Chapter 3

Transport
Layer

A note on the use of these ppt slides:

Computer
We’re making these slides freely available to all (faculty, students, readers).
They’re in PowerPoint form so you see the animations; and can add, modify,
Networking: A
and delete slides (including this one) and slide content to suit your needs.
They obviously represent a lot of work on our part. In return for use, we only
Top Down
ask the following:
 If you use these slides (e.g., in a class) that you mention their source
Approach
(after all, we’d like people to use our book!) 6th edition
 If you post any slides on a www site, that you note that they are adapted Jim Kurose, Keith
from (or perhaps identical to) our slides, and note our copyright of this
material.
Ross
Addison-Wesley
Thanks and enjoy! JFK/KWR
March 2012
All material copyright 1996-2012
J.F Kurose and K.W. Ross, All Rights Reserved

Transport Layer 3-1

Chapter 3: Transport Layer
our goals:
 understand  learn about Internet
principles behind transport layer
transport layer protocols:
services:  UDP: connectionless
 multiplexing, transport
demultiplexing  TCP: connection-
 reliable data oriented reliable
transfer transport
 flow control  TCP congestion
 congestion control
control

Transport Layer 3-2

Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-3
Transport services and
protocols applicatio
n
 provide logical transport

communication between network

data link
app processes running on physical

different hosts

lo
gi
ca
 transport protocols run in

enl
end systems

d-
en
 send side: breaks app

d
tr
messages into

a
ns
segments, passes to

po
network layer

rt
 rcv side: reassembles applicatio
n
segments into transport
network
messages, passes to data link
physical
app layer
 more than one transport
protocol available to apps
 Internet: TCP and UDP
Transport Layer 3-4
Transport vs. network
layer
 network layer: household analogy:
logical
communication 12 kids in Ann’s house
sending letters to 12
between hosts kids in Bill’s house:
 transport layer:  hosts = houses

logical  processes = kids

communication  app messages =

letters in envelopes
between  transport protocol =
processes Ann and Bill who
 relies on, demux to in-house
enhances, siblings
network layer  network-layer
services protocol = postal
service
Transport Layer 3-5
Internet transport-layer
protocols applicatio
 reliable, in-order n
transport
delivery (TCP) network
data link
network
 congestion control physical

lo
network data link

gi
data link physical
 flow control

ca
physical
network

l en
 connection setup data link

d-
physical

en
 unreliable, network

d
data link

tr
unordered delivery:

a
physical

ns
network

po
UDP data link

r
physical

t
network
 no-frills extension of data link
physical
applicatio
n
“best-effort” IP network
data link transport
network
physical
 services not data link
physical

available:
 delay guarantees
 bandwidth
guarantees Transport Layer 3-6
Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-7
 Multiplexing/
demultiplexing
multiplexing at sender:
handle data from demultiplexing at receiver:
multiple use header info to deliver
sockets, add transport received segments to corre
header (later used for socket
demultiplexing)
application

application P1 P2 application socket

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

Transport Layer 3-8

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

TCP/UDP segment format

Transport Layer 3-9

Connectionless
demultiplexing
 recall: created socket  recall: when creating
has host-local port #: datagram to send
DatagramSocket mySocket1
= new
into UDP socket, must
DatagramSocket(12534); specify
 destination IP address
 destination port #
 when host receives IP datagrams with
UDP segment: same dest. port #,
 checks destination but different source
port # in segment IP addresses and/or
 directs UDP segment source port numbers
will be directed to
to socket with that
port # same socket at dest

Transport Layer 3-10

Connectionless demux:
example
DatagramSocket serverSocket
= new DatagramSocket
DatagramSocket (6428); DatagramSocket
mySocket2 = new mySocket1 = new
DatagramSocket DatagramSocket
(9157); application
(5775);
application P1 application
P3 P4
transport
transport transport
network
network link network
link physical link
physical physical

source port: 6428 source port: ?

dest port: 9157 dest port: ?

source port: 9157 source port: ?

dest port: 6428 dest port: ?
Transport Layer 3-11
Connection-oriented
demux
 TCP socket  server host may
identified by 4- support many
tuple: simultaneous TCP
 source IP address sockets:
 source port number  each socket identified
by its own 4-tuple
 dest IP address
 dest port number
 web servers have
different sockets for
 demux: receiver each connecting
uses all four values client
to direct segment  non-persistent HTTP
to appropriate will have different
socket socket for each
request
Transport Layer 3-12
 Connection-oriented demux:
example
application
application P4 P5 P6 application
P3 P2 P3
transport
transport transport
network
network link network
link physical link
physical server: physical
IP
address
B
host: IP source IP,port: B,80 host: IP
address dest IP,port: A,9157 source IP,port: C,5775 address
A dest IP,port: B,80 C
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 3-13
 Connection-oriented demux:
example
threaded server
application
application application
P4
P3 P2 P3
transport
transport transport
network
network link network
link physical link
physical server: physical
IP
address
B
host: IP source IP,port: B,80 host: IP
address dest IP,port: A,9157 source IP,port: C,5775 address
A dest IP,port: B,80 C
source IP,port: A,9157
dest IP, port: B,80
source IP,port: C,9157
dest IP,port: B,80

Transport Layer 3-14

Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-15
UDP: User Datagram Protocol
[RFC 768]
 “no frills,” “bare bones”  UDP use:
Internet transport  streaming
protocol
multimedia apps
 “best effort” service,
UDP segments may be: (loss tolerant, rate
 lost sensitive)
 delivered out-of-order  DNS
to app  SNMP
 connectionless:  reliable transfer
 no handshaking
between UDP sender, over UDP:
receiver  add reliability at
 each UDP segment application layer
handled  application-specific
independently of error recovery!
others

Transport Layer 3-16

UDP: segment header
length, in bytes of
32 bits UDP segment,
source port # dest port # including header

length checksum
why is there a UDP?
 no connection
application establishment (which
data can add delay)
(payload)  simple: no connection
state at sender,
receiver
 small header size
UDP segment format  no congestion control:
UDP can blast away as
fast as desired

Transport Layer 3-17

UDP checksum
Goal: detect “errors” (e.g., flipped bits) in
transmitted segment
sender: receiver:
 treat segment contents,  compute checksum of
including header fields, received segment
as sequence of 16-bit  check if computed
integers
checksum equals
 checksum: addition checksum field value:
(one’s complement  NO - error detected
sum) of segment
contents  YES - no error detected.
 sender puts checksum But maybe errors
value into UDP nonetheless? More later
checksum field ….

Transport Layer 3-18

 Internet checksum:
example
example: add two 16-bit integers
1 1 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0
1 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 1 0 1 1 1 0 1 1 1 0 1 1 1 1 0 0
checksum 1 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

Transport Layer 3-19

Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-20
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)

Transport Layer 3-21

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)

Transport Layer 3-22

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)

Transport Layer 3-23

 Reliable data transfer: getting
started
rdt_send(): called from above, deliver_data(): called
(e.g., by app.). Passed data to by rdt to deliver data to
deliver to receiver upper layer upper

send receive
side side

udt_send(): called by rdt, rdt_rcv(): called when packet

to transfer packet over arrives on rcv-side of channel
unreliable channel to
receiver
Transport Layer 3-24
Reliable data transfer: getting
started
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
event causing state transition
actions taken on state transition
state: when in this
“state” next state state state
uniquely 1 event
determined by 2
actions
next event

Transport Layer 3-25

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) Wait for rdt_rcv(packet)

call from call from extract (packet,data)
above packet = make_pkt(data) below deliver_data(data)
udt_send(packet)

sender receiver

Transport Layer 3-26

rdt2.0: channel with bit
errors
 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
How do humansinrecover
new mechanisms from rdt1.0
rdt2.0 (beyond “errors”
):
during conversation?
 error detection
 receiver feedback: control msgs (ACK,NAK) rcvr-
>sender

Transport Layer 3-27

rdt2.0: channel with bit
errors
 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

Transport Layer 3-28

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

rdt_rcv(rcvpkt) && isACK(rcvpkt)

Wait for

call from
below
sender
rdt_rcv(rcvpkt) &&
notcorrupt(rcvpkt)
extract(rcvpkt,data)
deliver_data(data)
udt_send(ACK)

Transport Layer 3-29

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

rdt_rcv(rcvpkt) && isACK(rcvpkt)

Wait for
 call from
below

rdt_rcv(rcvpkt) &&
notcorrupt(rcvpkt)
extract(rcvpkt,data)
deliver_data(data)
udt_send(ACK)

Transport Layer 3-30

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

rdt_rcv(rcvpkt) && isACK(rcvpkt)

Wait for
 call from
below

rdt_rcv(rcvpkt) &&
notcorrupt(rcvpkt)
extract(rcvpkt,data)
deliver_data(data)
udt_send(ACK)

Transport Layer 3-31

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

stop and wait

sender sends one
packet,
then waits for
receiver
response Transport Layer 3-32
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)


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)

Transport Layer 3-33

 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) && rdt_rcv(rcvpkt) &&
(corrupt(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)

Transport Layer 3-34

rdt2.1: discussion
sender: receiver:
 seq # added to pkt  must check if
 two seq. #’s (0,1)
received packet is
will suffice. Why? duplicate
 must check if  state indicates
received ACK/NAK whether 0 or 1 is
corrupted expected pkt seq
 twice as many #
states  note: receiver can
 state must not know if its last
“remember” whether ACK/NAK received
“expected” pkt
should have seq # of OK at sender
0 or 1
Transport Layer 3-35
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

Transport Layer 3-36

 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) || 
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 3-37
rdt3.0: channels with errors and
loss
new assumption: approach: sender waits
underlying “reasonable” amount
channel can also of time for ACK
lose packets  retransmits if no ACK
received in this time
(data, ACKs)  if pkt (or ACK) just
 checksum, seq. #, delayed (not lost):
ACKs,  retransmission will be
retransmissions duplicate, but seq. #’s
will be of help … already handles this
but not enough  receiver must specify
seq # of pkt being
ACKed
 requires countdown
timer
Transport Layer 3-38
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 
 Wait for Wait
for timeout
call 0from
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) 
rdt_rcv(rcvpkt) &&
sndpkt = make_pkt(1, data, checksum)
( corrupt(rcvpkt) || udt_send(sndpkt)
isACK(rcvpkt,0) ) start_timer


Transport Layer 3-39

 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
ack1 send ack1 loss
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 3-40
 rdt3.0 in
action sender receiver
sender receiver send pkt0 pkt0
send pkt0 pkt0 rcv pkt0
ack0 send ack0
rcv pkt0
send ack0 rcv ack0
ack0 send pkt1 pkt1
rcv ack0 rcv pkt1
send pkt1 pkt1
rcv pkt1 send ack1
ack1 ack1
send ack1
X
loss timeout
resend pkt1 pkt1
rcv pkt1
timeout
resend pkt1 pkt1 rcv ack1 pkt0 (detect duplicate)
rcv pkt1 send pkt0 send ack1
(detect duplicate) ack1
ack1 send ack1 rcv ack1 rcv pkt0
rcv ack1 ack0 send ack0
pkt0 send pkt0 pkt0
send pkt0 rcv pkt0
rcv pkt0 ack0 (detect duplicate)
ack0 send ack0 send ack0

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

Transport Layer 3-41

Performance of rdt3.0
 rdt3.0 is correct, but performance stinks
 e.g.: 1 Gbps link, 15 ms prop. delay, 8000 bit packet:

L 8000 bits
Dtrans = R = 9 = 8 microsecs
10 bits/sec

 U sender: utilization – fraction of time sender busy

sending L/R .008
U = 0.00027
sender = =
30.008
RTT + L / R
 if RTT=30 msec, 1KB pkt every 30 msec:
33kB/sec thruput over 1 Gbps link
 network protocol limits use of physical
resources!
Transport Layer 3-42
 rdt3.0: stop-and-wait
operation
sender receiver
first packet bit transmitted, t = 0
last packet bit transmitted, t = L / R

first packet bit arrives

RTT last packet bit arrives, send
ACK

ACK arrives, send next

packet, t = RTT + L / R

U L/R .008
sender = = = 0.00027
RTT + L / R 30.008

Transport Layer 3-43

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
Transport Layer 3-44
 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 3-45

Pipelined protocols:
overview
Go-back-N: Selective Repeat:
 sender can have  sender can have up
up to N unacked to N unack’ed
packets in pipeline packets in pipeline
 receiver only sends  rcvr sends individual
cumulative ack ack for each packet
 doesn’t ack packet
if there’s a gap
 sender has timer  sender maintains
for oldest unacked timer for each
packet unacked packet
 when timer expires,  when timer expires,
retransmit all retransmit only that
unacked packets unacked packet
Transport Layer 3-46
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
Transport Layer 3-47
GBN: sender extended FSM
rdt_send(data)
if (nextseqnum < base+N) {
sndpkt[nextseqnum] = make_pkt(nextseqnum,data,chksum)
udt_send(sndpkt[nextseqnum])
if (base == nextseqnum)
start_timer
nextseqnum++
}
 else
refuse_data(data)
base=1
nextseqnum=1
timeout
start_timer
Wait
udt_send(sndpkt[base])
rdt_rcv(rcvpkt) udt_send(sndpkt[base+1])
&& corrupt(rcvpkt) …
udt_send(sndpkt[nextseqnum-
1])
rdt_rcv(rcvpkt) &&
notcorrupt(rcvpkt)
base = getacknum(rcvpkt)+1
If (base == nextseqnum)
stop_timer
else
start_timer
Transport Layer 3-48
GBN: receiver extended
FSM
default
udt_send(sndpkt) rdt_rcv(rcvpkt)
&& notcurrupt(rcvpkt)
 && hasseqnum(rcvpkt,expectedseqnum)
expectedseqnum=1 Wait extract(rcvpkt,data)
sndpkt = deliver_data(data)
make_pkt(expectedseqnum,ACK,chksum) sndpkt = make_pkt(expectedseqnum,ACK,chksum)
udt_send(sndpkt)
expectedseqnum++

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 #
Transport Layer 3-49
 GBN in action
sender window (N=4) sender receiver
012345678 send pkt0
012345678 send pkt1
send pkt2 receive pkt0, send ack0
012345678
send pkt3 Xloss receive pkt1, send ack1
012345678
(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 3-50

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

Transport Layer 3-51

Selective repeat: sender, receiver
windows

Transport Layer 3-52

Selective repeat
sender receiver
data from above: pkt n in [rcvbase,
rcvbase+N-1]
 if next available seq #
in window, send pkt  send ACK(n)
timeout(n):  out-of-order: buffer
 resend pkt n, restart
 in-order: deliver (also
timer deliver buffered, in-
order pkts), advance
ACK(n) in
[sendbase,sendbase+N]: window to next not-
 mark pkt n as yet-received pkt
received pkt n in [rcvbase-
N,rcvbase-1]
 if n smallest unACKed
pkt, advance window  ACK(n)
base to next unACKed otherwise:
seq #  ignore
Transport Layer 3-53
 Selective repeat in action
sender window (N=4) sender receiver
012345678 send pkt0
012345678 send pkt1
send pkt2 receive pkt0, send ack0
012345678
send pkt3 Xloss receive pkt1, send ack1
012345678
(wait)
receive pkt3, buffer,
012345678 rcv ack0, send pkt4 send ack3
012345678 rcv ack1, send pkt5 receive pkt4, buffer,
send ack4
record ack3 arrived receive pkt5, buffer,
send ack5
pkt 2 timeout
012345678 send pkt2
012345678 record ack4 arrived
012345678 rcv pkt2; deliver pkt2,
record ack4 arrived
012345678 pkt3, pkt4, pkt5; send ack2

Q: what happens when ack2 arrives?

Transport Layer 3-54

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

dilemma 0123012 pkt0

pkt1 0123012
0123012
0123012 pkt2 0123012
example: 0123012
0123012 pkt3
 seq #’s: 0, 1, 2, 3 0123012
X
 window size=3 pkt0 will accept packet
with seq number 0
(a) no problem
 receiver sees no
difference in two receiver can’t see sender side.
scenarios! receiver behavior identical in both cases!
something’s (very) wrong!
 duplicate data
accepted as new 0123012 pkt0
in (b) 0123012 pkt1 0123012
0123012 pkt2 0123012
X 0123012
Q: what relationship X
between seq # timeout
retransmit pkt0 X
size and window 0123012 pkt0
will accept packet
size to avoid (b) oops!
with seq number 0
problem in (b)?
Transport Layer 3-55
Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-56
TCP: Overview RFCs: 793,1122,1323,
2018, 2581

 point-to-point:  full duplex data:

 one sender, one  bi-directional data flow
receiver in same connection
 MSS: maximum
 reliable, in-order segment size
byte steam:  connection-oriented:
 no “message  handshaking
boundaries” (exchange of control
msgs) inits sender,
 pipelined: receiver state before
 TCP congestion and data exchange
flow control set  flow controlled:
window size  sender will not
overwhelm receiver

Transport Layer 3-57

 TCP segment structure
32 bits
URG: urgent data counting
(generally not used) source port # dest port #
by bytes
sequence number of data
ACK: ACK #
valid acknowledgement number (not segments!)
head not
PSH: push data now len used
UAP R S F receive window
(generally not used) # bytes
checksum Urg data pointer
rcvr willing
RST, SYN, FIN: to accept
options (variable length)
connection estab
(setup, teardown
commands)
application
Internet data
checksum (variable length)
(as in UDP)

Transport Layer 3-58

TCP seq. numbers, ACKs
outgoing segment from sender
sequence numbers: source port # dest port #
sequence number
 byte stream “number acknowledgement number
” of first byte in rwnd

segment’s data checksum 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

Q: how receiver handles ACKed yet
ACKed
but not usable
yet sent
out-of-order segments (“in-flight
”)
incoming segment to sender
 A: TCP spec doesn’t
source port # dest port #
say, - up to sequence number
implementor acknowledgement number
A rwnd
checksum urg pointer

Transport Layer 3-59

TCP seq. numbers, ACKs
Host A Host B

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

simple telnet scenario

Transport Layer 3-60

TCP round trip time,
timeout
Q: how to set TCP Q: how to estimate
timeout value? RTT?
 longer than RTT
 SampleRTT: measured
time from segment
 but RTT varies transmission until ACK
 too short: receipt
 ignore retransmissions
premature
timeout,
 SampleRTT will vary,
want estimated RTT
unnecessary “smoother”
retransmissions  average several
 too long: slow recent measurements,
reaction to not just current
SampleRTT
segment loss
Transport Layer 3-61
TCP round trip time,
timeout
EstimatedRTT = (1- )*EstimatedRTT + *SampleRTT
 exponential weighted moving average
 influence of past sample decreases
exponentially fast RTT: gaia.cs.umass.edu to fantasia.eurecom.fr
 typical value:  = 0.125 350

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

300
(milliseconds)
RTT

250
RTT (milliseconds)

200

sampleRTT
150

EstimatedRTT

100
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
time Transport Layer 3-62
SampleRTT Estimated RTT
TCP round trip time,
timeout
 timeout interval: EstimatedRTT plus “safety
margin”
 large variation in EstimatedRTT -> larger safety margin
 estimate SampleRTT deviation from EstimatedRTT:
DevRTT = (1-)*DevRTT +
*|SampleRTT-EstimatedRTT|
(typically,  = 0.25)

TimeoutInterval = EstimatedRTT + 4*DevRTT

estimated RTT “safety margin”

Transport Layer 3-63

Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-64
TCP reliable data transfer
 TCP creates rdt
service on top of
IP’s unreliable
service
 pipelined segments let’s initially consider
 cumulative acks simplified TCP
 single sender:
 ignore duplicate acks
retransmission timer
 ignore flow control,
 retransmissions congestion control
triggered by:
 timeout events
 duplicate acks

Transport Layer 3-65

TCP sender events:
data rcvd from app: timeout:
 create segment with  retransmit segment
seq # that caused timeout
 seq # is byte-  restart timer
stream number of ack rcvd:
first data byte in  if ack acknowledges
segment
 start timer if not
previously unacked
segments
already running  update what is
 think of timer as for
known to be ACKed
oldest unacked  start timer if there
segment
 expiration interval: are still unacked
segments
TimeOutInterval

Transport Layer 3-66

TCP sender (simplified)
data received from application above
create segment, seq. #: NextSeqNum
pass segment to IP (i.e., “send”)
NextSeqNum = NextSeqNum + length(data)
if (timer currently not running)
 start timer
NextSeqNum = InitialSeqNum wait
SendBase = InitialSeqNum for
event timeout
retransmit not-yet-acked
segment with
smallest seq. #
ACK received, with ACK field value y start timer

if (y > SendBase) {
SendBase = y
/* SendBase–1: last cumulatively ACKed byte */
if (there are currently not-yet-acked segments)
start timer
else stop timer
} Transport Layer 3-67
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

timeo

timeo
ACK=100
ut

ut
X
ACK=100
ACK=120

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

SendBase=100 bytes of data
SendBase=120
ACK=100
ACK=120

SendBase=120

lost ACK scenario premature timeout

Transport Layer 3-68
TCP: retransmission
scenarios
Host A Host B

Seq=92, 8 bytes of data

Seq=100, 20 bytes of data

ACK=100
timeo

X
ut

ACK=120

Seq=120, 15 bytes of data

cumulative ACK
Transport Layer 3-69
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 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 3-70

TCP fast
retransmit
 time-out period
often relatively TCP fast retransmit
long: if sender receives
 long delay before
resending lost packet
3 ACKs for same
data duplicate
(“triple
 detect lost
segments via ACKs”),
(“triple duplicate
duplicate ACKs. ACKs”), resend
 sender often sends unacked segment
many segments with smallest seq
back-to-back #
 if segment is lost,  likely that unacked
there will likely be
many duplicate ACKs. segment lost, so
don’t wait for
timeout
Transport Layer 3-71
TCP fast
retransmit
Host A Host B

Seq=92, 8 bytes of data

Seq=100, 20 bytes of data
X

ACK=100
timeo

ACK=100
ut

ACK=100
ACK=100
Seq=100, 20 bytes of data

fast retransmit after sender

receipt of triple duplicate ACK
Transport Layer 3-72
Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-73
TCP flow control
application
application may process
remove data from application
TCP socket buffers ….
TCP socket OS
receiver buffers
… slower than
TCP
receiver is TCP
delivering code
(sender is
sending)
IP
flow control code
receiver controls sender,
so sender won’t overflow
receiver’s buffer by from sender
transmitting too much,
receiver protocol stack
too fast

Transport Layer 3-74

TCP flow control
 receiver “advertises”
free buffer space by to application process
including rwnd value in
TCP header of receiver-
to-sender segments RcvBuffer buffered data
 RcvBuffer size set via
socket options (typical rwnd free buffer space
default is 4096 bytes)
 many operating systems
autoadjust RcvBuffer
TCP segment payloads
 sender limits amount of
unacked (“in-flight”)
data to receiver’s rwnd receiver-side buffering
value
 guarantees receive
buffer will not overflow
Transport Layer 3-75
Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-76
Connection Management
before exchanging data, sender/receiver
“handshake”:
 agree to establish connection (each knowing the
other willing to establish connection)
 agree on connection parameters
application application

connection state: connection state:

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

Socket clientSocket = Socket connectionSocket =

newSocket("hostname","port welcomeSocket.accept();
number");
Transport Layer 3-77
Agreeing to establish a
connection
2-way handshake:
Q: will 2-way handshake
always work in network?
 variable delays
Let’s talk  retransmitted messages
ESTAB (e.g. req_conn(x)) due to
OK message loss
ESTAB
 message reordering
 can’t “see” other side

choose x
req_conn(x)
ESTAB
acc_conn(x)
ESTAB

Transport Layer 3-78

 Agreeing to establish a
connection
2-way handshake failure scenarios:

choose x choose x
req_conn(x) req_conn(x)
ESTAB ESTAB
retransmit acc_conn(x) retransmit acc_conn(x)
req_conn( req_conn(
x) x)
ESTAB ESTAB
data(x+1) accept
req_conn(x)
retransmit data(x+1
data(x+1) )
connection connection
client x completes server x completes server
client
terminat forgets x terminat forgets x
es req_conn(x)
es

ESTAB ESTAB
data(x+1) accept
half open connection! data(x+1
(no client!) )
Transport Layer 3-79
 TCP 3-way handshake

client state server state

LISTEN 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 3-80

TCP 3-way
handshake: FSM
closed

Socket connectionSocket =
welcomeSocket.accept();

 Socket clientSocket =
SYN(x) newSocket("hostname","port
number");
SYNACK(seq=y,ACKnum=x+1)
create new socket for listen SYN(seq=x)
communication back to client

SYN SYN
rcvd sent

SYNACK(seq=y,ACKnum=x+1)
ESTAB ACK(ACKnum=y+1)
ACK(ACKnum=y+1)


Transport Layer 3-81

TCP: closing a 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 3-82

TCP: closing a connection
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

Transport Layer 3-83

Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-84
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!
Transport Layer 3-85
Causes/costs of congestion:
scenario 1
original data: in throughput:out
 two senders, two
receivers Host A
 one router, infinite unlimited shared
buffers output link buffers
 output link capacity:
R
 no retransmission
Host B

R/2

delay
out

in R/2 in R/2

 maximum per-  large delays as arrival
connection throughput: rate, in, approaches
R/2 capacity Transport Layer 3-86
Causes/costs of congestion:
scenario 2
 one router, finite buffers
 sender retransmission of timed-out packet
 application-layer input = application-layer output:in
= out
 transport-layer input includes retransmissions :in in
‘

in : original data

'in: original data, plus out
retransmitted data

Host A

finite shared output

Host B
link buffers
Transport Layer 3-87
Causes/costs of congestion:
scenario 2
R/2
idealization: perfect
knowledge

out
 sender sends only
when router buffers
available in R/2

in : original data

copy 'in: original data, plus out
retransmitted data

A free buffer space!

finite shared output

Host B
link buffers
Transport Layer 3-88
Causes/costs of congestion:
scenario 2
Idealization: known
loss packets can be
lost, dropped at router
due to full buffers
 sender only resends if
packet known to be lost

in : original data

copy out
'in: original data, plus
retransmitted data

A no buffer space!

Host B
Transport Layer 3-89
Causes/costs of congestion:
scenario 2
Idealization: known R/2
loss packets can be
lost, dropped at router when sending at R/2,
due to full buffers some packets are

out
retransmissions but
 sender only resends if asymptotic goodput
packet known to be lost is still R/2 (why?)

in R/2

in : original data

out
'in: original data, plus
retransmitted data

A free buffer space!

Host B
Transport Layer 3-90
Causes/costs of congestion:
scenario 2
Realistic: duplicates R/2
 packets can be lost,
dropped at router due when sending at R/2,
some packets are
to full buffers

out
retransmissions
including duplicated
 sender times out that are delivered!
prematurely, sending in R/2
two copies, both of
which are delivered
in
copy
timeo out
ut 'in

A free buffer space!

Host B
Transport Layer 3-91
Causes/costs of congestion:
scenario 2
Realistic: duplicates R/2
 packets can be lost,
dropped at router due when sending at R/2,
some packets are
to full buffers

out
retransmissions
including duplicated
 sender times out that are delivered!
prematurely, sending in R/2
two copies, both of
which are delivered
“costs” of congestion:
 more work (retrans) for given “goodput”
 unneeded retransmissions: link carries multiple
copies of pkt
 decreasing goodput

Transport Layer 3-92

Causes/costs of congestion:
scenario 3
 four senders Q: what happens as in
 multihop paths and in’ increase ?
A: as red in’ increases, all
 timeout/retransmit arriving blue pkts at upper
queue are dropped, blue
Host A
in : original throughput
data out 0
Host B
'in: original data, plus
retransmitted data
finite shared output
link buffers

Host D
Host C

Transport Layer 3-93

Causes/costs of congestion:
scenario 3
C/2
out

in’ C/2

another “cost” of congestion:

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

Transport Layer 3-94

Approaches towards congestion
control
two broad approaches towards congestion
control:
end-end network-assisted
congestion congestion
control: control:
 no explicit  routers provide
feedback from feedback to end
network systems
 congestion  single bit indicating
inferred from end- congestion (SNA,
system observed DECbit, TCP/IP ECN,
loss, delay ATM)
 approach taken by  explicit rate for
TCP sender to send at
Transport Layer 3-95
Case study: ATM ABR congestion
control
ABR: available bit RM (resource
rate: management) cells:
 sent by sender, interspersed
 “elastic service” with data cells
 if sender’s path  bits in RM cell set by
“underloaded”: switches (“network-assisted
 sender should ”)
 NI bit: no increase in rate
use available
(mild congestion)
bandwidth  CI bit: congestion
 if sender’s path indication
congested:  RM cells returned to sender
 sender throttled by receiver, with bits intact
to minimum
guaranteed rate
Transport Layer 3-96
Case study: ATM ABR congestion
control
RM cell data cell

 two-byte ER (explicit rate) field in RM cell

 congested switch may lower ER value in cell
 senders’ send rate thus max supportable rate on path
 EFCI bit in data cells: set to 1 in congested
switch
 if data cell preceding RM cell has EFCI set, receiver
sets CI bit in returned RM cell
Transport Layer 3-97
Chapter 3 outline
3.1 transport-layer 3.5 connection-
services oriented transport:
3.2 multiplexing TCP
 segment structure
and
demultiplexing  reliable data transfer
 flow control
3.3 connectionless  connection
transport: UDP management
3.4 principles of 3.6 principles of
reliable data congestion control
transfer
3.7 TCP congestion
control
Transport Layer 3-98
 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
additively increase window size …
after loss …. until loss occurs (then cut window in half)
congestion window size
cwnd: TCP sender

AIMD saw tooth

behavior: probing
for bandwidth

time
Transport Layer 3-99
TCP Congestion Control:
details
sender sequence number space
cwnd TCP sending rate:
 roughly: send
cwnd bytes, wait
last byte last byte RTT for ACKS,
ACKed sent, not- sent
yet then send more
ACKed
(“in-flight bytes cwnd
 sender limits
”) transmission: rate ~
~ bytes/sec
RTT
LastByteSent- < cwnd
LastByteAcked
 cwnd is dynamic, function of
perceived network congestion

Transport Layer 3-100

TCP Slow Start
Host A Host B
 when connection
begins, increase
rate exponentially one s e gm
ent
until first loss event:

RTT
 initially cwnd = 1 MSS two segm
en ts
 double cwnd every
RTT
 done by incrementing four segm
ents
cwnd for every ACK
received
 summary: initial rate
is slow but ramps up time
exponentially fast
Transport Layer 3-101
 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)

Transport Layer 3-102

TCP: switching from slow start
to CA
Q: when should the
exponential
increase switch to
linear?
A: when cwnd gets to
1/2 of its value
before timeout.

Implementation:
 variable ssthresh
 on loss event,
ssthresh is set to 1/2
of cwnd just before loss
event

Transport Layer 3-103

Summary: TCP Congestion
Control New
New ACK!
duplicate ACK
dupACKcount++
ACK!
new ACK
new ACK
.
cwnd = cwnd + MSS (MSS/cwnd)
dupACKcount = 0
cwnd = cwnd+MSS transmit new segment(s), as allowed
dupACKcount = 0
 transmit new segment(s), as allowed
cwnd = 1 MSS
ssthresh = 64 KB cwnd > ssthresh
dupACKcount = 0
slow  congestion
start timeout avoidance
ssthresh = cwnd/2
cwnd = 1 MSS duplicate ACK
timeout dupACKcount = 0 dupACKcount++
ssthresh = cwnd/2 retransmit missing segment
cwnd = 1 MSS
dupACKcount = 0
retransmit missing segment New
timeout
ACK!
ssthresh = cwnd/2
cwnd = 1 New ACK
dupACKcount = 0
retransmit missing segment cwnd = ssthresh dupACKcount == 3
dupACKcount == 3 dupACKcount = 0
ssthresh= cwnd/2 ssthresh= cwnd/2
cwnd = ssthresh + 3 cwnd = ssthresh + 3
retransmit missing segment retransmit missing segment
fast
recovery
duplicate ACK
cwnd = cwnd + MSS
transmit new segment(s), as allowed

Transport Layer 3-104

TCP throughput
 avg. 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
 avg. window size (# in-flight
3 W bytes) is ¾ W
avg TCP thruput = bytes/sec
 avg. thruput 4
is 3/4W per RTTRTT

W/2

Transport Layer 3-105

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

Transport Layer 3-106

TCP Fairness
fairness goal: if K TCP sessions share
same bottleneck link of bandwidth R,
each should have average rate of R/K

TCP connection 1

bottleneck
router
capacity R
TCP connection 2

Transport Layer 3-107

Why is TCP fair?
two competing sessions:
 additive increase gives slope of 1, as throughout
increases
 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
Transport Layer 3-108
Fairness (more)
Fairness and UDP Fairness, parallel TCP
 multimedia apps connections
 application can open
often do not use
TCP multiple parallel
 do not want rate connections between
throttled by two hosts
congestion control  web browsers do this
 instead use UDP:  e.g., link of rate R with 9
 send audio/video existing connections:
at constant rate,  new app asks for 1 TCP, gets
tolerate packet rate R/10
loss  new app asks for 11 TCPs, gets
R/2

Transport Layer 3-109

Chapter 3: summary
 principles behind transport
layer services:
 multiplexing,
demultiplexing next:
 reliable data transfer  leaving the
 flow control network “edge”
 congestion control (application,
 instantiation, transport layers)
implementation in the  into the network
Internet “core”
 UDP
 TCP

Transport Layer 3-110