COM210: Computer Networks (CN)

Chapter 3: TCP
Dr Ahmad Al-Zubi

King Saud University -

Dr Ahmad Al-Zubi1

TCP: Overview
Point-to-point:
one sender, one receiver

Reliable, in-order byte steam:

no message boundaries
But TCP chops it up into segments for
transmission internally

Pipelined (window) flow control:

Window size decided by receiver and network

Send & receive buffers

King Saud University -

Dr Ahmad Al-Zubi2

TCP: Overview

socket
door

a p p lic a tio n
w r ite s d a ta

a p p lic a t io n
re a d s d a ta

TCP
s e n d b u ffe r

TC P
r e c e iv e b u f f e r

socket
door

segm ent

King Saud University -

Dr Ahmad Al-Zubi3

TCP: Overview
Full duplex data:
bi-directional data flow in same connection
MSS: maximum segment size

Connection-oriented:
handshaking (exchange of control msgs)
inits sender, receiver state before data
exchange

Flow & Congestion Control:

sender will not overwhelm receiver or the
network
King Saud University -

Dr Ahmad Al-Zubi4

TCP segment structure

32 bits
URG: urgent data
(generally not used)
ACK: ACK #
valid

PSH: push data now

(generally not used)
RST, SYN, FIN:
connection estab
(setup, teardown
commands)
Internet
checksum
(as in UDP)

source port #

dest port #

sequence number
acknowledgement number
head not
U AP R S F
len used

checksum

rcvr window size

ptr urgent data

Options (variable length)

counting
by bytes
of data
(not segments!)
# bytes
rcvr willing
to accept

application
data
(variable length)

King Saud University -

Dr Ahmad Al-Zubi5

TCP seq. #s and ACKs (I)

Sequence Numbers:
byte stream number of first byte in segments
data
ACKs:
seq # of next byte expected from other side
cumulative ACK
Q: how receiver handles out-of-order segments
A: TCP spec doesnt say, - up to implementor

King Saud University -

Dr Ahmad Al-Zubi6

TCP Seq. #s and ACKs (II)

Host B

Host A
User
types
C

Seq=4
2, AC

79
Seq=

host ACKs
receipt
of echoed
C

Seq=4

K=79,
d

ata =
C

, data
3
4
=
, AC K

3, ACK

= C

host ACKs
receipt of
C, echoes
back C

=80

simple telnet scenario

King Saud University -

Dr Ahmad Al-Zubi7

Temporal Redundancy Model

Packets
Timeout
Status Reports

Sequence Numbers
CRC or Checksum
ACKs
NAKs,
SACKs
Bitmaps

Retransmissions
Packets
FEC information

King Saud University -

Dr Ahmad Al-Zubi8

Status Report Design

Cumulative acks:
Robust to losses on the reverse channel
Can work with go-back-N retransmission
Cannot pinpoint blocks of data which are
lost
The first lost packet can be pinpointed
because the receiver would generate
duplicate acks

King Saud University -

Dr Ahmad Al-Zubi9

TCP: reliable data transfer (I)

event: data received
from application above
create, send segment

wait
wait
for
for
event
event

event: timer timeout for

segment with seq # y
retransmit segment

one way data transfer

no flow, congestion
control

event: ACK received,

with ACK # y
ACK processing

King Saud University -

Dr Ahmad Al-Zubi
10

TCP:
reliable
data
transfer (II)

00 sendbase = initial_sequence number

01 nextseqnum = initial_sequence number
02
03 loop (forever) {
04
switch(event)
05
event: data received from application above
06
create TCP segment with sequence number nextseqnum
07
start timer for segment nextseqnum
08
pass segment to IP
09
nextseqnum = nextseqnum + length(data)
10
event: timer timeout for segment with sequence number y
11
retransmit segment with sequence number y
12
compute new timeout interval for segment y
13
restart timer for sequence number y
14
event: ACK received, with ACK field value of y
15
if (y > sendbase) { /* cumulative ACK of all data up to y */
16
cancel all timers for segments with sequence numbers < y
17
sendbase = y
18
}
19
else { /* a duplicate ACK for already ACKed segment */
20
increment number of duplicate ACKs received for y
21
if (number of duplicate ACKS received for y == 3) {
22
/* TCP fast retransmit */
23
resend segment with sequence number y
24
restart timer for segment y
25
}
26
} /* end of loop forever */

Simplified
TCP
sender
King Saud University -

Dr Ahmad Al-Zubi
11

TCP ACK generation

Event

TCP Receiver action

in-order segment arrival,

no gaps,
everything else already ACKed

delayed ACK. Wait up to 500ms

for next segment. If no next segment,
send ACK

in-order segment arrival,

no gaps,
one delayed ACK pending

immediately send single

cumulative ACK

out-of-order segment arrival

higher-than-expect seq. #
gap detected!

send duplicate ACK, indicating seq. #

of next expected byte

arrival of segment that

partially or completely fills gap

immediate ACK if segment starts

at lower end of gap

King Saud University -

Dr Ahmad Al-Zubi
12

TCP: retransmission scenarios

Host A

2, 8 b
ytes d
ata
=100
K
C
A

loss
Seq=9
2, 8

bytes
data

lost ACK scenario

Seq=
100,

8 byte
s data
20 by
t es d
at a

0
10
=
K
20
AC CK=1
A

Seq=9

2, 8 b
ytes d
ata
0
=12
K
AC

=100
ACK

time

Host B

Seq=9
2,

Seq=100 timeout
Seq=92 timeout

Seq=9

timeout

Host A

Host B

time

King Saud University -

premature timeout,
cumulative ACKs

Dr Ahmad Al-Zubi
13

TCP Flow Control

flow control
sender wont overrun
receivers buffers by
transmitting too much,
too fast
RcvBuffer = size or TCP Receive Buffer
RcvWindow = amount of spare room in Buffer

receiver: explicitly

informs sender of
free buffer space
RcvWindow field
in TCP segment
sender: keeps the
amount of
transmitted,
unACKed data less
than most recently
received
RcvWindow

receiver buffering

King Saud University -

Dr Ahmad Al-Zubi
14

Timeout and RTT Estimation

Timeout: for robust detection of
packet loss
Problem: How long should timeout
be ?
Too long => underutilization
Too short => wasteful
retransmissions
Solution: adaptive timeout: based on
estimate of max RTT
King Saud University -

Dr Ahmad Al-Zubi
15

How to estimate max RTT?

RTT = prop + queuing delay
Queuing delay highly variable
So, different samples of RTTs will give
different random values of queuing delay

Chebyshevs Theorem:
MaxRTT = Avg RTT + k*Deviation
Error probability is less than 1/(k**2)
Result true for ANY distribution of
samples
King Saud University -

Dr Ahmad Al-Zubi
16

Round Trip Time and Timeout (II)

Q: how to estimate RTT?
SampleRTT: measured time from segment
transmission until ACK receipt
ignore retransmissions, cumulatively
ACKed segments
SampleRTT will vary wildly => want
estimated RTT smoother
use several recent measurements, not
just current SampleRTT to calculate
AverageRTT
King Saud University -

Dr Ahmad Al-Zubi
17

TCP Round Trip Time and

Timeout (III)
AverageRTT = (1-x)*AverageRTT + x*SampleRTT

Exponential weighted moving average (EWMA)

influence of given sample decreases
exponentially fast; x = 0.1

Setting the timeout

AverageRTT plus safety margin proportional to
variation
Timeout = AverageRTT + 4*Deviation
Deviation = (1-x)*Deviation + x*|SampleRTT- AverageRTT|

King Saud University -

Dr Ahmad Al-Zubi
18

TCP Connection Management - 1

Recall: TCP sender, receiver establish connection
before exchanging data segments
initialize TCP variables:
seq. #s
buffers, flow control info (e.g. RcvWindow)

client: connection initiator

Socket clientSocket = new Socket("hostname","port number");

server: contacted by client

Socket connectionSocket = welcomeSocket.accept();

King Saud University -

Dr Ahmad Al-Zubi
19

TCP Connection Management - 2

Three way handshake:

Step 1: client end system sends TCP SYN control

segment to server
specifies initial seq #

Step 2: server end system receives SYN, replies

with SYNACK control segment

ACKs received SYN

allocates buffers
specifies server-> receiver initial seq. #

King Saud University -

Dr Ahmad Al-Zubi
20

TCP Connection Management - 3

Closing a connection:
client closes socket: clientSocket.close();

Step 1: client end system sends TCP

FIN control segment to server

Step 2: server receives FIN, replies with

ACK. Closes connection, sends FIN.
King Saud University -

Dr Ahmad Al-Zubi
21

TCP Connection Management - 4

Fddfdf
close

client

server
FIN

ACK

close

timed wait

FIN

ACK

closed

King Saud University -

Dr Ahmad Al-Zubi
22

TCP Connection Management - 5

Step 3: client receives FIN, replies with ACK.
Enters timed wait - will respond with
ACK to received FINs

Step 4: server, receives ACK.

Connection

closed.

Note: with small modification, can handle

simultaneous FINs.

King Saud University -

Dr Ahmad Al-Zubi
23

TCP Connection Management - 6

TCP client lifecycle

King Saud University -

Dr Ahmad Al-Zubi
24

TCP Connection Management - 7

TCP server lifecycle

King Saud University -

Dr Ahmad Al-Zubi
25

Recap: Stability of a Multiplexed System

Average Input Rate > Average Output Rate
=> system is unstable!

How to ensure stability ?

1. Reserve enough capacity so that
demand is less than reserved capacity
2. Dynamically detect overload and adapt
either the demand or capacity to resolve
overload
King Saud University -

Dr Ahmad Al-Zubi
26

Congestion Problem in Packet Switching

10 Mbs
Ethernet

A
B

statistical multiplexing

1.5 Mbs
queue of packets
waiting for output
link

45 Mbs

Cost: self-descriptive header per-packet,

buffering and delays for applications.

Need to either reserve resources or

dynamically detect/adapt to overload for stability

King Saud University -

Dr Ahmad Al-Zubi
27

The
Congestion Problem

outstrips
available capacity
Problem: demand
1
Demand

Capacity

If information about i , and is

known in a central location where
control of i or can be effected
with zero time delays,
the congestion problem is solved!
King Saud University -

Dr Ahmad Al-Zubi
28

The Congestion Problem

(Continued)
Problems:
Incomplete information (eg: loss
indications)
Distributed solution required
Congestion and control/measurement
locations different
Time-varying, heterogeneous timedelay

King Saud University -

Dr Ahmad Al-Zubi
29

The Congestion Problem

Static fixes may not solve congestion
a) Memory becomes cheap (infinite memory)

No buffer

Too late

b) Links become cheap (high speed links)?

Replace with 1 Mb/s
All links 19.2 kb/s

SS

SS

SS

SS

File Transfer time = 5 mins

SS

SS

SS

SS

File Transfer Time = 7 hours

King Saud University -

Dr Ahmad Al-Zubi
30

The Congestion Problem

(Continued)
c) Processors become cheap
(fast routers & switches)
A
B

C
D

Scenario: All links 1 Gb/s.

A & B send to C
=> high-speed congestion!!
(lose more packets faster!)
King Saud University -

Dr Ahmad Al-Zubi
31

Principles of Congestion Control

Congestion:
informally: too many sources sending too
much data too fast for network to handle
different from flow control (receiver overload)!
manifestations:
lost packets (buffer overflow at routers)
long delays (queuing in router buffers)
a top-10 problem!

King Saud University -

Dr Ahmad Al-Zubi
32

Causes/costs of congestion:
scenario 1
two senders, two
receivers
one router,
infinite buffers
no
retransmission
large delays
when congested
maximum
achievable
throughput

King Saud University -

Dr Ahmad Al-Zubi
33

Causes/costs of congestion:
scenario 2
one router, finite buffers
sender retransmission of lost packet

King Saud University -

Dr Ahmad Al-Zubi
34

Causes/costs of congestion:
scenario 2 (continued)

Costs of congestion:
More work (retrans) for given goodput
Unneeded retransmissions: link carries
multiple copies of pkt due to spurious
timeouts
King Saud University -

Dr Ahmad Al-Zubi
35

Causes/costs of congestion:
scenario 3

Another cost of congestion:

when packet dropped, any upstream
transmission capacity used for that packet
was wasted!
King Saud University -

Dr Ahmad Al-Zubi
36

Approaches towards
congestion control - 1
Two broad approaches towards
congestion control:

End-end congestion control:

no explicit feedback from network
congestion inferred from end-system
observed loss, delay
approach taken by TCP
King Saud University -

Dr Ahmad Al-Zubi
37

Approaches towards
congestion control - 2
Network-assisted congestion
control:
routers provide feedback to end
systems
single bit indicating congestion
(SNA, DECbit, TCP/IP ECN, ATM)
explicit rate sender should send at
King Saud University -

Dr Ahmad Al-Zubi
38

TCP congestion control - 1

end-end control (no network assistance)
transmission rate limited by congestion
window size, Congwin, over segments:

King Saud University -

Dr Ahmad Al-Zubi
39

TCP congestion control - 2

w segments, each with MSS
bytes sent in one RTT:
throughput =

w * MSS
RTT

King Saud University -

Bytes/sec

Dr Ahmad Al-Zubi
40

TCP congestion control - 3

Probing for usable bandwidth:
Window flow control: avoid receiver
overrun
Dynamic window congestion control:
avoid/control network overrun

Policy:
Increase Congwin until loss (congestion)
Loss => decrease Congwin, then begin
probing (increasing) again
King Saud University -

Dr Ahmad Al-Zubi
41

Additive Increase/Multiplicative
Decrease (AIMD) Policy
For stability:
rate-of-decrease > rate-of-increase
Decrease performed enough times as long as congestion exists
AIMD policy satisfies this condition, provided packet loss is
congestion indicator

window

time
King Saud University -

Dr Ahmad Al-Zubi
42

Fairness
Fairness goal: if N TCP sessions
share same bottleneck link, each
should get 1/N of link capacity
TCP connection 1

TCP
connection 2

bottleneck
router
capacity R

King Saud University -

Dr Ahmad Al-Zubi
43

Fairness Analysis

King Saud University -

Dr Ahmad Al-Zubi
44

AIMD Converges to Fairness

King Saud University -

Dr Ahmad Al-Zubi
45

TCP congestion control - 4

TCP uses AIMD policy in steady state
Two phases
Transient phase: aka Slow start
Steady State: aka Congestion avoidance

Important variables:
Congwin
threshold: defines threshold between two
slow start phase, congestion avoidance
phase

King Saud University -

Dr Ahmad Al-Zubi
46

TCP Slowstart - 1
Slowstart algorithm
initialize: Congwin = 1
for (each segment ACKed)
Congwin++
until (loss event OR
CongWin > threshold)

Exponential increase (per RTT) in window

size (not so slow!)
Loss event: timeout (Tahoe TCP) and/or
three duplicate ACKs (Reno TCP)
King Saud University -

Dr Ahmad Al-Zubi
47

TCP Slowstart - 2

RTT

asdf

Host A

Host B
one segm
en

two segm
ents

four segm

ents

time

King Saud University -

Dr Ahmad Al-Zubi
48

TCP Dynamics
1st RTT

2nd RTT

3rd RTT

4th RTT

Rate of acks determines rate of

packets : Self-clocking property.
100 Mbps

10 Mbps

Router

King Saud University -

Dr Ahmad Al-Zubi
49

TCP Congestion Avoidance - 1

Congestion avoidance
/* slowstart is over
*/
/* Congwin > threshold */
Until (loss event) {
every w segments ACKed:
Congwin++
}
threshold = Congwin/2
Congwin = 1
1
perform slowstart
1: TCP Reno skips slowstart (aka fast recovery) after
three duplicate ACKs and performs close to AIMD
King Saud University Dr Ahmad Al-Zubi
50

TCP Congestion Avoidance - 2

King Saud University -

Dr Ahmad Al-Zubi
51

TCP window dynamics (more)

Congestion
Window
(cwnd)

Timeout

Receiver Window
Idle
Interval

ssthresh

1
Time (units of RTTs)

King Saud University -

Dr Ahmad Al-Zubi
52

TCP latency modeling - 1

Q: How long does it take to receive
an object from a Web server after
sending a request?
TCP connection establishment
data transfer delay

King Saud University -

Dr Ahmad Al-Zubi
53

TCP latency modeling - 2

Notation, assumptions:
Assume one link between client and
server of rate R
Assume: fixed congestion window, W
segments
S: MSS (bits)
O: object size (bits)
no retransmissions (no loss, no
corruption)
King Saud University -

Dr Ahmad Al-Zubi
54

TCP latency modeling - 3

Two cases to consider:
WS/R > RTT + S/R: ACK for first
segment in window returns
before windows worth of data
sent
WS/R < RTT + S/R: wait for ACK
after sending windows worth of
data sent
King Saud University -

Dr Ahmad Al-Zubi
55

TCP latency modeling - 4

Case 1: latency = 2RTT + O/R

King Saud University -

Dr Ahmad Al-Zubi
56

TCP latency modeling - 5

K = O/WS

Case 2: latency = 2RTT + O/R

+ (K-1)[S/R + RTT - WS/R]
King Saud University -

Dr Ahmad Al-Zubi
57

TCP latency modeling:

slow start - 1
Now suppose window grows according to
slow start.
Will show that the latency of one object of
size O is:
O
S
S

P
Latency 2 RTT P RTT (2 1)
R
R
R

where P is the number of times TCP stalls at

server:
King Saud University -

Dr Ahmad Al-Zubi
58

TCP latency modeling:

slow start - 2
P min{Q, K 1}
- where Q is the number of times the server
would stall if the object were of infinite
size.
- and K is the number of windows that cover
the object.
King Saud University -

Dr Ahmad Al-Zubi
59

TCP latency modeling:

slow start - 3
in it ia t e T C P
c o n n e c t io n

Example:

re q u e s t
o b je c t

O/S = 15 segments

f ir s t w in d o w
= S /R
RTT

s e c o n d w in d o w
= 2 S /R

K = 4 windows

t h ir d w in d o w
= 4 S /R

Q=2
f o u r t h w in d o w
= 8 S /R

P = min{K-1,Q} = 2
Server stalls P=2 times.

c o m p le t e
t r a n s m is s io n

o b je c t
d e liv e r e d
t im e a t
c lie n t

King Saud University -

t im e a t
s e rv e r

Dr Ahmad Al-Zubi
60

TCP latency modeling:

slow start - 4
S
RTT time from when server starts to send segment
R
until server receives acknowledgement

k 1

S
time to transmit the kth window
R

S
k 1 S
R RTT 2 R

stall time after the kth window

King Saud University -

Dr Ahmad Al-Zubi
61

TCP latency modeling:

slow start - 5
P
O
latency 2 RTT stallTime p
R
p 1
P
O
S
k 1 S
2 RTT [ RTT 2
]
R
R
k 1 R
O
S
S
P
2 RTT P[ RTT ] (2 1)
R
R
R

King Saud University -

Dr Ahmad Al-Zubi
62

Sample Results
R

Minimum Latency:
O/R + 2 RTT

Latency with slow

start

O/R

28 Kbps

28.6 sec

28.8 sec

28.9 sec

100 Kbps

8 sec

8.2 sec

8.4 sec

1 Mbps

800
msec

1 sec

1.5 sec

10 Mbps

80 msec

0.28 sec

0.98 sec

King Saud University -

Dr Ahmad Al-Zubi
63

Summary: Chapter 3

Principles behind transport layer services:

multiplexing/demultiplexing
reliable data transfer
flow control
congestion control
Instantiation and implementation in the
Internet
UDP, TCP
King Saud University -

Dr Ahmad Al-Zubi
64