
&& has_seq0(rcvpkt)
extract(rcvpkt,data)
deliver_data(data)
sndpkt = make_pkt(ACK, chksum)
udt_send(sndpkt)
rdt_rcv(rcvpkt) && (corrupt(rcvpkt) rdt_rcv(rcvpkt) && (corrupt(rcvpkt)
sndpkt = make_pkt(NAK, chksum) sndpkt = make_pkt(NAK, chksum)
udt_send(sndpkt) udt_send(sndpkt)
Wait for Wait for
rdt_rcv(rcvpkt) && 0 from 1 from rdt_rcv(rcvpkt) &&
not corrupt(rcvpkt) && below below not corrupt(rcvpkt) &&
has_seq1(rcvpkt) has_seq0(rcvpkt)
sndpkt = make_pkt(ACK, chksum) sndpkt = make_pkt(ACK, chksum)
udt_send(sndpkt) udt_send(sndpkt)
rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)
&& has_seq1(rcvpkt)

extract(rcvpkt,data)
deliver_data(data)
sndpkt = make_pkt(ACK, chksum)
udt_send(sndpkt)
Summary: reliable data transfer
Version Channel Mechanism
rdt1.0 Reliable nothing
channel
rdt2.0 bit errors (1)error detection via checksum
(no loss) (2)receiver feedback (ACK/NAK)
(3)retransmission upon NAK
rdt2.1 Same as 2.0 handling fatal flaw with rdt 2.0:
(4)need seq #. for each packet

Transport Layer: 3-60

rdt2.1: 1-bit seq # is enough!
Sender receiver sender receiver
send pkt0 pkt0 send pkt0 pkt0
rcv pkt0 rcv pkt0
ack send ack ack send ack
rcv ack rcv ack
send pkt1 pkt1 send pkt1
rcv pkt1 pkt1
ack send ack rcv garbled pkt1
rcv ack1 drop pkt1
send pkt0 pkt0 NACK
rcv pkt0 rcv NACK send NACK
(new pkt!) send ack
ack resend pkt1 pkt1
rcv pkt1
ack send ack
rcv ack
send pkt0 pkt0
(a) no error (new pkt!) rcv pkt0
ack send ack

(b) packet with bit errors

Transport Layer: 3-63
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

As we will see, TCP uses this approach to be NAK-free

Transport Layer: 3-64

rdt2.2: NAK-free
sender receiver sender receiver
send pkt0 pkt0 send pkt0 pkt0
rcv pkt0 rcv pkt0
ack0 send ack0 ack0 send ack0
rcv ack0 rcv ack0
send pkt1 pkt1 send pkt1
rcv pkt1 pkt1
ack1 send ack1 rcv garbled pkt1
rcv garbled drop pkt 1
resend pkt1 pkt1 rcv pkt1 (dup) ack0
rcv dup ack0 send ack0
drop dup pkt1
ack1 resend pkt1 pkt1
send ack1 rcv pkt1
rcv ack1
send pkt0 pkt0 ack1 send ack1
rcv pkt0 rcv ack1
ack0 send ack0 send pkt0 pkt0
rcv pkt0
ack0 send ack0

(a) Corrupted ack (b) dup ack for garbled pkt

Transport Layer: 3-65
rdt2.2: sender, receiver fragments
rdt_send(data)
sndpkt = make_pkt(0, data, checksum)
udt_send(sndpkt)
rdt_rcv(rcvpkt) &&
( corrupt(rcvpkt) ||
Wait for Wait for
ACK isACK(rcvpkt,1) )
call 0 from
above 0 udt_send(sndpkt)
sender FSM
fragment rdt_rcv(rcvpkt)
&& notcorrupt(rcvpkt)
rdt_rcv(rcvpkt) && && isACK(rcvpkt,0)
(corrupt(rcvpkt) || L
has_seq1(rcvpkt)) Wait for receiver FSM
0 from
udt_send(sndpkt) below fragment
rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)
&& has_seq1(rcvpkt)
extract(rcvpkt,data)
deliver_data(data)
sndpkt = make_pkt(ACK1, chksum)
udt_send(sndpkt) Transport Layer: 3-66
Summary: reliable data transfer
Version Channel Mechanism
rdt1.0 Reliable nothing
channel
rdt2.0 bit errors (1)error detection via checksum
(no loss) (2)receiver feedback (ACK/NAK)
(3)retransmission upon NAK
rdt2.1 Same as 2.0 (4)seq# (1 bit, 0/1) for each pkt
(fatal flaw)
rdt2.2 Same as 2.0 A variant to rdt2.1 (no NAK)
Duplicate ACK = NAK
67 Transport Layer: 3-67
rdt3.0: channels with errors and loss
New channel assumption: underlying channel can also lose
packets (data, ACKs)
• checksum, sequence #s, ACKs, retransmissions will be of help …
but not quite enough

Q: How do humans handle lost sender-to-

receiver words in conversation?

Transport Layer: 3-68

rdt3.0: channels with errors and loss
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 seq #s already handles this!
• receiver must specify seq # of packet being ACKed
§ use countdown timer to interrupt after “reasonable” amount
of time
timeout

Transport Layer: 3-69

rdt3.0 sender
rdt_send(data)
sndpkt = make_pkt(0, data, checksum)
udt_send(sndpkt)
start_timer

Wait for Wait

call 0 from for
above ACK0
rdt_rcv(rcvpkt)
&& notcorrupt(rcvpkt) rdt_rcv(rcvpkt)
&& isACK(rcvpkt,1) && notcorrupt(rcvpkt)
stop_timer && isACK(rcvpkt,0)
stop_timer
Wait Wait for
for call 1 from
ACK1 above

rdt_send(data)
sndpkt = make_pkt(1, data, checksum)
udt_send(sndpkt)
start_timer

Transport Layer: 3-71

rdt3.0 sender
rdt_send(data)
rdt_rcv(rcvpkt) &&
sndpkt = make_pkt(0, data, checksum) ( corrupt(rcvpkt) ||
udt_send(sndpkt) isACK(rcvpkt,1) )
rdt_rcv(rcvpkt) start_timer L
L Wait for Wait
for timeout
call 0 from
ACK0 udt_send(sndpkt)
above
start_timer
rdt_rcv(rcvpkt)
&& notcorrupt(rcvpkt) rdt_rcv(rcvpkt)
&& isACK(rcvpkt,1) && notcorrupt(rcvpkt)
stop_timer && isACK(rcvpkt,0)
stop_timer
Wait Wait for
timeout for call 1 from
udt_send(sndpkt) ACK1 above
start_timer rdt_rcv(rcvpkt)
rdt_send(data) L
rdt_rcv(rcvpkt) &&
( corrupt(rcvpkt) || sndpkt = make_pkt(1, data, checksum)
isACK(rcvpkt,0) ) udt_send(sndpkt)
start_timer
L

Transport Layer: 3-72

Example: rdt3.0 in action
sender receiver sender receiver
send pkt0 pkt0 send pkt0 pkt0
rcv pkt0 rcv pkt0
ack0 send ack0 ack0 send ack0
rcv ack0 rcv ack0
send pkt1 pkt1 send pkt1 pkt1
rcv pkt1 X
loss
ack1 send ack1
rcv ack1
send pkt0 pkt0
rcv pkt0 timeout
ack0 send ack0 resend pkt1 pkt1
rcv pkt1
ack1 send ack1
rcv ack1
send pkt0 pkt0
(a) no loss rcv pkt0
ack0 send ack0

(b) packet loss

Transport Layer: 3-73
rdt3.0 in action
sender receiver
sender receiver send pkt0
pkt0
rcv pkt0
send pkt0 pkt0 send ack0
ack0
rcv pkt0 rcv ack0
ack0 send ack0 send pkt1 pkt1
rcv ack0 rcv pkt1
send pkt1 pkt1 send ack1
rcv pkt1 ack1
ack1 send ack1
X timeout
loss resend pkt1
pkt1 rcv pkt1
timeout
resend pkt1 pkt1
rcv pkt1 rcv ack1 (detect duplicate)
send pkt0 pkt0 send ack1
(detect duplicate)
ack1 send ack1 ack1 rcv pkt0
rcv ack1 rcv ack1 send ack0
send pkt0 pkt0 (ignore) ack0
rcv pkt0
ack0 send ack0 pkt1

(c) ACK loss (d) premature timeout/ delayed ACK

Transport Layer: 3-74
Summary: reliable data transfer
Version Channel Mechanism
rdt1.0 Reliable channel nothing

rdt2.0 bit errors (1)error detection via checksum

(no loss) (2)receiver feedback (ACK/NAK)
(3)retransmission upon NAK
rdt2.1 Same as 2.0 (4)seq# (1 bit) for each pkt
rdt2.2 Same as 2.0 A variant to rdt2.1 (no NAK)
Unexpected ACK = NAK
ACK0 = ACK for pkt0, NAK for pkt1
Rdt3.0 Bit errors + (5) retransmission upon timeout
loss No NAK, only ACK 76 Transport Layer: 3-76
Performance of rdt3.0 (stop-and-wait)
§ U sender: utilization – fraction of time sender busy sending

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

• time to transmit packet into channel:
L 8000 bits
Dtrans = R = 9 = 8 microsecs
10 bits/sec

Transport Layer: 3-77

rdt3.0: stop-and-wait operation
sender receiver
first packet bit transmitted, t = 0

first packet bit arrives

RTT last packet bit arrives, send ACK

ACK arrives, send next

packet, t = RTT + L / R

Transport Layer: 3-78

rdt3.0: stop-and-wait operation
sender receiver

L/R L/R
Usender=
RTT + L / R
.008 RTT
=
30.008
= 0.00027

§ rdt 3.0 protocol performance stinks!

§ Protocol limits performance of underlying infrastructure (channel)

Transport Layer: 3-79

 Mechanisms for reliable data transfer
§ Error detection
• via algorithms such as Internet checksum (in UDP), CRC (later in Chapter 6)
§ Receiver feedback via (ACK + sequence #)
• Duplicate ACK = negative acknowledgment
§ Timer & sequence # for each transmitted packet
• Number of seq. #: ≥ 2 for stop & wait protocol
• Timeout not too small, not too big (≈ 𝑅𝑇𝑇)
§ Retransmission upon timeout or duplicate ACK (i.e., negative ACK)

Transport Layer: 3-80

rdt3.0: pipelined protocols operation
pipelining: sender allows multiple, “in-flight”, yet-to-be-acknowledged
packets
• range of sequence numbers must be increased
• buffering at sender and/or receiver

Transport Layer: 3-81

Pipelining: increased utilization
sender receiver
first packet bit transmitted, t = 0
last bit transmitted, t = L / R

first packet bit arrives

RTT last packet bit arrives, send ACK
last bit of 2nd packet arrives, send ACK
last bit of 3rd packet arrives, send ACK
ACK arrives, send next
packet, t = RTT + L / R
3-packet pipelining increases
utilization by a factor of 3!

U 3L / R .0024
sender = = = 0.00081
RTT + L / R 30.008

Transport Layer: 3-82

Go-Back-N: sender
§ sender: “window” of up to N, consecutive transmitted but unACKed pkts
• k-bit seq # in pkt header

§ cumulative ACK: ACK(n): ACKs all packets up to, including seq # n

• on receiving ACK(n): move window forward to begin at n+1
§ timer for oldest in-flight packet
§ timeout(n): retransmit packet n and all higher seq # packets in window
Transport Layer: 3-84
Go-Back-N: receiver
§ ACK-only: always send ACK for correctly-received packet so far, with
highest in-order seq #
• may generate duplicate ACKs
• need only remember rcv_base
§ on receipt of out-of-order packet:
• can discard (don’t buffer) or buffer: an implementation decision
• re-ACK pkt with highest in-order seq #

Receiver view of sequence number space:

received and ACKed

… … Out-of-order: received but not ACKed

rcv_base
Not received
Transport Layer: 3-85
Go-Back-N in action: No loss
sender window (N=4) sender receiver
012345678 send pkt0
012345678 send pkt1
012345678 send pkt2 receive pkt0, send ack0
012345678 send pkt3 receive pkt1, send ack1
(wait) receive pkt2, send ack2
receive pkt3, send ack3
012345678 rcv ack0, send pkt4
012345 678 rcv ack1, send pkt5
0123456 78 rcv ack2, send pkt6 receive pkt4, send ack4
01234567 8 rcv ack3, send pkt7 receive pkt5, send ack5
receive pkt6, send ack6
pkt0 timeout
receive pkt7, send ack7
pkt1 timeout

pkt2 timeout

pkt3 timeout
pkt4 timeout
Transport Layer: 3-86
Go-Back-N in action: Loss
sender window (N=4) sender receiver
012345678 send pkt0
012345678 send pkt1
send pkt2 receive pkt0, send ack0
012345678
012345678 send pkt3 Xloss receive pkt1, send ack1
(wait)
receive pkt3, discard,
012345678 rcv ack0, send pkt4 (re)send ack1
012345678 rcv ack1, send pkt5 receive pkt4, discard,
(re)send ack1
ignore duplicate ACK receive pkt5, discard,
(re)send ack1
pkt 2 timeout
012345678 send pkt2
012345678 send pkt3
012345678 send pkt4 rcv pkt2, deliver, send ack2
012345678 send pkt5 rcv pkt3, deliver, send ack3
rcv pkt4, deliver, send ack4
rcv pkt5, deliver, send ack5

Transport Layer: 3-87

Selective repeat
§receiver individually acknowledges all correctly received packets
• buffers packets, as needed, for eventual in-order delivery to upper
layer
§sender times-out/retransmits individually for unACKed packets
• sender maintains timer for each unACKed pkt
§sender window
• N consecutive seq #s
• limits seq #s of sent, unACKed packets

Transport Layer: 3-88

Selective repeat: sender, receiver windows

Transport Layer: 3-89

Selective repeat: sender and receiver
sender receiver
data from above: packet n in [rcvbase, rcvbase+N-1]
§ if next available seq # in § send ACK(n)
window, send packet § out-of-order: buffer
timeout(n): § in-order: deliver (also deliver
buffered, in-order packets),
§ resend packet n, restart timer advance window to next not-yet-
ACK(n) in [sendbase,sendbase+N]: received packet
§ mark packet n as received packet n in [rcvbase-N,rcvbase-1]
§ if n smallest unACKed packet, § ACK(n)
advance window base to next otherwise:
unACKed seq # § ignore

Transport Layer: 3-90

Selective Repeat in action
sender window (N=4) sender receiver
012345678 send pkt0
012345678 send pkt1
012345678 send pkt2 receive pkt0, send ack0
012345678 send pkt3 Xloss receive pkt1, send ack1
(wait)
receive pkt3, buffer,
012345678 rcv ack0, send pkt4 send ack3
012345678 rcv ack1, send pkt5
receive pkt4, buffer,
record ack3 arrived send ack4
receive pkt5, buffer,
pkt 2 timeout send ack5
012345678 send pkt2
012345678 (but not 3,4,5)
012345678 rcv pkt2; deliver pkt2,
012345678 pkt3, pkt4, pkt5; send ack2

Q: what happens when ack2 arrives?

Transport Layer: 3-91

Selective repeat:
sender window receiver window
(after receipt) (after receipt)

a dilemma!
0123012 pkt0
0123012 pkt1 0123012
0123012 pkt2 0123012

example:
0123012
0123012 pkt3
X
0123012
§ seq #s: 0, 1, 2, 3 (base 4 counting) pkt0 will accept packet
with seq number 0
§ window size=3 (a) no problem

0123012 pkt0
0123012 pkt1 0123012
0123012 pkt2 X 0123012
X 0123012
X
timeout
retransmit pkt0
0123012 pkt0
will accept packet
with seq number 0
(b) oops!
Transport Layer: 3-92
Selective repeat:
sender window receiver window
(after receipt) (after receipt)

a dilemma!
0123012 pkt0
0123012 pkt1 0123012
0123012 pkt2 0123012

example:
0123012
0123012 pkt3
X
§ seq #s: 0, 1, 2, 3 (base 4 counting) § receiver can’t
0123012
pkt0 will accept packet
see sender side with seq number 0
§ window size=3 (a) no problem
§ receiver
behavior
identical in both
cases!
§0something’s
123012 pkt0
Q: what relationship is needed 0(very)
1 2 3 0 1wrong!
2 pkt1 0123012

between sequence # size and 0123012 pkt2 X

X
0123012

window size to avoid problem

0123012
X
in scenario (b)? timeout
retransmit pkt0
0123012 pkt0
will accept packet
with seq number 0
(b) oops!
Transport Layer: 3-93
 sender window receiver window
Selective repeat: (after receipt) (after receipt)

dilemma (N+1) 0123012 pkt0

pkt1
0123012 0123012
0123012 pkt2 0123012
example: 0123012
0123012 pkt3
§ window size=3 0123012
X
pkt0 will accept packet
§ seq #’s: 0, 1, 2, 3 (a) no problem
with seq number 0
v receiver sees no
difference in two receiver can’t see sender side.
scenarios! receiver behavior identical in both cases!
something’s (very) wrong!
v duplicate data
accepted as new in 0123012 pkt0
(b) 0123012 pkt1 0123012
0123012 pkt2 0123012
X 0123012
Q: what relationship X
between seq # size timeout
retransmit pkt0 X
and window size to 0123012 pkt0
will accept packet
avoid problem in (b)? with seq number 0
(b) oops!
2N Transport Layer: 3-94
 Summary: reliable data transfer
Version Channel Mechanism
rdt1.0 No error/loss nothing
rdt2.0 bit errors (1)error detection via checksum
(no loss) (2)receiver feedback (ACK/NAK)
(3)retransmission upon NAK
rdt2.1 Same as 2.0 (4)seq# (1 bit) for each pkt
rdt2.2 Same as 2.0 (no NAK): Unexpected ACK = NAK
Rdt3.0 errors + loss (5)Retransmission upon timeout; ACK-only

Performance issue: low utilization

Same as 3.0 N sliding window (pipeline)
Goback-N Discard out-of-order pkts (recovery)
Selective Same as 3.0 N sliding window,
Repeat selective recovery 95 Transport Layer: 3-95
Chapter 3: roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
• segment structure
• reliable data transfer
• flow control
• connection management
§ Principles of congestion control
§ TCP congestion control
Transport Layer: 3-96
TCP: overview RFCs: 793,1122, 2018, 5681, 7323
§ point-to-point: § cumulative ACKs
• one sender, one receiver § pipelining:
§ reliable, in-order byte • TCP congestion and flow control
steam: set window size
• no “message boundaries" § connection-oriented:
§ full duplex data: • handshaking (exchange of control
• bi-directional data flow in messages) initializes sender,
same connection receiver state before data exchange
• MSS: maximum segment size § flow controlled:
• sender will not overwhelm receiver

Transport Layer: 3-97

 TCP segment structure
32 bits

source port # dest port # segment seq #: counting

ACK: seq # of next expected sequence number bytes of data into bytestream
byte; A bit: this is an ACK (not segments!)
acknowledgement number
length (of TCP header) head not
len used C EUAP R SF receive window flow control: # bytes
Internet checksum checksum Urg data pointer receiver willing to accept

options (variable length)

C, E: congestion notification
TCP options
application data sent by
RST, SYN, FIN: connection data application into
management (variable length) TCP socket

Transport Layer: 3-98

TCP sequence numbers, ACKs
outgoing segment from sender
Sequence numbers: source port # dest port #
sequence number
• byte stream “number” of acknowledgement number

first byte in segment’s data checksum

rwnd
urg pointer

window size
Acknowledgements: N

• seq # of next byte expected

from other side sender sequence number space

• cumulative ACK sent sent, not- usable not

ACKed yet ACKed but not usable
Q: how receiver handles out-of- (“in-flight”) yet sent

order segments outgoing segment from receiver

• A: TCP spec doesn’t say, - up

source port # dest port #
sequence number

to implementor acknowledgement number

A rwnd
checksum urg pointer
Transport Layer: 3-99
TCP sequence numbers, ACKs
Host A Host B

User types‘C’
Seq=42, ACK=79, data = ‘C’
host ACKs receipt
of‘C’, echoes back ‘C’
Seq=79, ACK=43, data = ‘C’
host ACKs receipt
of echoed ‘C’
Seq=43, ACK=80

simple telnet scenario

Transport Layer: 3-100
TCP round trip time, timeout
Q: how to set TCP timeout Q: how to estimate RTT?
value? § SampleRTT:measured time
§ longer than RTT, but RTT varies! from segment transmission until
ACK receipt
§ too short: premature timeout,
• ignore retransmissions
unnecessary retransmissions
§ SampleRTT will vary, want
§ too long: slow reaction to estimated RTT “smoother”
segment loss • average several recent
measurements, not just current
SampleRTT

Transport Layer: 3-101

TCP round trip time, timeout
EstimatedRTT = (1- a)*EstimatedRTT + a*SampleRTT
§ exponential weighted moving average (EWMA)
§ influence of past sample decreases exponentially fast
RTT: gaia.cs.umass.edu to fantasia.eurecom.fr
§ typical value: a = 1/8 350

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

(milliseconds)
300

250

RTT (milliseconds)
RTT
200

sampleRTT
150

EstimatedRTT

100
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
time (seconds)
SampleRTT Estimated RTT
Transport Layer: 3-102
TCP round trip time, timeout
§ timeout interval: EstimatedRTT plus “safety margin”
• large variation in EstimatedRTT: want a larger safety margin
TimeoutInterval = EstimatedRTT + 4*DevRTT

estimated RTT “safety margin”

§ DevRTT: EWMA of SampleRTT deviation from EstimatedRTT:

DevRTT = (1-b)*DevRTT + b*|SampleRTT-EstimatedRTT|
(typically, b = 1/4)

* Check out the online interactive exercises for more examples: http://gaia.cs.umass.edu/kurose_ross/interactive/
Transport Layer: 3-103
TCP Sender (simplified)
event: data received from event: timeout
application § retransmit segment that
caused timeout
§ create segment with seq #
§ restart timer
§ seq # is byte-stream number
of first data byte in segment
event: ACK received
§ start timer if not already
running § if ACK acknowledges
previously unACKed segments
• think of timer as for oldest
unACKed segment • update what is known to be
ACKed
• expiration interval:
TimeOutInterval • start timer if there are still
unACKed segments
Transport Layer: 3-104
TCP Receiver: ACK generation [RFC 5681]
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 seq. # . indicating seq. # of next expected byte
Gap detected

arrival of segment that immediate send ACK, provided that

partially or completely fills gap segment starts at lower end of gap

Transport Layer: 3-105

TCP: retransmission scenarios
Host A Host B Host A Host B

SendBase=92
Seq=92, 8 bytes of data Seq=92, 8 bytes of data

Seq=100, 20 bytes of data

timeout

timeout
ACK=100
X
ACK=100
ACK=120

Seq=92, 8 bytes of data Seq=92, 8

SendBase=100 bytes of data send cumulative
SendBase=120 ACK for 120
ACK=100
ACK=120

SendBase=120

lost ACK scenario premature timeout

Transport Layer: 3-106

TCP: retransmission scenarios
Host A Host B

Seq=92, 8 bytes of data

Seq=100, 20 bytes of data

ACK=100
X
ACK=120

Seq=120, 15 bytes of data

cumulative ACK covers

for earlier lost ACK

Transport Layer: 3-107

TCP fast retransmit
Host A Host B
TCP fast retransmit
if sender receives 3 additional
ACKs for same data (“triple Seq=92
Seq=1
,8 bytes
of data
duplicate ACKs”), resend unACKed 00, 20
bytes
of data
segment with smallest seq # X
§ likely that unACKed segment lost,
=100
so don’t wait for timeout A CK

=100

timeout
A CK
=100
A CK
=100
Receipt of three duplicate ACKs A CK

indicates 3 segments received Seq=100, 20 bytes of data

after a missing segment – lost

segment is likely. So retransmit!

Transport Layer: 3-108

Chapter 3: roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
• segment structure
• reliable data transfer
• flow control
• connection management
§ Principles of congestion control
§ TCP congestion control
Transport Layer: 3-109
TCP flow control application
application may process
remove data from application
TCP socket buffers ….
TCP socket OS
receiver buffers
… slower than TCP
receiver is delivering
(sender is sending) TCP
code

IP
flow control code
receiver controls sender, so
sender won’t overflow
receiver’s buffer by from sender

transmitting too much, too fast receiver protocol stack

TCP flow control
§ receiver “advertises” free
buffer space by including
rwnd value in TCP header of to application process
receiver-to-sender segments
• RcvBuffer size set via socket RcvBuffer buffered data
options (typical default is 4096
bytes) rwnd free buffer space
• many operating systems
autoadjust RcvBuffer
§ sender limits amount of TCP segment payloads
unacked (“in-flight”) data to
receiver’s rwnd value receiver-side buffering
§ guarantees receive buffer will
not overflow
TCP connection management
before exchanging data, sender/receiver “handshake”:
§ agree to establish connection (each knowing the other willing to establish connection)
§ agree on connection parameters (e.g., starting seq #s)

application application

connection state: ESTAB connection state: ESTAB

connection variables: connection Variables:
seq # client-to-server seq # client-to-server
server-to-client server-to-client
rcvBuffer size rcvBuffer size
at server,client at server,client

network network

Socket clientSocket = Socket connectionSocket =

newSocket("hostname","port number"); welcomeSocket.accept();
Transport Layer: 3-118
Agreeing to establish a connection
2-way handshake:

Q: will 2-way handshake always

Let’s talk work in network?
ESTAB
ESTAB
OK § variable delays
§ retransmitted messages (e.g.
req_conn(x)) due to message loss
§ message reordering
choose x § can’t “see” other side
req_conn(x)
ESTAB
acc_conn(x)
ESTAB

Transport Layer: 3-119

 TCP 3-way handshake
Server state
Client state
serverSocket = socket(AF_INET,SOCK_STREAM)
serverSocket.bind((‘’,serverPort))
serverSocket.listen(1)
clientSocket = socket(AF_INET, SOCK_STREAM) connectionSocket, addr = serverSocket.accept()
LISTEN
clientSocket.connect((serverName,serverPort)) LISTEN
choose init seq num, x
send TCP SYN msg
SYNSENT SYNbit=1, Seq=x
choose init seq num, y
send TCP SYNACK
msg, acking SYN SYN RCVD
SYNbit=1, Seq=y
ACKbit=1; ACKnum=x+1
received SYNACK(x)
ESTAB indicates server is live;
send ACK for SYNACK;
this segment may contain ACKbit=1, ACKnum=y+1
client-to-server data
received ACK(y)
indicates client is live
ESTAB

Transport Layer: 3-123

How to set SYNC, ACK bit?
32 bits

source port # dest port #

sequence number
ACK: ACK #
valid acknowledgement number
head not
len used U A P R S F receive window
checksum Urg data pointer

RST, SYN, FIN: options (variable length)

connection estab
(setup, teardown
commands)
application
data
(variable length)
Closing a TCP connection
§ 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

Transport Layer: 3-125

Closing TCP connection (i.e., two 1-way
subconnections)
client state server state
ESTAB ESTAB
clientSocket.close()

FIN_WAIT_1 can no longer FINbit=1, seq=x

send but can
receive data CLOSE_WAIT
ACKbit=1; ACKnum=x+1
can still
FIN_WAIT_2 wait for server send data
close

LAST_ACK
FINbit=1, seq=y
TIMED_WAIT can no longer
send data
ACKbit=1; ACKnum=y+1
timed wait
for 2*max CLOSED
segment lifetime

CLOSED Makes the client wait for a duration long enough for an ACK to be lost
and a FIN to arrive. If a FIN arrives, restart the timer 2*max-segment-lifetime
Drop any delayed segments during timer=2*max-segment-time (2min default) Transport Layer: 3-126
Chapter 3: roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
§ Principles of congestion control
§ TCP congestion control
§ Evolution of transport-layer
functionality
Transport Layer: 3-127
Principles of congestion control
Congestion:
§ informally: “too many sources sending too much data too fast for
network to handle”
§ manifestations:
• long delays (queueing in router buffers)
• packet loss (buffer overflow at routers)
§ different from flow control! congestion control:
§ a top-10 problem! too many senders,
sending too fast

flow control: one sender

too fast for one receiver
Transport Layer: 3-128
Approaches towards congestion control

End-end congestion control:

§ no explicit feedback from
network
§ congestion inferred from data data
ACKs
observed loss, delay ACKs

§ approach taken by TCP

Transport Layer: 3-139

Approaches towards congestion control
Network-assisted congestion
control: explicit congestion info
§ routers provide direct feedback
to sending/receiving hosts with data data
ACKs
flows passing through congested ACKs

router
§ may indicate congestion level or
explicitly set sending rate
§ TCP ECN, ATM, DECbit protocols
Transport Layer: 3-140
Chapter 3: roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
§ Principles of congestion control
§ TCP congestion control
§ Evolution of transport-layer
functionality
Transport Layer: 3-141
TCP Congestion Control
vIdea
§ Assumes best-effort network
§ Each source determines network capacity for itself
§ Implicit feedback via ACKs or timeout events
§ Feedback control system in practice
§ ACKs pace transmission (self-clocking)

vChallenge
§ Determining initial available capacity
§ Adjusting to changes in capacity in a timely manner

Transport Layer: 3-142

TCP Congestion Control
§ Assumptions for congestion control
• TCP pipelined reliable data transfer (SR in the common cases)
• Works with TCP flow control
• All losses of TCP segments are due to Internet congestion
• Ignore the transmission errors (since link quality is good in general)
§ Mechanism: Window-based congestion control
• Adjust the window size for SR to change the TCP sending rate
§ Changes in congestion window size (cwnd)
• Slow increases to absorb new bandwidth
• Quick decreases to eliminate congestion

Transport Layer: 3-143

TCP Congestion Control
r sender limits transmission: How does sender perceive
LastByteSent-LastByteAcked congestion?
£ cwnd r loss event = timeout or 3
sender sequence number space duplicate acks
cwnd
r TCP sender reduces rate
(cwnd) after loss event
last byte last byte three mechanisms:
ACKed sent, not-yet
sent
ACKed
(“in-flight”)
m AIMD: how to grow cwnd
m slow start: startup
r cwnd is dynamic, function of
m conservative after loss
perceived network congestion (timeout, duplicate ACKs)
events
AIMD Rule: additive increase, multiplicative
decrease
r Approach: increase transmission rate (window size),
probing for usable bandwidth, until loss occurs
m additive increase: increase cwnd by 1 MSS every
RTT until loss detected
m multiplicative decrease: cut cwnd by 50% after
congestion
window

loss
congestion window size
24 Kbytes

Saw tooth
behavior: probing
16 Kbytes

for bandwidth
8 Kbytes

time
time
What AIMD? TCP Fairness
Two competing sessions:
r Additive increase gives slope of 1, as throughout
increases
r multiplicative decrease decreases throughput
proportionally
R equal bandwidth share
Connection 2 throughput

loss: decrease window by factor of 2

congestion avoidance: additive increase
loss: decrease window by factor of 2
congestion avoidance: additive increase

Connection 1 throughput R
TCP Congestion Control (RFC 5681)

How to implement TCP Congestion Control?

Multiple algorithms work together:

r slow start: how to jump-start
r congestion avoidance: additive increase
r fast retransmit/fast recovery: recover from
single packet loss: multiplicative decrease
r retransmission upon timeout: conservative
loss/failure handling
TCP Slow Start
§ When connection • When connection
begins, cwnd £ 2 MSS, begins, increase rate
typically, set cwnd = exponentially fast until
1MSS cwnd reaches a
• Example: MSS = 500
bytes & RTT = 200 msec threshold value: slow-
• initial rate = 20 kbps start-threshold
§ available bandwidth ssthresh
may be >> MSS/RTT m cwnd < ssthresh
• desirable to quickly ramp
up to respectable rate
TCP Slow Start (more)
§When connection Host A Host B
begins, increase rate
exponentially when one segm
ent

RTT
cwnd<ssthresh two segm
• Goal: double cwnd ents

every RTT by setting

• Action: cwnd += 1 MSS four segm
ents
for every ACK received
§Summary: initial rate is
slow but ramps up
exponentially fast time
Congestion Avoidance
§ Goal: increase cwnd by 1 MSS per RTT until congestion (loss) is
detected

• Conditions: when cwnd > ssthresh and no loss occurs

• Actions: cwnd += (MSS/cwnd)*MSS (bytes) upon every incoming non-

duplicate ACK
TCP Congestion Control
Algoritms condition Design action

Slow Start cwnd <= ssthresh; cwnd doubles per RTT cwnd+=1MSS per ACK

Congestion cwnd++ per RTT cwnd+=1/cwnd * MSS per

Avoidance cwnd > ssthresh (additive increase) ACK

Transport Layer: 3-163

When loss occurs
§Detecting losses and reacting to them:

• through duplicate ACKs

• fast retransmit / fast recovery
• Goal: multiplicative decrease cwnd upon loss

• through retransmission timeout

• Goal: reset everything
Fast Retransmit/Fast Recovery
§ fast retransmit: to detect and Philosophy:
repair loss, based on incoming q 3 dup ACKs to infer losses
duplicate ACKs and differentiate from
• use 3 duplicate ACKs to infer transient out-of-order
packet loss
delivery
• set ssthresh = max(cwnd/2, q What about only 1 or 2
2MSS)
dup ACKs?
• cwnd = ssthresh + 3MSS q Do nothing; this allows for
• retransmit the lost packet transient out-of-order
delivery
§ fast recovery: governs the
transmission of new data until a q receiving each duplicate
non-duplicate ACK arrives ACK indicates one more
• increase cwnd by 1 MSS upon every packet left the network and
duplicate ACK
arrived at the receiver
Transport Layer: 3-165
Algorithm for fast rexmit/fast
recovery
§ Initially, fastretx = false;
§ If upon 3rd duplicate ACK
• ssthresh = max (cwnd/2, 2*MSS)
• cwnd = ssthresh + 3*MSS
• why add 3 packets here?
• retransmit the lost TCP packet
• Set fastretx = true;
§ If fastretx == true; upon each additional duplicate ACK
• cwnd += 1 MSS
• transmit a new packet if allowed
• by the updated cwnd and rwnd
§ If fastretx == true; upon a new (i.e., non-duplicate) ACK
• cwnd = ssthresh
• Fastretx = false; // After fast retx/fast recovery, cwnd decreases by half

Transport Layer: 3-166

Retransmission Timeout
when retransmission timer expires
• ssthresh = max ( cwnd/2, 2*MSS)
• cwnd should be flight size to be more accurate
• see RFC 2581

• cwnd = 1 MSS

• retransmit the lost TCP packet

§ why resetting?
• heavy loss detected

Transport Layer: 3-167

TCP Congestion Window Trace

Transport Layer: 3-168

TCP Congestion Control Summary
Algoritms condition Design action
Slow Start cwnd <= ssthresh; cwnd doubles per RTT cwnd+=1MSS per ACK
Congestion cwnd++ per RTT (additive cwnd+=(MSS/cwnd) * MSS
Avoidance cwnd > ssthresh increase) per ACK
ssthresh = max(cwnd/2,2MSS)
fast reduce the cwnd by half cwnd = ssthresh + 3 MSS;
retransmit 3 duplicate ACK (multicative decreasing) retx the lost packet
finish the 1/2 reduction of
receiving a new ACK cwnd in fast retx/fast cwnd = ssthresh;
fast recovery after fast retx recovery tx if allowed by cwnd
upon a dup ACK after cwnd +=1MSS;
fast retx before fast ("transition phrase) Note: it is different from slow
recovery start.
ssthresh = max(cwnd/2,2MSS)
cwnd = 1MSS;
RTO timeout time out Reset everything retx the lost packet Transport Layer: 3-169
Putting Things Together in TCP
How Selective repeat, congestion control, flow control work together:
§ use selective repeat to do reliable data transfer for a window of
packets win at any time
§ update win = min (cwnd, rwnd)
• cwnd is updated by TCP congestion control
• rwnd is updated by TCP flow control
§ Example: cwnd = 20; rwnd = 10
• Then win=10

Transport Layer: 3-170

Illustrative Example

Transport Layer: 3-171

Example Setting
§ Use all following TCP congestion control algorithms:
• Slow start
• Congestion avoidance (CA)
• Fast retransmit/fast recovery
• Retransmission timeout (say, RTO=500ms)
§ When cwnd=ssthresh, use slow start algorithm (instead of CA)
§ Assume rwnd is always large enough, then the send window size min(rwnd,cwnd) =cwnd
§ Assume 1 acknowledgement per packet (i.e., no delayed ACK is used), and we use TCP cumulative ACK (i.e.,
ACK # = (largest sequence # received in order at the receiver + 1) )
§ Assume each packet size is 1 unit (1B) for simple calculation
§ TCP sender has infinite packets to send, 1, 2, 3, 4, 5,….
§ Assume packet #5 is lost once
§ Assume that the receiver will buffer out of order packets (like selective repeat)

We will how TCP congestion control algorithms work together

Transport Layer: 3-172

 CC algorithm SR after algo runs
1
slow start cwnd =1 pkt 1
ssh =4 1 1
ack2
slow start (upon ack2) cwnd =1+1=2 2 pkt 2
3
ssh =4 2 3 pkt 3 2
3
ack3
slow start (upon ack3) cwnd =2+1=3 4 pkt 4
ssh =4 3 4 5 5 pkt 5 4
ack4
slow start (upon ack4) cwnd =3+1=4 6 X
7 pkt 6
ssh =4 4 5 6 7 ack5 pkt 7 6
slow start (upon ack5 ) cwnd =4+1=5 8 7
ssh =4 9
ack5 (1 dup) pkt 8
st
5 6 7 8 9 pkt 9 8
ack5 (2 dup)
nd
Do nothing upon ack5 (1st dup ) 9
Do nothing upon ack5 (2nd| dup )
ack5 (3 dup)
rd

Fast retransmit (upon 3 dup ack5 )

5

Pkt 5
)
ack5 (4 dup
th
5 6 7 8 9
5
Fast recovery w/ additional dup ACK (upon 4th dup)
10
ssh = 2, cwnd = 5 +1 =6 Pkt 10
send pkt 10
ack10 10
5 6 7 8 9 10 ack11

Transport Layer: 3-173

 CC algorithm cwnd after algo runs
Pkt 5
Fast recovery w/ additional dup ACK (upon 4th dup) 5
ssh = 2, cwnd = 5 +1 =6; send pkt 10 10
5 6 7 8 9 10
Pkt 10
Fast recovery w/ a new ACK (upon ack10)
ssh =2 cwnd = ssh = 2 CK)
10
Fast retx/fast recovery is over Ack10 (new A

10 11 11
Pkt 11
11
Slow start also upon ack10 12 pkt 12
12
ssh =2 cwnd =2 + 1 = 3
Send new packet 12
10 11 12 ack 11
Pkt 13
Congestion avoidance upon ack11 13 13
ssh =2
Ack12
11 12 13
Pkt 14
Congestion avoidance upon ack 12 14 14
ssh =2
Ack13
12 13 14
Pkt 15
Congestion avoidance upon ack 13 15
16 Pkt 16 15
ssh =2 cwnd = 3 + 3/3=4 16
Send packets 15, 16 Ack14
16
Ack15 Ack
13 14 15 16 Ack17

Transport Layer: 3-174

Transport layer: roadmap
§ Transport-layer services
§ Multiplexing and demultiplexing
§ Connectionless transport: UDP
§ Principles of reliable data transfer
§ Connection-oriented transport: TCP
§ Principles of congestion control
§ TCP congestion control
§ Evolution of transport-layer
functionality
Transport Layer: 3-179
Evolving transport-layer functionality
§ TCP, UDP: principal transport protocols for 40 years
§ different “flavors” of TCP developed, for specific scenarios:
Scenario Challenges
Long, fat pipes (large data Many packets “in flight”; loss shuts down
transfers) pipeline
Wireless networks Loss due to noisy wireless links, mobility;
TCP treat this as congestion loss
Long-delay links Extremely long RTTs
Data center networks Latency sensitive
Background traffic flows Low priority, “background” TCP flows

§ moving transport–layer functions to application layer, on top of UDP

• HTTP/3: QUIC
Transport Layer: 3-180
QUIC: Quick UDP Internet Connections
§ application-layer protocol, on top of UDP
• increase performance of HTTP
• deployed on many Google servers, apps (Chrome, mobile YouTube app)

HTTP/2 HTTP/2 (slimmed)

Application HTTP/3
TLS QUIC

Transport TCP UDP

Network IP IP

HTTP/2 over TCP HTTP/2 over QUIC over UDP

Transport Layer: 3-181

QUIC: Quick UDP Internet Connections
adopts approaches we’ve studied in this chapter for
connection establishment, error control, congestion control
• error and congestion control: “Readers familiar with TCP’s loss
detection and congestion control will find algorithms here that parallel
well-known TCP ones.” [from QUIC specification]
• connection establishment: reliability, congestion control,
authentication, encryption, state established in one RTT

§ multiple application-level “streams” multiplexed over single QUIC

connection
• separate reliable data transfer, security
• common congestion control
Transport Layer: 3-182
QUIC: Connection establishment

TCP handshake
(transport layer) QUIC handshake

data
TLS handshake
(security)
data

TCP (reliability, congestion control QUIC: reliability, congestion control,

state) + TLS (authentication, crypto authentication, crypto state
state)
§ 1 handshake
§ 2 serial handshakes

Transport Layer: 3-183

 QUIC: streams: parallelism, no HOL blocking
HTTP HTTP
GET GET HTTP
GET
HTTP HTTP
application

GET GET
HTTP
GET QUIC QUIC QUIC QUIC QUIC QUIC
encrypt encrypt encrypt encrypt encrypt encrypt
QUIC QUIC QUIC QUIC QUIC QUIC
TLS encryption TLS encryption RDT RDT RDT RDT
error!
RDT RDT

QUIC Cong. Cont. QUIC Cong. Cont.

TCP RDT TCP
error! RDT
transport

TCP Cong. Contr. TCP Cong. Contr. UDP UDP

(a) HTTP 1.1 (b) HTTP/2 with QUIC: no HOL blocking

Transport Layer: 3-184
Chapter 3: summary
§ principles behind transport Up next:
layer services: § leaving the network
• multiplexing, demultiplexing “edge” (application,
• reliable data transfer transport layers)
• flow control § into the network “core”
• congestion control
§ two network-layer
§ instantiation, implementation chapters:
in the Internet • data plane
• UDP • control plane
• TCP

Transport Layer: 3-185