Content uploaded by Bhargavi Goswami
Author content
All content in this area was uploaded by Bhargavi Goswami on Nov 25, 2017
Content may be subject to copyright.
Proceedings of International Conference on Computer Science Networks and Information Technology
Held on 23rd-24th January 2016, in Pattaya, ISBN: 9788193137338
50
PERFORMANCE EVALUATION OF WIDELY IMPLEMENTED
CONGESTION CONTROL ALGORITHMS OVER DIVERSIFIED
NETWORKING SITUATIONS
Bhargavi Goswami
MCA Department, Sunshine
Group of Institutions,
Rajkot, Gujarat, India
Saleh Asadollahi
MSc.IT and MCA Department,
Saurashtra University,
Rajkot, Gujarat, India
Ian Asare
MSc.IT and MCA Department,
Saurashtra University,
Rajkot, Gujarat, India
Abstract— Gone are the days when the network just
used to be either wired or wireless. In today’s era of
Internet of Things, every single machine is connected
to Internet. Each device is having wireless
connectivity to the Access Point to get connected to
the world of Internet which, may be Fiber or
Microwave signals further, reaching to the end host
by means of again wireless Access Point. The fragile
packet when enters the complex network of Internet,
face brutal blockage, hindering its timely arrival on
the destination. Multiple researchers have been
working on this problem for the last a decades and
have achieved success in improving the situation by
means of one or the other congestion control
algorithms. But, when the demand of the Internet is
now not limited to just the computers and smart
phones, increasing its proximity including every
single device we use, how efficient and robust these
algorithms are required to be gauged. In this paper
we have evaluated the performance of existing stellar
congestion control algorithms in the diversified
heterogeneous networking circumstances. This paper
is an effort towards demonstrating how the existing
mechanisms fit the current requirement and fix the
problem and to what extent.
Keywords— Cubic TCP, XCP, RCP, RCP+, RCP++,
Congestion Control, NS2, Heterogeneous Network,
Wireless Network, Throughput.
I. INTRODUCTION
Widely implemented topology in Internet is
Heterogeneous networks which are a combination of
wired and wireless networks. Due to high bandwidth
availability at wired links and high delayed products
at wireless links, bottleneck situation is generated at
Access Point and it has to face severe Congestion
consequences. Some of the most known and recent
protocols developed to provide faster and lighter
congestion control are TCP (Transmission Control
Protocol), Rate Control Protocol (RCP), Explicit
Congestion Control Protocol (XCP), RCP Plus (RCP+)
and RCP Plus Plus (RCP++). Whenever, traffic exceeds
the network capacity, congestion is likely to take
place. Consider the networking situation from ISP’s
angle. Customers may create bottleneck condition
when all the internet users start sending data at their full
speed. In metro cities, much of the corporate work is
depended on Internet especially during crowded office
hours. ISP’s networks, when combined for the entire
city, makes it a large crowded network, where end users
are generally wireless making it even a larger scale
wireless network. The intriguing part of data network is
that, the traffic is bursty and in most situations,
bursty traffics are unpredictable and random by
behavior. The irregular pattern of traffic makes the
situation even more difficult to control and
networking processes induce complexity into the
computation as well. In such a situation of traffic
bursts under large wireless networks, it is
advantageous that nodes are able to buildup
connectivity with any node of its choice being wireless
by nature and traffic patterns can be settled tactfully in
favor of congestion control.
This raise question that whether the existing
congestion control mechanisms are enough to handle the
current communication demands and does it fit the
current technology scenario? And therefore, it is
desirable to survey upon congestion control mechanisms
available and already deployed and how much are they
capable of handling the current situations and demands.
This paper is a small effort in this direction which helps
the researchers to know the mechanisms evolved to
control congestion, how they benefit the current
communication scenario and how much still they are
lagging behind of fulfilling demands of current internet.
In this paper we have taken most adapted
congestion control protocols that are efficient to
certain extent and weak at certain parameters.
The paper is categorized in following sections. First of
all, we introduce problem statement in Introduction. In
second section, we provide short description of every
single congestion control algorithm taken into
consideration for performance evaluation. Section three
is all about the methodology used and further followed
Proceedings of International Conference on Computer Science Networks and Information Technology
Held on 23rd-24th January 2016, in Pattaya, ISBN: 9788193137338
51
by experimental configurations and set up details
provided with explanation on what is used and why.
Section four is all about obtained results and detailed
analysis. Section five is all about conclusion and further
followed by future aspects, acknowledgement and
references before we end.
II. EXISTING CONGESTION CONTROL ALGORITHMS:
In this section we are going to discuss the basic
behavior of esteemed congestion control algorithms. We
start with TCP, XCP, RCP, RCP+ and RCP++.
1) TCP:
TCP is the protocol widely accepted and
implemented in the network world in all the
diversified network situations, compatible with each
networking device and each network mechanism[1].
Researchers prefer experimenting with TCP as it is the
only protocol in real world with all the simulators
support and easy equipments configuration for model
based or real time experimentation testing. But, few
problems are observed in TCP protocol. While the
global network access speeds increased dramatically
on average in the past decade, the standard value of
TCP’s initial congestion window has remained
unchanged. In spite of avoiding congestion in the
early stages on the Internet, TCP is finding it
increasingly difficult to cope with the growth of
communication network capacities and applications.
TCP’s inability to properly utilize network links is
one of the problems of TCP [6]. Any delay or packet
loss can cause TCP to dramatically reduce its
congestion window and hence under-utilize the link
[6]. There have be many improvements proposed for
TCP to change its behaviour for slow start and AIMD
behaviour like High Speed TCP, Scalable TCP,
Cubic TCP, etc.
2) XCP:
The fundamental algorithm of XCP has been
developed and published by Dina Katabi of MIT and
Mark Handley of ICIR. XCP is a feedback-based
congestion control system that uses direct, explicit,
router feedback to avoid congestion in the
network.Explicit Congestion Control Protocol is another
famous solution to congestion that extracts congestion
information directly from routers and achieves fairness,
maximum link utilization and does efficient use of
bandwidth, which is by nature scalable too. XCP
maintains per flow congestion state in packet. This
needs changes in all the routers and this require large
queuing capacity in routers which is difficult to deploy.
Simulations indicate that XCP is powerful and
scalable. XCP delivers the highest possible
application performance over a broad range of
network infrastructure, including extremely high
speed and very high delay links that are not well
served by TCP[5]. By doing so, it achieves
maximum link utilizations and wastes no bandwidth due
to packet loss as proved by the authors of [5]. XCP
is novel in separating the efficiency and fairness
policies of congestion control, enabling routers to
quickly make use of available bandwidth while
conservatively managing the allocation of bandwidth to
flows[5].
3) RCP:
Nandita Dukkipatti proposed RCP (Rate Control
Protocol) that enables typical large Internet Sized
flows to complete in one to two orders of magnitude
faster than the existing (TCP SACK) and XCP
congestion control algorithms. In RCP, the function of
providing same rate to all the flows based on the
above stated two congestion control mechanisms is
done by Router. And to complete the flows faster, RCP
provides with extreme quick computation the excessive
bandwidth to the flows. This is the reason behind faster
completion of flows by RCP. It is observed through
experimentation that proposed RCP is reducing Flow
Completion Time (FCT) for typical Internet Size
flows by one to two orders of magnitude which in
comparison of TCP and XCP are performing
remarkable. Talking about the mechanism used by
RCP, an improved congestion control mechanism,
RCP achieves it by simply emulating processor
sharing at each router.
4) RCP+:
Research team of Dr. Bhargavi Goswami, Dr. Atul
Gosai and Uditnarayan Kar came up with a congestion
control mechanism and named it RCP+. RCP+ wins
upon existing congestion control algorithm with three
characteristics, 1) RCP+ is more flexible to be
implemented on wireless networks in comparison of
RCP. 2) It allows multiple variants of TCP to co-
exist in the same networks with RCP+ and work with
each other smoothly. 3) And it performs well in large
scale wireless networking scenario too. RCP+ is a
congestion window based congestion control algorithm
where the base equation is similar to RCP and
implementation has far more benefits to the world of
networks in co-existence with RCP and TCP. RCP+ is
most suitable for networking condition where, multiple
flows are having diverse characteristics and their
demands vary frequently. RCP+ is having the added
advantage of coexistence with other wired and
wireless TCP, XCP, RCP and DCCP protocols.
RCP+ is flexible like TCP and so is expected to
have wide implementation over current demands of
Internet. RCP+ approach is quite simple in
comparison of both, RCP and TCP. Its behaviour is a
pure combination of RCP and TCP.
5) RCP++:
Research team of Ishani Mehta, Uditnarayan Kar and
Dr. Atul Gosai came up with an improved version of
RCP+ and named it RCP++. RCP++ is based on
Improved AIMD and RCP+ algorithm. They use
congestion window mechanism of improved AIMD
algorithm to use the spare capacity of congestion
window after occurrence of congestion event. Here
they used Improved AIMD mechanism as well as
Proceedings of International Conference on Computer Science Networks and Information Technology
Held on 23rd-24th January 2016, in Pattaya, ISBN: 9788193137338
52
modified rate change equation of RCP+. The only
limitation of this protocol observed on simulation
environment is that it works for smaller no of
nodes. If the number of nodes increases, its
performance degrades to great extent. To make
RCP++ work for more no of nodes, further
improvements are necessary[4,7].
III. EXPERIMENTAL SETUP
We have implemented proposed model in NS-2
simulator and version is 2.35. Following given Table-
1 is the list of configuration settings of our modeling
with respect to topology specified by our
experiment. The scenario consist s of network of
wireless nodes starting from 10 to 50 nodes. The flows
present in the network are variable by nature, few of
them are large and few of them are small.
Figure. 1 Overview of Networking Scenario.
We have built the simulation model with few
fixed parameters stated in the Table 1 with constant
values; whereas, few are varying for the entire
experiment , for example, number of nodes. The access
point is presumed to be the gateway/router. The
Packet Size is fixed to 1500 bytes for larger flows and
smaller flows are given the size up to 1000 bytes starting
from 100 bytes.
TABLE 1: CONFIGURATION TABLE
Application
FTP
FTP over TCP Agent
and RCP+
Configuration
Network Area
1000 nodes
No. Of Nodes
1 to 50, Incremental
Queue Length
Varying from 50
Queue Type
Drop -tail/Priqueue and
SFQ
Network
Type
Uni-cast
Sources
9<Nodes<51
Uni-cast
Destination
9<Nodes<51
Intermediate
Nodes
Nodes may work as
AP(Access Point) or
End Host
Mobility
Delay
10m/s
Maximum
Speed
54mbps as per
specification of
802.11g respectively.
Pause Time
10 s
Routing
Protocol
DSDV
M AC
IEEE 802.11g
54 mbps
PHY
Propagation
Model
Two Ray Ground
Antenna
Omni Antenna
Figure 2: Simulation of 50 nodes transmitting while moving.
Proceedings of International Conference on Computer Science Networks and Information Technology
Held on 23rd-24th January 2016, in Pattaya, ISBN: 9788193137338
53
Graph 1: Throughput of network on variable nodes from 10 to 50 for
each congestion control protocol.
IV. RESULT ANALYSIS:
We present the results for TCP, XCP, RCP, RCP+
and RCP++ as incremental number of nodes v/s
throughput plots. Here, x-axis is number of nodes and
y -axis is throughput. Graph 1 represents the overall
throughput observed with respect to number of
nodes starting from 10 nodes to 50 nodes. It was
observed that RCP+ is well defined networking
algorithm that reaches to its optimum data rate in small
duration of time in comparison of all other existing
congestion control algorithms. Analysis indicates that
the algorithm is fair initially and tries to stay stable
even when the throughput is degrading. Thus, for
medium range of networks, RCP+ is behaving stable and
performs to its optimum for longer duration.
V. CONCLUSION
In this paper, we have evaluated the performance of
the entire list of prominent congestion control algorithm
in today’s era having impact widely upon the world of
internet. We have evaluated performance of TCP, XCP,
RCP, RCP+ and RCP++. The matrix of evaluation was
Throughput. It was observed during the experimentation
that the behaviour of all the nodes working on
RCP+ and other protocols was affected by their basic
characteristics. Routing protocols affect performance
of networks because it works well under AODV
protocol in comparison of DSDV protocol. It is also
affected somewhat by mobility and synchronization
between moving nodes as there is freedom to get
connected to any node in the wireless range in absence of
guided media. RCP+ wins upon existing congestion
control algorithm with three characteristics, 1) RCP+
is more flexible to be implemented on wireless
networks in comparison of RCP. 2) It allows multiple
variants of TCP co-exist in the same networks with
RCP+ and work with each other smoothly. 3) And
it performs well in large scale wireless networking
scenario too. Our results, and many others in the
literature, indicate that there is no existing single
congestion-control method that is the best in all
situations. Then too, by presenting our work, we
would like to show the direction to networking research
community that brings the era where the congestion
control is prolific in wireless and heterogeneous
networks which is social enough to co-exist with all
other congestion control protocols.
VI. REFERENCES
[1] Dr. Atul Gonsai, Bhargavi Goswami & Uditnarayan Kar,
“Newly developed Algorithm RCP+ for Congestion Control
on large scale Wireless Networks”, International Journal of
Innovations & Advancement in Computer Science IJIACS
ISSN 2347 – 8616 Volume 3, Issue 2 Apri 2014
[2] Hardik Gohel,” Introduction to Network & Cyber Security”,
LAP Lambert Academic Publishing, 2015
[3] Dr. Atul Gonsai, Bhargavi Goswami, Uditnarayan Kar,
“Design of Congestion Control Protocol for Wireless
Networks with Small Flow Completion Time” Proceedings
on National Conference on Emerging Trends in Information &
Communication Technology. 2013
[4] Dr. Atul Gonsai, Bhargavi Goswami, Uditnarayan Kar,
“Evolution of Congestion Control Mechanisms for TCP and
Non TCP Protocols” Matrix Academic International Online
Journal Of Engineering And Technology © MAIOJET. 2013
[5] Hardik A Gohel,” Cyber Security and Social Media”, CSI
Communications – Knowledge Digest for IT Community,
39(5), 2015
[6] Dr. AtulGonsai, BhargaviGoswami, “Network Simulator for
Efficient Performance Parameter Testing & Evaluation” in
National Journal of System and Information Techno logy -
2012, volume 5 Issue 1, pg. 89 -105
[7] Dr. Atul Gonsai, Bhargavi Goswami, “Experimental
Performance Testing of T CP and UDP Protocol over
WLAN 802.11b and802.11g” Karpagam Journal of Computer
Science, Volume 07, Issue 03, March April2013, pg. no. 168 t o
183
[8] Hardik Gohel, “Looking Back at the Evolution of the Internet”,
CSI Communications-Knowledge Digest for IT Community, 38
(6), 2014
[9] Dr. Atul Gonsai, Udit narayan Kar, Bhargavi Goswami,
"Experiment al based performance Analysis of IEEE
802.11/g Hybrid Net work" in Journal of Network and
Information Security(JNIS), Volume 1 Issue 1 Pg. No. 61 -66,
October 2013
[10] Hardik Gohel, Forum Kalyani, "Li-Fi Technology–A survey on
Current IT Trends", International Journal on Advances in
Engineering Technology and Science Volume: 1, Issue: 2,
December 2015
[11] Dr. Bhargavi Goswami, Saleh Asadollahi, “Enhancement Of
LAN Infrastructure Performance For Data Center In Presence
Of Network Security” in proceedings of CSI-2015, Springer,
December 2015.
0
4000
8000
12000
1 2 3 4 5
Nodes#
TCP
XCP
RCP
RCP+
RCP++
Throughput
TCP
RCP+