A Survey on Congestion Control & Avoidance

raggedsquadΔίκτυα και Επικοινωνίες

30 Οκτ 2013 (πριν από 3 χρόνια και 9 μήνες)

121 εμφανίσεις

Available ONLINE
www.vsrdjournals.com





VSRD
-
IJCSIT, Vol. 2 (9
), 201
2
,
1
-
5



____________________________

1
,2
Research Scholar,
Singhania University,
Rajsthan
, INDIA. *Correspondence :
*****


R
R
R
E
E
E
S
S
S
E
E
E
A
A
A
R
R
R
C
C
C
H
H
H



C
C
C
O
O
O
M
M
M
M
M
M
U
U
U
N
N
N
I
I
I
C
C
C
A
A
A
T
T
T
I
I
I
O
O
O
N
N
N



A Survey on Congestion Control & Avoidance

1
Sapna Gupta
*

and

2
Nitin Kumar Sharma

ABSTRACT

C
ongestion has been considered as one of the basic important issue in
packet switched network. Congestion
control refers to the mechanisms and techniques to control the congestion and keep the load below the capacity.

This paper provides an overview of category provided by congestion control. It also includes how TCP uses
co
ngestion control to avoid congestion or alleviate congestion in network. Computer networks have experienced
an explosive growth over the past few years and with that growth have come severe congestion problems.


The paper also concentrates on avoidance of
congestion. This scheme allows a network to operate in the region
of low delay and high throughput. Finally, it gives a summary of Congestion control & avoidance that’s
provides a good foundation for understanding

Keywords :

congestion control, congestion
avoidance.

1.

INTRODUCTION

Congestion in a network may occur if the load on the network
-
the number of packets sent to the network
-
is
grater than the capacity of the network
-
the number of packets a network can handle.

Network congestion

occurs when a link or
node is carrying so much data that its

quality of service

deteriorates.
Typical effects include

queuing,

packet loss

or the

blocking

of new connections. Congestion control is a method
used for monitoring the process of regulating the total amount of data ent
ering the network .so as to keep traffic
levels at an acceptable value. This is done in order to avoid the telecommunication network reaching what is
termed

congest
ive collapse
.

Modern networks use

congestion control

and

net
work congestion avoidance

techniques to try to avoid
congestion collapse. These include:

exponential back off

in protocols such as

802.11
's

CSMA/CA

and the
original

Ethernet
,

window

reduction in

TCP
, and

fair queuing

in devices such as

routers

Congestion
avoidance
Sapna gupta
et al

/ VSRD
International Journal of CS &
IT Vol. 2 (9), 2012


Page
2

of 9


techniques monitor network traffic loads in an effort to anticipate and avoid congestion at common network
bottlenecks.

2.

CONCEPT

In this section, we define the basic concept of congestion control and congestion

avoidance.

These concepts are
relate
d to each other but distinct. They are related because all these solve the problem of resource management
in the network. They are distinct because they solve resource problem either in different parts of or in a different
manner.

2.1.

Congestion Control

Congestion control mostly applies to packet
-
switching network. A wide variety of approaches have been
proposed, however the "objective is to maintain the number of packets within the network below the level at
which performance falls off dramatically."

The
re are two transport layer protocols where congestion control is implemented.



Transmission Control Protocol



User Datagram Protocol

Congestion control mechanisms are divided into two broad categories:
-



Open

loop congestion control

(
Prevention)



Closed
-
loop
congestion control

(removal)

2.1.1.

Open loop congestion control

This policy is applied to prevent congestion before it happens. In this mechanism, congestion control is handled
by either the source or the destination.

A.

Prevention: Different policies at various l
ayers can affect congestion, and these are summarised in the
table.

B.

Congestion prevention tries to design these Policies carefully to minimise congestion

in the first place.

Layer

Policies

TRANSPORT



Retransmission Policy



Out
-
of
-
order catching policy



Acknowledgment policy



Flow control policy

Sapna gupta
et al

/ VSRD
International Journal of CS &
IT Vol. 2 (9), 2012


Page
3

of 9


NETWORK



packet queuing service policy



packet discard policy



routing algorithm



packet
-
lifetime management

DATA LINK



Retransmission Policy



Out
-
of
-
order catching policy



Acknowledgment

Policy




Flow control policy

Table: Congestion Prevention Policies
.


2.1.1.1.

Transport &
D
ata
L
ink
L
ayer policies

A.)
Retransmission Policy
:
-

If sender feels that a sent packet is lost or corrupted, the packet needs to be
retransmitted. A sender that’s times out quickly and retransmit all packets using go back
-
N will put load on the
system then the sender uses selective repeat.

Retransmission i
n general may increase congestion in the network. However, a good retransmission policy can
prevent congestion. It deals with how fast a sender times out and what it transmit upon timeout.

B.)
Out
-
of
-
order catching policy:
-

If the receiver discards all ou
t
-
of
-
order packets, these packets will have to
be transmitted again later, creating extra load
.

C.)
Acknowledgment

Policy
:
-
This

Policy imposed by the receiver may also affect congestion. If the receiver
does not acknowledge every packet it receives, it may

slow down the sender and help prevent congestion. A
receiver may send an acknowledgment only if it has a packet to be sent or a special timer expires.

If each packet is acknowledged immediately, the

acknowledgment packet generate
extra load. However, if
a
cknowledgments are saved up to piggyback onto reserve traffic, extra time out and retransmissions may result.

D.) Flow control policy

Flow control is a scheme for the control of the data flow with in an OSI layer .In other words, its limit an
amount of dat
a transmitted by the send
ing transport entity to a level
, or rate,

that the receiver can manage. It
involves a direct feedback from receiver to sender.

2.1.1.2.

Network Layer Policy

A.)
Packet Queuing Service policy
:

-

Router may have one queue per input line, one

queue per output line or
both. It also relates to the order packets are processed.

B.)
Packet Discarding Policy
:
-
A good discarding policy by the routers may prevent congestion and at the
Sapna gupta
et al

/ VSRD
International Journal of CS &
IT Vol. 2 (9), 2012


Page
4

of 9


same time may not harm the integrity of the transmission .It tell w
hich a packet is drop when there is no place.

C.)
Routing Algorithm
:

-

In
routing algorithms, Destination / next hop associations tell a router that a
particular destination can be gained optimally by sending the packet to a particular router representing
the "next
hop" on the way to the final destination. When a router receives an incoming packet, it checks the destination
address and attempts to associate this address with a next hop.


D.)
Packet life time management
:

-

It deals with how long a packet ma
y

live before being discarded
If it is too
long,

lost packet waste the network bandwidth. If it is too short, packet may be discarded before reaching their
destination.

2.1.2.

Closed loop congestion control

This mechanism tries to alleviate congestion after it
happens. It monitor the system to detect congestion, pass
this information to where action can be taken, and adjust system operation to correct the problem (detect,
feedback and correct). There are provided some mechanisms have been used by different proto
col.

A.)
Backpressure:

-

The technique refers to a congestion control mechanism in which a congested node stops
receiving data from the immediate upstream node or nodes. This may cause the upstream node or nodes to
become congested and they, in turn, rejec
t data from their upstream nodes. This is a node
-
to
-
node congestion
control that starts with a node and propagates, in the opposite direction of data flow, to the source. The
backpressure technique can be applied only to virtual circuit networks, in which
each node knows the upstream
node from which a flow of data is coming.

B.)
Choke Packet
:
-

A Choke packet is a packet sent by a node to the source to inform it of congestion. There is
a difference between the backpressure and choke packet methods. In backp
ressure, the warning is from one
node to its upstream node, although the warning may eventually reach the source station. In choke packet
method, the warning is from the router, which has encountered congestion, to the source station directly. The
intermed
iate nodes through which the packet has travelled are not warned.

C.)
Implicit Signalling
:
-

In implicit signalling, there is no communication between the congested node or
nodes and the source. The source guesses that there is congestion somewhere in the

network from other
symptoms. For example, when a source sends several packets and there is no acknowledgment for a while, one
assumption is that the network is congested. The delay in receiving an acknowledgment is interpreted as
congestion in the network
; the source should slow down
.

D.)
Explicit Signalling
:

-

The node that experiences congestion can explicitly send a signal to the source or
destination .The explicitly signalling method is different from the choke packet method. In the choke packet
method
, a separate packet is used for this purpose; in the explicit

signalling method, the signal is included in the
packets that carry data.


2.2.

Congestion A
voidance

Congestion avoidance techniques monitor network traffic loads in an effort to anticipate and avoid congestion at
Sapna gupta
et al

/ VSRD
International Journal of CS &
IT Vol. 2 (9), 2012


Page
5

of 9


common network bottlenecks. Congestion avoidance is achieved through packet dropping. Among the more
commonly used congestion avoidance mechani
sms is Random Early Detection (RED), which is optimum for
high
-
speed transit networks. Cisco IOS QoS includes an implementation of RED that, when configured, controls
when the router drops packets. If you do not configure Weighted Random Early Detection (W
RED), the router
uses the cruder default packet drop mechanism called tail drop.

Router uses the cruder default packet drop mechanism called tail drop.

There is some description
of the kinds of
congestion avoidance mechanisms provided by the Cisco

IOS QoS
features. It discusses the following features:

2.2.1.

Trail Drop


Tail drop treats all traffic equally and does not differentiate between classes of service. Queues fill during
periods of congestion. When the output queue is full and tail drop is in effect, packe
ts are dropped until the
congestion is eliminated and the queue is no longer full.

Weighted Random Early Detection

WRED combines the capabilities of the RED algorithm with the IP
Precedence feature to provide for preferential traffic handling of higher pri
ority packets.

WRED avoids the globalization problems that occur when tail drop is used as the congestion avoidance
mechanism on the router. Global synchronization occurs as waves of congestion crest only to be followed by
troughs during which the transmis
sion link is not fully utilized.

In
WRED, the following related features are discussed:
-



Flow
-
based WRED
:
-

Flow
-
based WRED is a feature that forces WRED to afford greater fairness to
all flows on an interface in regard to how packets are dropped.



Uses of

Flow based WRED
:
-

it helps to think about how WRED (without flow
-
based WRED
configured) affects different kinds of packet flows. Even before flow
-
based WRED classifies packet
flows, flows can be thought of as belonging to one of the following categories

1.

No adaptive flows, which are flows that do not respond to congestion.

2.

Robust flows, which on average have a uniform data rate and slow down in response to congestion.

3.

Fragile flows, which, though congestion
-
aware, have fewer packets buffered at a gateway t
han do
robust flows
.



DiffServ Compliant WRED
:
-
. DiffServ Compliant WRED extends WRED to support Differentiated
Services (DiffServ) and Assured Forwarding (AF) Per Hop Behaviour (PHB). This feature enables
customers to implement AF PHB by colouring packets
according to differentiated services code point
(DSCP) values and then assigning preferential drop probabilities to those packets.

2.3.

Congestion Avoidance Vs. Congestion Control

The distinction between congestion control and congestion avoidance is similar to

that between deadlock
Sapna gupta
et al

/ VSRD
International Journal of CS &
IT Vol. 2 (9), 2012


Page
6

of 9


recovery and deadlock avoidance. Congestion control procedures are cures and the avoidance procedures are
preventive in nature. A congestion control scheme tries to bring the network back to an operating state, while a
congestion avo
idance scheme tries to keep the network at an optimal state. Without congestion control a
network may cease operating whereas networks have been operating without congestion avoidance for a long
time. The point at which a congestion control scheme is calle
d upon depends on the amount of memory
available in the routers, whereas, the point at which a congestion avoidance scheme is invoked is independent of
the memory size. A congestion avoidance scheme may continuously oscillate slightly around its goal witho
ut
significant degradation in performance, whereas, congestion control scheme tries to minimize the chances of
going above the limit.


Fig.

1

:

The rate of control and

Feedback delays are related.

2.4.

Fundamental Principle of Control

Congestion control and
congestion avoidance are

dynamic system control issues.
All other control schemes they
consist of two parts: a feedback mechanism and a control mechanism. The feedback mechanism allows the
system to inform the users of the current state of the system. The
control mechanism allows the users to adjust
their load on the system.

As shown in above figure, if the control is faster than the feedback, the system will have oscillations and
instability. On the other hand, if the control is slower than the feedback,
the system will be tardy and slow to
respond to changes. In designing

congestion schemes it is important to apply this principle and to carefully se
-
lect the control interval.

A number of feedback mechanisms have been proposed. If we extend those mechanis
ms to signal operations
around the knee rather than the cli
ff

we obtain a congestion avoidance scheme. Of course, the control
mechanism will also have to be adjusted to help the network operate around the knee rather than the cli
ff

for the
feedback mechani
sms we have the following alternatives:

1.

Congestion feedback via packets sent from routers to sources.

2.

Feedback included in the routing messages exchanged among routers.

3.

End
-
to
-
end probe packets sent by sources.

4.

Each packet contains a congestion feedback

f
ie
ld that is
fi
lled

by routers in packets going in the reverse
Sapna gupta
et al

/ VSRD
International Journal of CS &
IT Vol. 2 (9), 2012


Page
7

of 9


direction.

5.

A congestion feedback field is filled by routers in packets going in the forward direction.

The 1
st

alternative is popularly known as choke packet or source quench message in ARP
ANET. It requires
introducing additional tra

c in the network during congestion, which may not be desirable
.

The second alternative, increasing the cost of congested path, has been tried before in ARPANET delay
sensitive routing.

The third alternative, pro
be packets from the disadvantage of added overhead unless probe packets had a dual
role of carrying other information in them. There would be no reason not to use every packet going through the
network as a probe packet. We may achieve this by reserving a
fi
eld in the packet that is used by the network to
signal congestion. This leads us to the last two alternatives.

The fourth alternative, reverse feedback requires routers to piggyback the signal on the packets going in the
direction opposite the congestion. This alternative has the advantage in that the feedback reaches the source
faster. However, the forward and rev
erse tra

c are not always related

The
f
if
th alternative, forward feedback sends the signal in the packets going in the forward direction (direction
of congestion). The destination either asks

the source to adjust the load or

returns the signal back to the
source
in the packets (or acknowledgments) going in the reverse direction
.

The congestion avoidance algorithms and concepts can be easily modi
fi
ed for other forms of
fl
ow control such
as rate
-
based
fl
ow control in which the sources must send below a rate s
peci
ali
zed by the destination. In this
case, the users would adjust rates based on the signals received from the network
.

3.

CONCLUSION

The paper presents the study of congestion control and elaborates various issues related with it. As the
congestion control

is the most important factor of any packet switching network, the whole performance and
accuracy of network is directly related to it, the congestion control becomes more important.

The proposed study helps while implementing any packet switching network
and reflects a clear picture of
differences among various methods of congestion avoidance and congestion control.

4.

REFERENCES

[1]

http://en.wikibooks.org/wiki/Computer_Networks/Congestion_Control

[2]

http://en.wikipedia.org/wiki/Network_congestion

[3]

D. Bansal, H.
Balakrishnan, and S. S. S. Floyd. Dynamic behavior of slowly
-
responsive congestion control
algorithms
.

[4]

Behrouz A Forouzan . Data communications and networking , page
-
765
-
768.

[5]

V. Jacobson. Congestion avoidance and control. ACM Computer Communication Review;

Proceedings of
Sapna gupta
et al

/ VSRD
International Journal of CS &
IT Vol. 2 (9), 2012


Page
8

of 9


the

Sigcomm ’88

Symposium, 18, 4:314
-
329, Aug. 1988.

[6]

B. T. Doshi and H. Q. Nguyen, "Congestion Con
trol in ISDN Frame
-
Relay Networks," AT&T Tech
nical

Journal, November/December 1988, pp. 35
-
46.

[7]

R. Jain and K. K. Ramakrishnan, "Congestion A
voidance

in

Computer Networks

with a Con
nectionless
Network Layer: Concepts, Goals, and Methodology," Proc. IEEE Computer
.
Networking Symposium,
Washington, D. C, April 1988.

[8]

Jain, R. Ramakrishna, and Chiu, "Congestion avoidance in computer networks with
connectionless network
layer." Digital Equipment Corporation DEC
-
TR
-
506, 1987.

[9]

W. Stevens, "TCP Slow Start, Congestion Avoidance, Fast Retransmit, and Fast Recovery Algorithms."
IETF RFC2001, 1997.

