SlideShare a Scribd company logo
1 of 7
Download to read offline
IOSR Journal of Computer Engineering (IOSRJCE)
ISSN: 2278-0661 Volume 3, Issue 1 (July-Aug. 2012), PP 13-19
www.iosrjournals.org
www.iosrjournals.org 13 | Page
Performance Evaluation of High Speed Congestion Control
Protocols
Mohamed Ali Mani and Rachid Mbarek
ABSTRACT: Computer networks are facing several technological and scientific challenges. Among these
challenges is congestion control. It limits the quantity of information input at a rate less important than that of
the transmission one to ensure good performance as well as protect against overload and blocking of the
network. Researchers have done a great deal of work on improving congestion control protocols, especially on
high speed networks.
In this paper, we will be studying the congestion control alongside low and high speed congestion control
protocols. We will be also simulating, evaluating, and comparing seven of high speed congestion control
protocols : Bic TCP, Cubic TCP, Hamilton TCP, HighSpeed TCP, Illinois TCP, Scalable TCP and YeAH TCP,
with multiple flows.
KEYWORDS: HIGH-SPEED, TCP PROTOCOLS, CONGESTION CONTROL, PERFORMANCE, MULTIPLE FLOWS.
I. INTRODUCTION
Due to the ease of finding information, data access and utilization by a simple mouse move in
multimedia pages, Internet has become necessary in our daily lives. In fact, the number of Internet users is
growing rapidly, and according to the ITU (International Telecommunication Union) it reached only 500 million
users in the beginning of 2000s while it was 1.6 billion in 2008 to rise to the number of 2 billions in 2011.
This increase must undergo a parallel development of both of hardware and software at the same time.
Early, throughput was about 56 kbps, however, now we talk about Gigabit Internet aiming to meet the usersโ€™
needs mainly in astronomy and video conference fields.
In Internet, there are some machines responsible for transmitting data packets, and other that take off
these packets from the queue. If the sender machine sends a data packet rate much more important than the rate
of the receiver, network congestion is produced, hence, a congestion control will limit the quantity of input
information with a lower rate than the transmission one to guarantee a good performance as well as a network
protection against overloading and blocking.
The high number of high speed congestion control protocols led us to prepare this research work,
which focuses on evaluating and comparing high speed congestion control protocols.
This paper is organized as follows:
In the second section we study the state of the art. In the third one, we present the architectures used as
well as the curves and the performances evaluation for different high speed congestion control protocols.
STATE OF THE ART
Researchers have worked on the enhancement of high speed congestion control protocols. Practically
every year, one or two protocols are implemented having for each one of them its own specific strengths and
weaknesses.
Recently, the research works are interested in evaluating the performance of these protocols by
comparing among 2 to 5 protocols with 1 to 12 flows [1] [2] [3] [4] [5].
HS-TCP [6]: If the congestion window is low, HS-TCP behaves just like the standard TCP, once the congestion
window exceeds the max_ssthresh, it rises in an aggressive way. The algorithm of HS-TCP is an AIMD
Additive Increment cwnd = cwnd + a(cwnd) Multiplicative Decrement cwnd = (1- b(cwnd)) x cwnd, the values
used for HSTCP are a(cwnd) in the interval of [1,72], b(cwnd) in the interval of [0.1,0.5] and low_window
equals 38 packets. when cwnd is lower than low_window a(cwnd) = 1 and b(cwnd) = 0.5 thus the algorithm acts
just like a standard TCP, in the case of a packet loss, the cwnd is reduced to 50%, when the cwnd reaches the
high_window a(cwnd ) = 72 and b(cwnd)=1.
If there is an acknowledgment:
๐‘ค1 =
1 + ฮฒ
2
ร— cwnd cwnd < ๐‘ค1
๐‘ค1 = cwnd cwnd โ‰ฅ ๐‘ค1
w2 = cwnd
Scalable TCP [7]: For this protocol the cwnd increases with 0.001, if a congestion is detected the cwnd reduces
with 0.125 independently from the throughput and the packet size. This algorithm is a MIMD; Multiplicative
Performance Evaluation Of High Speed Congestion Control Protocols
www.iosrjournals.org 14 | Page
Increase since the cwnd increases with 0.01 per acknowledgment, Therefore there is a total increase of 0.01
cwnd per RTT, and Multiplicative Decrease since the cwnd decreases with 0.125 per packet loss.
BIC TCP [8]: uses a function that allows to have a rapid increase of cwnd when its value is far from the
maximal threshold Smax and to slowly decrease when having a value close to w_1.
cwnd = (1 โ€“ ฮฒ) ร— cwnd
With :
fฮฑ (ฮด, cwnd) =
ฮฒ
ฮด
ฮด โ‰ค 1 and cwnd < w1
or w1 โ‰ค cwnd < w1 + ฮฒ
ฮด 1 < ฮด โ‰ค Smax and cwnd < w1
w1
ฮฒโˆ’1
ฮฒ โ‰ค cwnd โˆ’ w1 < Smax 1 โˆ’ ฮฒ
Smax otherwise
cwnd : Current congestion window
Smax : Maximal threshold
Smin : Minimal threshold
ฮ’ : Decrease factor usually equals 0.875
CUBIC TCP [9]: The name CUBIC refers to the function of the congestion window increase which is cubic,
when receiving an ACK:
cwnd = C (T โ€“ K) ยณ + Maxcwnd
When having a packet loss:
cwnd = ฮฒ Maxcwnd
With :
HTCP [10]: uses ฮ” the duration since the last congestion rather than cwnd as information about the product
delay bandwidth BDP. The increase parameter of AIMD (additive-increase/multiplicative-decrease) varies
depending on ฮ”. It also depends on the value of RTT which has provoked the unfairness among the competitor
flows with different RTTs.
In case of receiving an ACK:
cwnd = cwnd +
2 1โˆ’ฮฒ fฮฑ(ฮ”)
cwnd
In case of packet loss:
cwnd = gฮฒ(B) ร— cwnd
With :
fฮฑ(ฮ”) =
1 ฮ” โ‰ค ฮ”L
max fฮฑ ฮ” Tmin ,1 ฮ” > ฮ”L
gฮฒ(B) =
0.5
B K+1 โ€“ B(k)
B(K)
> ฮ”B
min
Tmin
Tmax
,0.8 otherwise
cwnd*:*Current congestion window
ฮฑ*:*Increase parameter
ฮฒ*:*Decrease parameter
ฮ”*:*Duration of time since the last congestion
ฮ”L*:*Threshold to change the mode low speed to highspeed
Tmin *:*Minimal duration of sending and receiving packets
Tmax *:*Maximal duration of sending and receiving packets
B K + 1 *:*Maximal throughput reached before a congestion
fฮฑ ฮ” *=*1 +10 (ฮ” - ฮ”L) + 0,25 (ฮ” - ฮ”L)ยฒ
Cwnd*:*Current congestion window
C*:*Constant
T*:*Duration of time since the last congestion
Maxcwnd *:*Size of the last congestion window
K*:* Maxcwnd ฮฒ/C3
Performance Evaluation Of High Speed Congestion Control Protocols
www.iosrjournals.org 15 | Page
Illinois TCP [11] : Similar to standard TCP, Illinois-TCP increases the congestion window by using two
parameters ฮฑ and ฮฒ, it is based on packet loss to define the congestion window value, however the values of ฮฑ
and ฮฒ are not constant by using the delays.
When receiving an ACK:
cwnd = cwnd +
ฮฑ
๐‘๐‘ค๐‘›๐‘‘
When losing packets:
cwnd = (1- ฮฒ) ร— cwnd
With :
ฮฑ = ๐‘“1(๐‘‘ ๐‘Ž ) =
ฮฑ ๐‘š๐‘Ž๐‘ฅ ๐‘‘ ๐‘Ž โ‰ค ๐‘‘1
๐‘˜1
๐‘˜2+ ๐‘‘ ๐‘Ž
๐‘œ๐‘กโ„Ž๐‘’๐‘Ÿ๐‘ค๐‘–๐‘ ๐‘’
ฮฒ = ๐‘“2(๐‘‘ ๐‘Ž ) =
ฮฒ ๐‘š๐‘–๐‘›
๐‘‘ ๐‘Ž โ‰ค ๐‘‘2
๐‘˜3 + ๐‘˜4 ๐‘‘ ๐‘Ž ๐‘‘2 < ๐‘‘ ๐‘Ž < ๐‘‘3
ฮฒ ๐‘š๐‘Ž๐‘ฅ
๐‘œ๐‘กโ„Ž๐‘’๐‘Ÿ๐‘ค๐‘–๐‘ ๐‘’
๐‘˜1 =
(๐‘‘ ๐‘š โˆ’ ๐‘‘1) ฮฑ ๐‘š๐‘–๐‘› ฮฑ ๐‘š๐‘Ž๐‘ฅ
ฮฑ ๐‘š๐‘Ž๐‘ฅ โˆ’ ฮฑ ๐‘š๐‘–๐‘›
*
๐‘˜2 =
(๐‘‘ ๐‘š โˆ’ ๐‘‘1) ฮฑ ๐‘š๐‘–๐‘›
ฮฑ ๐‘š๐‘Ž๐‘ฅ โˆ’ ฮฑ ๐‘š๐‘–๐‘›
- ๐‘‘1
๐‘˜3 =
ฮฒ ๐‘š๐‘–๐‘› ๐‘‘3โˆ’ ฮฒ ๐‘š๐‘Ž๐‘ฅ ๐‘‘2
๐‘‘3โˆ’ ๐‘‘2
*๐‘˜4 =
ฮฒ ๐‘š๐‘Ž๐‘ฅ โˆ’ ฮฒ ๐‘š๐‘–๐‘›
๐‘‘3โˆ’ ๐‘‘2
ฮฑ ๐‘š๐‘–๐‘› =
๐‘˜1
๐‘˜2+ ๐‘‘ ๐‘š๐‘œ๐‘ฆ
, ฮฑ ๐‘š๐‘Ž๐‘ฅ =
๐‘˜1
๐‘˜2+ ๐‘‘1
ฮฒ ๐‘š๐‘–๐‘›
= ๐‘˜3 + ๐‘˜4 ๐‘‘2 , ฮฒ ๐‘š๐‘Ž๐‘ฅ
=๐‘˜3 + ๐‘˜4 ๐‘‘3
YeAH TCP [12] the algorithm starts with the Slow Start if cwnd < ssthresh and Scalable TCP once cwnd
reaches the ssthresh. This protocol uses two modes, fast mode and slow mode. In the case of the fast one, the
congestion window increases aggressively, for the second mode, the protocol behaves similarly to Reno TCP.
The mode is chosen according to the queue status. Having RTTBASE the minimal RTT measured by the sender
and RTTMIN the minimal RTT estimated from the current window. The estimated delay is:
RTTqueue = RTTmin โˆ’ RTTbase
Thanks to this value, the number of packet on hold in the queue can be defined as follows:
Q = RTTqueue ร— G = RTTqueue ร—
cwnd
RTTmin
And G is the throughput.
II. SIMULATION AND PERFORMANCES EVALUATION
2.1. Architecture
In order to simulate the high speed congestion control protocols, we have chosen a topology (as shown
in figure 1) composed of a sender and a receiver linked together with two routers by a line of 1Gbps of
bandwidth, the delay is 1ms. The routers are linked to each other with a line having a bandwidth of 200Mbps,
the delay is equal to 92 ms and the queue capacity is exactly 100 packets [13]. The MSS size equals 1460 bytes.
The differences among the bandwidths capacities provoke congestion.
Figure 1. Basic topology
To evaluate the performances, we have used 2, 4, 6, 8, 10, 12, 16 and 24 identical flows (as shown in
figure 2) with a link capacity of 1 Gbps and a delay of 1 ms.
Figure 2. Topology with multiple flows
2.2. TCP-Friendly
To test the TCP Friendly of different protocols, we have used a multiple flows topology with 4 input
flows triggered at the same time. The results are shown in the following graphs:
Performance Evaluation Of High Speed Congestion Control Protocols
www.iosrjournals.org 16 | Page
Figure 3.a. HS TCP : Throughput variation for 4 flows
*
Figure 3.b. Scalable TCP : Throughput variation for 4 flows
Figure 3.c. Bic TCP : Throughput variation for 4 flows
*
Figure 3.d. Hamilton TCP : Throughput variation for 4 flows
Figure 3.e. Cubic TCP : Throughput variation for 4 flows
*
Figure 3.f. Illinois TCP : Throughput variation for 4 flows
Figure 3.g. YeAH TCP : Throughput variation for 4 flows
Performance Evaluation Of High Speed Congestion Control Protocols
www.iosrjournals.org 17 | Page
2.3. Efficiency
The efficiency (rate of utilization or performance) of a network is the percentage of utilization [14].
Mathematically, it is the division of the average throughput by the optimal throughput.
An efficient network uses the maximum of capacity.
The following variable ๐‘ž๐‘– is used as the average throughput of a flow i.
The throughput of a source with n flows is:
๐‘„ =
1
๐‘›
๐‘ž๐‘–
๐‘›
๐‘–=1
We calculate the ratio between the average throughput and the optimal one which is the result of an ideal
network performance:
๐ธ =
๐‘„
๐‘„ ๐‘œ๐‘๐‘ก
We have used these two topologies to simulate the seven protocols for 1, 2, 4, 6, 8, 10, 12, 16 and 24 flows.
The efficiency calculus results are as mentioned in the following graph:
Figure 4. Efficiency for different flow numbers
2.4. Fairness
The fairness is the attempt of sharing the network capacities among users in a fair way.
For the purpose of measuring the fairness, one method is used in the networks field which is called the Maximin
law proposed by Raj Jain [15]. Here is the procedure that allows us to calculate the fairness of a proposed
algorithm:
Having an algorithm that provides the distribution vi = [x1, x2,โ€ฆ,xn] instead of the optimal distribution vopt =
[x1,opt , x2,opt , โ€ฆ , xn,opt ]. We calculate the standardized distribution for every source as follows:
๐‘‹๐‘– =
๐‘ฅ๐‘–
๐‘ฅ ๐‘œ๐‘๐‘ก
Thus, the fairness index F equals the sum of distributions squared and divided by the square of sums:
๐น =
( ๐‘‹๐‘–
๐‘›
๐‘–=1 )ยฒ
๐‘› ๐‘‹๐‘–ยฒ๐‘›
๐‘–=1
We have used these two topologies to simulate the seven protocols for 1, 2, 4, 6, 8, 10, 12, 16 and 24 flows.
The fairness calculus results are as mentioned in the following graph:
Figure 5. Fairness for different flow numbers
2.5. Performance
The performance of a congestion control algorithm is the relation between efficiency and fairness:
Performance = ฮฑ ร— E + (1 โ€“ ฮฑ) ร— F
With ฮฑ = [0,1.. 0,9], E the efficiency and F the fairness.
For the different algorithms, we have calculated the performances for a network rather efficient with ฮฑ = 0,8, a
network rather fair with ฮฑ = 0,2 and a balanced network when ฮฑ = 0,5. The results for 1, 2, 4, 6, 8, 10, 12, 16 and
24 flows are as follows:
Performance Evaluation Of High Speed Congestion Control Protocols
www.iosrjournals.org 18 | Page
Figure 6.a. Performance for different flow numbers (ฮฑ = 0,8)
Figure 6.b. Performance for different flow numbers (ฮฑ = 0,5)
Figure 6.c. Performance for different flow numbers (ฮฑ = 0,2)
2.6. Link Utilisation per RTT
We varied the RTT for different congestion control algorithms to calculate the link utilization. The
used RTT values are: 20 ms, 60 ms, 100 ms, 140 ms, 180 ms and 220 ms.
We used the basic topology with only one flow; the results are described in the following graph:
Figure 7. Link utilization for different RTTs
2.7. Link utilization for different queue sizes
Certainly the queue size has an important impact on the efficiency of the network, but which algorithm
benefits the more of this change? We varied the queue size for different congestion control algorithms to
calculate the link utilization. The used sizes are: 50, 100, 150, 200, 250 and 300 packets.
We used the basic topology with one flow, the results are as follows:
Figure 8. Link utilization for different queue sizes
III. CONCLUSIONS
The high number of congestion control algorithms never satisfied the researchers. Since this field is
always having changes and enhancements due to the usersโ€™ needs and the evolution of both of hardware and
software in Telecommunications. Indeed, it is difficult to conceive an algorithm that gives satisfying results for
all the architectures.
Performance Evaluation Of High Speed Congestion Control Protocols
www.iosrjournals.org 19 | Page
During this research work, we simulated seven high speed congestion control protocols for 1, 2, 4, 6, 8,
10, 12, 16 and 24 flows. We evaluated their performances by calculating the efficiency, the fairness as well as
the performance while varying the RTT and the queue size.
Basing on the simulation and the high speed protocols performance evaluation, we can conclude that:
โ€ข Some protocols perform well in some defined cases, but weak in others.
โ€ข The network architecture has got an important impact on the protocols performance.
โ€ข For 24 flows, all protocols are unfair.
โ€ข For 24 flows, the protocols: Bic TCP, Cubic TCP, Hamilton TCP, Scalable TCP and YeAH TCP are
efficient.
โ€ข Scalable is not only TCP-friendly, but also selfish, since the first flow gets the most important throughput,
and doesnโ€™t leave its place for any other flow such as the case of Cubic TCP, Hamilton TCP and Illinois
TCP.
โ€ข The protocols Bic TCP, HighSpeed TCP and YeAH TCP are TCP-friendly
โ€ข For low values of RTT, all the protocols use the network capacities in a bad way.
โ€ข For low values of the queue size, Cubic TCP is the most efficient.
โ€ข YeAH TCP and Illinois TCP give the best results while increasing the queue size.
โ€ข Although his aggressiveness, YeAH TCP is quite TCP-friendly. It is also quite performant, but it doesnโ€™t
suit all the architectures.
Further work is required to evaluate the performances of these protocols by using Relentless TCP
which will be the topic for a multitude of research works in the near future, particularly to define if the protocols
that have more stability and take more time before falling another time in a new congestion such as Cubic TCP
and Illinois TCP in order to check if they will well exploit the law that forms the basis of Relentless TCP which
is the reduction of the congestion window with the number of lost segments. It will be also interesting to create a
model that changes dynamically the congestion control protocols used in terms of flow number to better exploit
the network capacities as well as studying the QoS.
REFERENCES
[1] R. Mbarek, M. T. BenOthman and S. Nasri. โ€œFairness of High-Speed TCP Protocols with Different Flow Capacitiesโ€. Journal Of
Networks, Vol. 4, No. 3, pp. 163-169, May 2009.
[2] S. Hemminger. โ€œTCP testing โ€“ Preventing global Internet warmingโ€. LinuxConf Europe, Sep. 2007.
[3] S. Ha, L. Le, I. Rhee and L. Xu. โ€œImpact of Background Traffic on Performance of High-Speed TCP Variant Protocolsโ€. The
International Journal of Computer and Telecommunications Networking, Vol.51, May, 2007.
[4] M. Nabeshima and K. Yata. โ€œPerformance Evaluation and Comparison of Transport Protocols for Fast Long-Distance Networksโ€.
IEICE Trans. Commun., Vol.E89-B, No.4, pp. 1273-1283, Apr. 2006.
[5] M. Weigle, P. Sharma and J. Freeman. โ€œPerformance of Competing High-Speed TCP Flowsโ€. Networking 2006, LNCS 3976, pp.
476-487, 2006.
[6] S. Floyd. โ€œHigh-Speed TCP for Large Congestion Windowsโ€. RFC 3649, Experimental, Dec. 2003.
[7] T. Kelly. โ€œScalable TCP : Improving Performance in High-Speed Wide Area Networksโ€. In Proceedings of PFLDnet, Geneva,
Switzerland, Feb. 2003.
[8] L. Xu, K. Harfoush, and I. Rhee. โ€œBinary Increase Congestion Control for Fast, Long Distance Networksโ€. In Proceedings of IEEE
INFOCOM, Hong Kong, Mar. 2004.
[9] I. Rhee and L. Xu. โ€œCUBIC: A New TCP-Friendly High-Speed TCP Variantโ€. In Proceedings of PFLDnet, Lyon, France, Feb.
2005.
[10] R. N. Shorten and D. J. Leith. โ€œH-TCP : TCP for high-speed and long-distance networksโ€. In Proceedings of PFLDnet, Argonne,
Feb. 2004.
[11] S. Liu, T. Basar and R. Srikant.,"TCP-Illinois : A Loss- and Delay-based Congestion Control Algorithm for High-Speed Networks",
Performance Evaluation, 65(6-7) : 417-440, Mar. 2008.
[12] A. Baiocchi, A. P. Castellani and F. Vacirca. โ€œYeAH-TCP: Yet Another Highspeed TCPโ€. PFLDnet, Roma, Italy, Feb. 2007.
[13] J. Semke, J. Mahdavi, and M. Mathis. "Automatic TCP Buffer Tuning", Computer Communications Review, a publication of ACM
SIGCOMM, Vol. 28, No. 4, Oct. 1998.
[14] R. Mbarek, M. T. BenOthman and S. Nasri. โ€œPerformance Evaluation of Competing High-Speed TCP Protocolsโ€. IJCSNS
International Journal of Computer Science and Network Security, VOL.8 No.6, pp. 99-105, June 2008.
[15] R. Jain. โ€œCongestion Control and Traffic Management in ATM Networks : Recent Advances and A Surveyโ€. Computer networks
and ISDN Systems, Nov. 1995.
Authors
M. Mohamed Ali Mani received his M.S degree from ISITCOM, Hammam Sousse, Sousse University Tunisia, in 2011.
He is currently a lecturer in the Department of Computer Science in ISTLS in the same University. His research interests
lie in the general area of computer networks, including high speed transports protocols, flow control, congestion control in
high-speed networks and wireless networks.
Dr. Rachid Mbarek received his Ph.D from the College of Engineering (ENIS), Sfax University, Tunisia, received his
M.Sc. degree in Computer Engineering in 2003 from the same University. He is currently a Lecturer in the Department of
Telecommunication in the ISITCOM at Sousse University, Tunisia. His research interests lie in the general area of
computer communication networks, including high-speed transports protocols, flow control and congestion control in
high-speed networks, network performance analysis, peer to peer networking.

More Related Content

What's hot

TCP-FIT: An Improved TCP Congestion Control Algorithm and its Performance
TCP-FIT: An Improved TCP Congestion Control Algorithm and its PerformanceTCP-FIT: An Improved TCP Congestion Control Algorithm and its Performance
TCP-FIT: An Improved TCP Congestion Control Algorithm and its PerformanceKevin Tong
ย 
Tcp congestion control
Tcp congestion controlTcp congestion control
Tcp congestion controlAbdo sayed
ย 
A Study on MPTCP for Tolerating Packet Reordering and Path Heterogeneity in W...
A Study on MPTCP for Tolerating Packet Reordering and Path Heterogeneity in W...A Study on MPTCP for Tolerating Packet Reordering and Path Heterogeneity in W...
A Study on MPTCP for Tolerating Packet Reordering and Path Heterogeneity in W...Communication Systems & Networks
ย 
Tcp congestion control (1)
Tcp congestion control (1)Tcp congestion control (1)
Tcp congestion control (1)Abdo sayed
ย 
Scheduling and Resource allocation in 802.11ax (WIFI 6)
Scheduling and Resource allocation in 802.11ax (WIFI 6)Scheduling and Resource allocation in 802.11ax (WIFI 6)
Scheduling and Resource allocation in 802.11ax (WIFI 6)RESHAN FARAZ
ย 
Lect9
Lect9Lect9
Lect9Abdo sayed
ย 
Congestion control in tcp
Congestion control in tcpCongestion control in tcp
Congestion control in tcpsamarai_apoc
ย 
Congetion Control.pptx
Congetion Control.pptxCongetion Control.pptx
Congetion Control.pptxNaveen Dubey
ย 
Ad Hoc Probe
Ad Hoc ProbeAd Hoc Probe
Ad Hoc Probenutikumar
ย 
Computer network (13)
Computer network (13)Computer network (13)
Computer network (13)NYversity
ย 
Available network bandwidth schema to improve performance in tcp protocols
Available network bandwidth schema to improve performance in tcp protocolsAvailable network bandwidth schema to improve performance in tcp protocols
Available network bandwidth schema to improve performance in tcp protocolsIJCNCJournal
ย 
PERFORMANCE EVALUATION OF SELECTED E2E TCP CONGESTION CONTROL MECHANISM OVER ...
PERFORMANCE EVALUATION OF SELECTED E2E TCP CONGESTION CONTROL MECHANISM OVER ...PERFORMANCE EVALUATION OF SELECTED E2E TCP CONGESTION CONTROL MECHANISM OVER ...
PERFORMANCE EVALUATION OF SELECTED E2E TCP CONGESTION CONTROL MECHANISM OVER ...ijwmn
ย 
Tcp Congestion Avoidance
Tcp Congestion AvoidanceTcp Congestion Avoidance
Tcp Congestion AvoidanceRam Dutt Shukla
ย 

What's hot (20)

TCP-FIT: An Improved TCP Congestion Control Algorithm and its Performance
TCP-FIT: An Improved TCP Congestion Control Algorithm and its PerformanceTCP-FIT: An Improved TCP Congestion Control Algorithm and its Performance
TCP-FIT: An Improved TCP Congestion Control Algorithm and its Performance
ย 
HIGH SPEED NETWORKS
HIGH SPEED NETWORKSHIGH SPEED NETWORKS
HIGH SPEED NETWORKS
ย 
Ch24
Ch24Ch24
Ch24
ย 
TCP Congestion Control
TCP Congestion ControlTCP Congestion Control
TCP Congestion Control
ย 
presentation
presentationpresentation
presentation
ย 
Tcp congestion control
Tcp congestion controlTcp congestion control
Tcp congestion control
ย 
A Study on MPTCP for Tolerating Packet Reordering and Path Heterogeneity in W...
A Study on MPTCP for Tolerating Packet Reordering and Path Heterogeneity in W...A Study on MPTCP for Tolerating Packet Reordering and Path Heterogeneity in W...
A Study on MPTCP for Tolerating Packet Reordering and Path Heterogeneity in W...
ย 
Tcp congestion control (1)
Tcp congestion control (1)Tcp congestion control (1)
Tcp congestion control (1)
ย 
Scheduling and Resource allocation in 802.11ax (WIFI 6)
Scheduling and Resource allocation in 802.11ax (WIFI 6)Scheduling and Resource allocation in 802.11ax (WIFI 6)
Scheduling and Resource allocation in 802.11ax (WIFI 6)
ย 
Lect9
Lect9Lect9
Lect9
ย 
Chapter 23
Chapter 23Chapter 23
Chapter 23
ย 
Ch23
Ch23Ch23
Ch23
ย 
Congestion control in tcp
Congestion control in tcpCongestion control in tcp
Congestion control in tcp
ย 
Congetion Control.pptx
Congetion Control.pptxCongetion Control.pptx
Congetion Control.pptx
ย 
Ad Hoc Probe
Ad Hoc ProbeAd Hoc Probe
Ad Hoc Probe
ย 
Ch23
Ch23Ch23
Ch23
ย 
Computer network (13)
Computer network (13)Computer network (13)
Computer network (13)
ย 
Available network bandwidth schema to improve performance in tcp protocols
Available network bandwidth schema to improve performance in tcp protocolsAvailable network bandwidth schema to improve performance in tcp protocols
Available network bandwidth schema to improve performance in tcp protocols
ย 
PERFORMANCE EVALUATION OF SELECTED E2E TCP CONGESTION CONTROL MECHANISM OVER ...
PERFORMANCE EVALUATION OF SELECTED E2E TCP CONGESTION CONTROL MECHANISM OVER ...PERFORMANCE EVALUATION OF SELECTED E2E TCP CONGESTION CONTROL MECHANISM OVER ...
PERFORMANCE EVALUATION OF SELECTED E2E TCP CONGESTION CONTROL MECHANISM OVER ...
ย 
Tcp Congestion Avoidance
Tcp Congestion AvoidanceTcp Congestion Avoidance
Tcp Congestion Avoidance
ย 

Viewers also liked

A Critical Review of the Nigerian Energy Scenario
A Critical Review of the Nigerian Energy ScenarioA Critical Review of the Nigerian Energy Scenario
A Critical Review of the Nigerian Energy ScenarioIOSR Journals
ย 
A Novel Approach for Tracking with Implicit Video Shot Detection
A Novel Approach for Tracking with Implicit Video Shot DetectionA Novel Approach for Tracking with Implicit Video Shot Detection
A Novel Approach for Tracking with Implicit Video Shot DetectionIOSR Journals
ย 
J o b s new sky 1 unit 36
J o b s new sky 1 unit 36J o b s new sky 1 unit 36
J o b s new sky 1 unit 36Aleksandra Nikolic
ย 
Investigation on Using Fractal Geometry for Classification of Partial Dischar...
Investigation on Using Fractal Geometry for Classification of Partial Dischar...Investigation on Using Fractal Geometry for Classification of Partial Dischar...
Investigation on Using Fractal Geometry for Classification of Partial Dischar...IOSR Journals
ย 
Growth Performance of Rats Maintained On Citrullus colocynthis Seed Coat-base...
Growth Performance of Rats Maintained On Citrullus colocynthis Seed Coat-base...Growth Performance of Rats Maintained On Citrullus colocynthis Seed Coat-base...
Growth Performance of Rats Maintained On Citrullus colocynthis Seed Coat-base...IOSR Journals
ย 
ะŸั€ะพัั‚ั‹ะต ะธ ัะพัั‚ะฐะฒะฝั‹ะต ั‡ะธัะปะธั‚ะตะปัŒะฝั‹ะต
ะŸั€ะพัั‚ั‹ะต ะธ ัะพัั‚ะฐะฒะฝั‹ะต ั‡ะธัะปะธั‚ะตะปัŒะฝั‹ะตะŸั€ะพัั‚ั‹ะต ะธ ัะพัั‚ะฐะฒะฝั‹ะต ั‡ะธัะปะธั‚ะตะปัŒะฝั‹ะต
ะŸั€ะพัั‚ั‹ะต ะธ ัะพัั‚ะฐะฒะฝั‹ะต ั‡ะธัะปะธั‚ะตะปัŒะฝั‹ะตNatalya Dyrda
ย 
Application of Comparators in Modern Power System Protection and Control
Application of Comparators in Modern Power System Protection and ControlApplication of Comparators in Modern Power System Protection and Control
Application of Comparators in Modern Power System Protection and ControlIOSR Journals
ย 
Threshold Secure B2B Model
Threshold Secure B2B ModelThreshold Secure B2B Model
Threshold Secure B2B ModelIOSR Journals
ย 
Rp presentation
Rp presentationRp presentation
Rp presentationKhalid Ak
ย 
Dewis cv
Dewis cvDewis cv
Dewis cvdewisnatra
ย 
Survey and Analysis of Medium Access Control Protocols for Wireless Sensor Ne...
Survey and Analysis of Medium Access Control Protocols for Wireless Sensor Ne...Survey and Analysis of Medium Access Control Protocols for Wireless Sensor Ne...
Survey and Analysis of Medium Access Control Protocols for Wireless Sensor Ne...IOSR Journals
ย 
Survey of different Web Application Attacks & Its Preventive Measures
Survey of different Web Application Attacks & Its Preventive MeasuresSurvey of different Web Application Attacks & Its Preventive Measures
Survey of different Web Application Attacks & Its Preventive MeasuresIOSR Journals
ย 
Abecedario
AbecedarioAbecedario
Abecedariohernan1951
ย 
ะปะธั†ะตะน ะฟัƒัˆะบะธะฝะฐ
ะปะธั†ะตะน ะฟัƒัˆะบะธะฝะฐะปะธั†ะตะน ะฟัƒัˆะบะธะฝะฐ
ะปะธั†ะตะน ะฟัƒัˆะบะธะฝะฐNatalya Dyrda
ย 
Signal Processing Approach for Recognizing Identical Reads From DNA Sequencin...
Signal Processing Approach for Recognizing Identical Reads From DNA Sequencin...Signal Processing Approach for Recognizing Identical Reads From DNA Sequencin...
Signal Processing Approach for Recognizing Identical Reads From DNA Sequencin...IOSR Journals
ย 
Introtothemes
IntrotothemesIntrotothemes
Introtothemesmelissataggart
ย 
ะ—ะฝะฐะตั‚ะต ะปะธ ะฒั‹ ัะบะฐะทะบะธ ะ.ะก. ะŸัƒัˆะบะธะฝะฐ
ะ—ะฝะฐะตั‚ะต ะปะธ ะฒั‹ ัะบะฐะทะบะธ ะ.ะก. ะŸัƒัˆะบะธะฝะฐะ—ะฝะฐะตั‚ะต ะปะธ ะฒั‹ ัะบะฐะทะบะธ ะ.ะก. ะŸัƒัˆะบะธะฝะฐ
ะ—ะฝะฐะตั‚ะต ะปะธ ะฒั‹ ัะบะฐะทะบะธ ะ.ะก. ะŸัƒัˆะบะธะฝะฐNatalya Dyrda
ย 

Viewers also liked (20)

A Critical Review of the Nigerian Energy Scenario
A Critical Review of the Nigerian Energy ScenarioA Critical Review of the Nigerian Energy Scenario
A Critical Review of the Nigerian Energy Scenario
ย 
A Novel Approach for Tracking with Implicit Video Shot Detection
A Novel Approach for Tracking with Implicit Video Shot DetectionA Novel Approach for Tracking with Implicit Video Shot Detection
A Novel Approach for Tracking with Implicit Video Shot Detection
ย 
J o b s new sky 1 unit 36
J o b s new sky 1 unit 36J o b s new sky 1 unit 36
J o b s new sky 1 unit 36
ย 
Investigation on Using Fractal Geometry for Classification of Partial Dischar...
Investigation on Using Fractal Geometry for Classification of Partial Dischar...Investigation on Using Fractal Geometry for Classification of Partial Dischar...
Investigation on Using Fractal Geometry for Classification of Partial Dischar...
ย 
Growth Performance of Rats Maintained On Citrullus colocynthis Seed Coat-base...
Growth Performance of Rats Maintained On Citrullus colocynthis Seed Coat-base...Growth Performance of Rats Maintained On Citrullus colocynthis Seed Coat-base...
Growth Performance of Rats Maintained On Citrullus colocynthis Seed Coat-base...
ย 
ะŸั€ะพัั‚ั‹ะต ะธ ัะพัั‚ะฐะฒะฝั‹ะต ั‡ะธัะปะธั‚ะตะปัŒะฝั‹ะต
ะŸั€ะพัั‚ั‹ะต ะธ ัะพัั‚ะฐะฒะฝั‹ะต ั‡ะธัะปะธั‚ะตะปัŒะฝั‹ะตะŸั€ะพัั‚ั‹ะต ะธ ัะพัั‚ะฐะฒะฝั‹ะต ั‡ะธัะปะธั‚ะตะปัŒะฝั‹ะต
ะŸั€ะพัั‚ั‹ะต ะธ ัะพัั‚ะฐะฒะฝั‹ะต ั‡ะธัะปะธั‚ะตะปัŒะฝั‹ะต
ย 
Application of Comparators in Modern Power System Protection and Control
Application of Comparators in Modern Power System Protection and ControlApplication of Comparators in Modern Power System Protection and Control
Application of Comparators in Modern Power System Protection and Control
ย 
Threshold Secure B2B Model
Threshold Secure B2B ModelThreshold Secure B2B Model
Threshold Secure B2B Model
ย 
Rp presentation
Rp presentationRp presentation
Rp presentation
ย 
Dewis cv
Dewis cvDewis cv
Dewis cv
ย 
A0350107
A0350107A0350107
A0350107
ย 
Survey and Analysis of Medium Access Control Protocols for Wireless Sensor Ne...
Survey and Analysis of Medium Access Control Protocols for Wireless Sensor Ne...Survey and Analysis of Medium Access Control Protocols for Wireless Sensor Ne...
Survey and Analysis of Medium Access Control Protocols for Wireless Sensor Ne...
ย 
Survey of different Web Application Attacks & Its Preventive Measures
Survey of different Web Application Attacks & Its Preventive MeasuresSurvey of different Web Application Attacks & Its Preventive Measures
Survey of different Web Application Attacks & Its Preventive Measures
ย 
Bangkok
BangkokBangkok
Bangkok
ย 
Abecedario
AbecedarioAbecedario
Abecedario
ย 
ะปะธั†ะตะน ะฟัƒัˆะบะธะฝะฐ
ะปะธั†ะตะน ะฟัƒัˆะบะธะฝะฐะปะธั†ะตะน ะฟัƒัˆะบะธะฝะฐ
ะปะธั†ะตะน ะฟัƒัˆะบะธะฝะฐ
ย 
Signal Processing Approach for Recognizing Identical Reads From DNA Sequencin...
Signal Processing Approach for Recognizing Identical Reads From DNA Sequencin...Signal Processing Approach for Recognizing Identical Reads From DNA Sequencin...
Signal Processing Approach for Recognizing Identical Reads From DNA Sequencin...
ย 
Introtothemes
IntrotothemesIntrotothemes
Introtothemes
ย 
10____
  10____  10____
10____
ย 
ะ—ะฝะฐะตั‚ะต ะปะธ ะฒั‹ ัะบะฐะทะบะธ ะ.ะก. ะŸัƒัˆะบะธะฝะฐ
ะ—ะฝะฐะตั‚ะต ะปะธ ะฒั‹ ัะบะฐะทะบะธ ะ.ะก. ะŸัƒัˆะบะธะฝะฐะ—ะฝะฐะตั‚ะต ะปะธ ะฒั‹ ัะบะฐะทะบะธ ะ.ะก. ะŸัƒัˆะบะธะฝะฐ
ะ—ะฝะฐะตั‚ะต ะปะธ ะฒั‹ ัะบะฐะทะบะธ ะ.ะก. ะŸัƒัˆะบะธะฝะฐ
ย 

Similar to Performance Evaluation of High Speed Congestion Control Protocols

ANALYSIS AND EXPERIMENTAL EVALUATION OF THE TRANSMISSION CONTROL PROTOCOL CON...
ANALYSIS AND EXPERIMENTAL EVALUATION OF THE TRANSMISSION CONTROL PROTOCOL CON...ANALYSIS AND EXPERIMENTAL EVALUATION OF THE TRANSMISSION CONTROL PROTOCOL CON...
ANALYSIS AND EXPERIMENTAL EVALUATION OF THE TRANSMISSION CONTROL PROTOCOL CON...IRJET Journal
ย 
A Packet Drop Guesser Module for Congestion Control Protocols for High speed ...
A Packet Drop Guesser Module for Congestion Control Protocols for High speed ...A Packet Drop Guesser Module for Congestion Control Protocols for High speed ...
A Packet Drop Guesser Module for Congestion Control Protocols for High speed ...ijcseit
ย 
A packet drop guesser module for congestion Control protocols for high speed ...
A packet drop guesser module for congestion Control protocols for high speed ...A packet drop guesser module for congestion Control protocols for high speed ...
A packet drop guesser module for congestion Control protocols for high speed ...ijcseit
ย 
tcp congestion .pptx
tcp congestion .pptxtcp congestion .pptx
tcp congestion .pptxECE01AJAYS
ย 
Lecture 19 22. transport protocol for ad-hoc
Lecture 19 22. transport protocol for ad-hoc Lecture 19 22. transport protocol for ad-hoc
Lecture 19 22. transport protocol for ad-hoc Chandra Meena
ย 
U01725129138
U01725129138U01725129138
U01725129138IOSR Journals
ย 
Iaetsd an effective approach to eliminate tcp incast
Iaetsd an effective approach to eliminate tcp incastIaetsd an effective approach to eliminate tcp incast
Iaetsd an effective approach to eliminate tcp incastIaetsd Iaetsd
ย 
chapter 3.2 TCP.pptx
chapter 3.2 TCP.pptxchapter 3.2 TCP.pptx
chapter 3.2 TCP.pptxTekle12
ย 
IMPACT OF CONTENTION WINDOW ON CONGESTION CONTROL ALGORITHMS FOR WIRELESS ADH...
IMPACT OF CONTENTION WINDOW ON CONGESTION CONTROL ALGORITHMS FOR WIRELESS ADH...IMPACT OF CONTENTION WINDOW ON CONGESTION CONTROL ALGORITHMS FOR WIRELESS ADH...
IMPACT OF CONTENTION WINDOW ON CONGESTION CONTROL ALGORITHMS FOR WIRELESS ADH...cscpconf
ย 
P2885 jung
P2885 jungP2885 jung
P2885 jungganeshgss1
ย 
Studying_the_TCP_Flow_and_Congestion_Con.pdf
Studying_the_TCP_Flow_and_Congestion_Con.pdfStudying_the_TCP_Flow_and_Congestion_Con.pdf
Studying_the_TCP_Flow_and_Congestion_Con.pdfIUA
ย 
TCP Traffic Control Chapter12
TCP Traffic Control Chapter12TCP Traffic Control Chapter12
TCP Traffic Control Chapter12daniel ayalew
ย 
Congestion Control: A Dynamic Approach
Congestion Control: A Dynamic ApproachCongestion Control: A Dynamic Approach
Congestion Control: A Dynamic ApproachIOSR Journals
ย 
Congestion Control: A Dynamic Approach
Congestion Control: A Dynamic ApproachCongestion Control: A Dynamic Approach
Congestion Control: A Dynamic ApproachIOSR Journals
ย 
"Performance Evaluation and Comparison of Westwood+, New Reno and Vegas TCP ...
 "Performance Evaluation and Comparison of Westwood+, New Reno and Vegas TCP ... "Performance Evaluation and Comparison of Westwood+, New Reno and Vegas TCP ...
"Performance Evaluation and Comparison of Westwood+, New Reno and Vegas TCP ...losalamos
ย 
TCP_Congestion_Control.ppt
TCP_Congestion_Control.pptTCP_Congestion_Control.ppt
TCP_Congestion_Control.ppt19UCSA032ASANJAYKUMA
ย 
DCN 5th ed. slides ch24 Transport-Layer Protocols.pdf
DCN 5th ed. slides ch24 Transport-Layer Protocols.pdfDCN 5th ed. slides ch24 Transport-Layer Protocols.pdf
DCN 5th ed. slides ch24 Transport-Layer Protocols.pdfBilal Munir Mughal
ย 

Similar to Performance Evaluation of High Speed Congestion Control Protocols (20)

ANALYSIS AND EXPERIMENTAL EVALUATION OF THE TRANSMISSION CONTROL PROTOCOL CON...
ANALYSIS AND EXPERIMENTAL EVALUATION OF THE TRANSMISSION CONTROL PROTOCOL CON...ANALYSIS AND EXPERIMENTAL EVALUATION OF THE TRANSMISSION CONTROL PROTOCOL CON...
ANALYSIS AND EXPERIMENTAL EVALUATION OF THE TRANSMISSION CONTROL PROTOCOL CON...
ย 
A Packet Drop Guesser Module for Congestion Control Protocols for High speed ...
A Packet Drop Guesser Module for Congestion Control Protocols for High speed ...A Packet Drop Guesser Module for Congestion Control Protocols for High speed ...
A Packet Drop Guesser Module for Congestion Control Protocols for High speed ...
ย 
A packet drop guesser module for congestion Control protocols for high speed ...
A packet drop guesser module for congestion Control protocols for high speed ...A packet drop guesser module for congestion Control protocols for high speed ...
A packet drop guesser module for congestion Control protocols for high speed ...
ย 
tcp congestion .pptx
tcp congestion .pptxtcp congestion .pptx
tcp congestion .pptx
ย 
Lecture 19 22. transport protocol for ad-hoc
Lecture 19 22. transport protocol for ad-hoc Lecture 19 22. transport protocol for ad-hoc
Lecture 19 22. transport protocol for ad-hoc
ย 
U01725129138
U01725129138U01725129138
U01725129138
ย 
Iaetsd an effective approach to eliminate tcp incast
Iaetsd an effective approach to eliminate tcp incastIaetsd an effective approach to eliminate tcp incast
Iaetsd an effective approach to eliminate tcp incast
ย 
chapter 3.2 TCP.pptx
chapter 3.2 TCP.pptxchapter 3.2 TCP.pptx
chapter 3.2 TCP.pptx
ย 
IMPACT OF CONTENTION WINDOW ON CONGESTION CONTROL ALGORITHMS FOR WIRELESS ADH...
IMPACT OF CONTENTION WINDOW ON CONGESTION CONTROL ALGORITHMS FOR WIRELESS ADH...IMPACT OF CONTENTION WINDOW ON CONGESTION CONTROL ALGORITHMS FOR WIRELESS ADH...
IMPACT OF CONTENTION WINDOW ON CONGESTION CONTROL ALGORITHMS FOR WIRELESS ADH...
ย 
P2885 jung
P2885 jungP2885 jung
P2885 jung
ย 
Studying_the_TCP_Flow_and_Congestion_Con.pdf
Studying_the_TCP_Flow_and_Congestion_Con.pdfStudying_the_TCP_Flow_and_Congestion_Con.pdf
Studying_the_TCP_Flow_and_Congestion_Con.pdf
ย 
Fairness in Transfer Control Protocol for Congestion Control in Multiplicativ...
Fairness in Transfer Control Protocol for Congestion Control in Multiplicativ...Fairness in Transfer Control Protocol for Congestion Control in Multiplicativ...
Fairness in Transfer Control Protocol for Congestion Control in Multiplicativ...
ย 
TCP Traffic Control Chapter12
TCP Traffic Control Chapter12TCP Traffic Control Chapter12
TCP Traffic Control Chapter12
ย 
M017137072
M017137072M017137072
M017137072
ย 
Congestion Control: A Dynamic Approach
Congestion Control: A Dynamic ApproachCongestion Control: A Dynamic Approach
Congestion Control: A Dynamic Approach
ย 
Congestion Control: A Dynamic Approach
Congestion Control: A Dynamic ApproachCongestion Control: A Dynamic Approach
Congestion Control: A Dynamic Approach
ย 
"Performance Evaluation and Comparison of Westwood+, New Reno and Vegas TCP ...
 "Performance Evaluation and Comparison of Westwood+, New Reno and Vegas TCP ... "Performance Evaluation and Comparison of Westwood+, New Reno and Vegas TCP ...
"Performance Evaluation and Comparison of Westwood+, New Reno and Vegas TCP ...
ย 
TCP_Congestion_Control.ppt
TCP_Congestion_Control.pptTCP_Congestion_Control.ppt
TCP_Congestion_Control.ppt
ย 
HIGH SPEED NETWORKS
HIGH SPEED NETWORKSHIGH SPEED NETWORKS
HIGH SPEED NETWORKS
ย 
DCN 5th ed. slides ch24 Transport-Layer Protocols.pdf
DCN 5th ed. slides ch24 Transport-Layer Protocols.pdfDCN 5th ed. slides ch24 Transport-Layer Protocols.pdf
DCN 5th ed. slides ch24 Transport-Layer Protocols.pdf
ย 

More from IOSR Journals

More from IOSR Journals (20)

A011140104
A011140104A011140104
A011140104
ย 
M0111397100
M0111397100M0111397100
M0111397100
ย 
L011138596
L011138596L011138596
L011138596
ย 
K011138084
K011138084K011138084
K011138084
ย 
J011137479
J011137479J011137479
J011137479
ย 
I011136673
I011136673I011136673
I011136673
ย 
G011134454
G011134454G011134454
G011134454
ย 
H011135565
H011135565H011135565
H011135565
ย 
F011134043
F011134043F011134043
F011134043
ย 
E011133639
E011133639E011133639
E011133639
ย 
D011132635
D011132635D011132635
D011132635
ย 
C011131925
C011131925C011131925
C011131925
ย 
B011130918
B011130918B011130918
B011130918
ย 
A011130108
A011130108A011130108
A011130108
ย 
I011125160
I011125160I011125160
I011125160
ย 
H011124050
H011124050H011124050
H011124050
ย 
G011123539
G011123539G011123539
G011123539
ย 
F011123134
F011123134F011123134
F011123134
ย 
E011122530
E011122530E011122530
E011122530
ย 
D011121524
D011121524D011121524
D011121524
ย 

Recently uploaded

Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businesspanagenda
ย 
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data DiscoveryTrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data DiscoveryTrustArc
ย 
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CVReal Time Object Detection Using Open CV
Real Time Object Detection Using Open CVKhem
ย 
Ransomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdfRansomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdfOverkill Security
ย 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Miguel Araรบjo
ย 
Navi Mumbai Call Girls ๐Ÿฅฐ 8617370543 Service Offer VIP Hot Model
Navi Mumbai Call Girls ๐Ÿฅฐ 8617370543 Service Offer VIP Hot ModelNavi Mumbai Call Girls ๐Ÿฅฐ 8617370543 Service Offer VIP Hot Model
Navi Mumbai Call Girls ๐Ÿฅฐ 8617370543 Service Offer VIP Hot ModelDeepika Singh
ย 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc
ย 
AXA XL - Insurer Innovation Award Americas 2024
AXA XL - Insurer Innovation Award Americas 2024AXA XL - Insurer Innovation Award Americas 2024
AXA XL - Insurer Innovation Award Americas 2024The Digital Insurer
ย 
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu SubbuApidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbuapidays
ย 
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...
"I see eyes in my soup": How Delivery Hero implemented the safety system for ..."I see eyes in my soup": How Delivery Hero implemented the safety system for ...
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...Zilliz
ย 
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...apidays
ย 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProduct Anonymous
ย 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024The Digital Insurer
ย 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...Martijn de Jong
ย 
FWD Group - Insurer Innovation Award 2024
FWD Group - Insurer Innovation Award 2024FWD Group - Insurer Innovation Award 2024
FWD Group - Insurer Innovation Award 2024The Digital Insurer
ย 
Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)wesley chun
ย 
Manulife - Insurer Transformation Award 2024
Manulife - Insurer Transformation Award 2024Manulife - Insurer Transformation Award 2024
Manulife - Insurer Transformation Award 2024The Digital Insurer
ย 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MIND CTI
ย 
DBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor PresentationDBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor PresentationDropbox
ย 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century educationjfdjdjcjdnsjd
ย 

Recently uploaded (20)

Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
ย 
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data DiscoveryTrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
ย 
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CVReal Time Object Detection Using Open CV
Real Time Object Detection Using Open CV
ย 
Ransomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdfRansomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdf
ย 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
ย 
Navi Mumbai Call Girls ๐Ÿฅฐ 8617370543 Service Offer VIP Hot Model
Navi Mumbai Call Girls ๐Ÿฅฐ 8617370543 Service Offer VIP Hot ModelNavi Mumbai Call Girls ๐Ÿฅฐ 8617370543 Service Offer VIP Hot Model
Navi Mumbai Call Girls ๐Ÿฅฐ 8617370543 Service Offer VIP Hot Model
ย 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
ย 
AXA XL - Insurer Innovation Award Americas 2024
AXA XL - Insurer Innovation Award Americas 2024AXA XL - Insurer Innovation Award Americas 2024
AXA XL - Insurer Innovation Award Americas 2024
ย 
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu SubbuApidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
ย 
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...
"I see eyes in my soup": How Delivery Hero implemented the safety system for ..."I see eyes in my soup": How Delivery Hero implemented the safety system for ...
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...
ย 
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
ย 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ย 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
ย 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
ย 
FWD Group - Insurer Innovation Award 2024
FWD Group - Insurer Innovation Award 2024FWD Group - Insurer Innovation Award 2024
FWD Group - Insurer Innovation Award 2024
ย 
Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)
ย 
Manulife - Insurer Transformation Award 2024
Manulife - Insurer Transformation Award 2024Manulife - Insurer Transformation Award 2024
Manulife - Insurer Transformation Award 2024
ย 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024
ย 
DBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor PresentationDBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor Presentation
ย 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century education
ย 

Performance Evaluation of High Speed Congestion Control Protocols

  • 1. IOSR Journal of Computer Engineering (IOSRJCE) ISSN: 2278-0661 Volume 3, Issue 1 (July-Aug. 2012), PP 13-19 www.iosrjournals.org www.iosrjournals.org 13 | Page Performance Evaluation of High Speed Congestion Control Protocols Mohamed Ali Mani and Rachid Mbarek ABSTRACT: Computer networks are facing several technological and scientific challenges. Among these challenges is congestion control. It limits the quantity of information input at a rate less important than that of the transmission one to ensure good performance as well as protect against overload and blocking of the network. Researchers have done a great deal of work on improving congestion control protocols, especially on high speed networks. In this paper, we will be studying the congestion control alongside low and high speed congestion control protocols. We will be also simulating, evaluating, and comparing seven of high speed congestion control protocols : Bic TCP, Cubic TCP, Hamilton TCP, HighSpeed TCP, Illinois TCP, Scalable TCP and YeAH TCP, with multiple flows. KEYWORDS: HIGH-SPEED, TCP PROTOCOLS, CONGESTION CONTROL, PERFORMANCE, MULTIPLE FLOWS. I. INTRODUCTION Due to the ease of finding information, data access and utilization by a simple mouse move in multimedia pages, Internet has become necessary in our daily lives. In fact, the number of Internet users is growing rapidly, and according to the ITU (International Telecommunication Union) it reached only 500 million users in the beginning of 2000s while it was 1.6 billion in 2008 to rise to the number of 2 billions in 2011. This increase must undergo a parallel development of both of hardware and software at the same time. Early, throughput was about 56 kbps, however, now we talk about Gigabit Internet aiming to meet the usersโ€™ needs mainly in astronomy and video conference fields. In Internet, there are some machines responsible for transmitting data packets, and other that take off these packets from the queue. If the sender machine sends a data packet rate much more important than the rate of the receiver, network congestion is produced, hence, a congestion control will limit the quantity of input information with a lower rate than the transmission one to guarantee a good performance as well as a network protection against overloading and blocking. The high number of high speed congestion control protocols led us to prepare this research work, which focuses on evaluating and comparing high speed congestion control protocols. This paper is organized as follows: In the second section we study the state of the art. In the third one, we present the architectures used as well as the curves and the performances evaluation for different high speed congestion control protocols. STATE OF THE ART Researchers have worked on the enhancement of high speed congestion control protocols. Practically every year, one or two protocols are implemented having for each one of them its own specific strengths and weaknesses. Recently, the research works are interested in evaluating the performance of these protocols by comparing among 2 to 5 protocols with 1 to 12 flows [1] [2] [3] [4] [5]. HS-TCP [6]: If the congestion window is low, HS-TCP behaves just like the standard TCP, once the congestion window exceeds the max_ssthresh, it rises in an aggressive way. The algorithm of HS-TCP is an AIMD Additive Increment cwnd = cwnd + a(cwnd) Multiplicative Decrement cwnd = (1- b(cwnd)) x cwnd, the values used for HSTCP are a(cwnd) in the interval of [1,72], b(cwnd) in the interval of [0.1,0.5] and low_window equals 38 packets. when cwnd is lower than low_window a(cwnd) = 1 and b(cwnd) = 0.5 thus the algorithm acts just like a standard TCP, in the case of a packet loss, the cwnd is reduced to 50%, when the cwnd reaches the high_window a(cwnd ) = 72 and b(cwnd)=1. If there is an acknowledgment: ๐‘ค1 = 1 + ฮฒ 2 ร— cwnd cwnd < ๐‘ค1 ๐‘ค1 = cwnd cwnd โ‰ฅ ๐‘ค1 w2 = cwnd Scalable TCP [7]: For this protocol the cwnd increases with 0.001, if a congestion is detected the cwnd reduces with 0.125 independently from the throughput and the packet size. This algorithm is a MIMD; Multiplicative
  • 2. Performance Evaluation Of High Speed Congestion Control Protocols www.iosrjournals.org 14 | Page Increase since the cwnd increases with 0.01 per acknowledgment, Therefore there is a total increase of 0.01 cwnd per RTT, and Multiplicative Decrease since the cwnd decreases with 0.125 per packet loss. BIC TCP [8]: uses a function that allows to have a rapid increase of cwnd when its value is far from the maximal threshold Smax and to slowly decrease when having a value close to w_1. cwnd = (1 โ€“ ฮฒ) ร— cwnd With : fฮฑ (ฮด, cwnd) = ฮฒ ฮด ฮด โ‰ค 1 and cwnd < w1 or w1 โ‰ค cwnd < w1 + ฮฒ ฮด 1 < ฮด โ‰ค Smax and cwnd < w1 w1 ฮฒโˆ’1 ฮฒ โ‰ค cwnd โˆ’ w1 < Smax 1 โˆ’ ฮฒ Smax otherwise cwnd : Current congestion window Smax : Maximal threshold Smin : Minimal threshold ฮ’ : Decrease factor usually equals 0.875 CUBIC TCP [9]: The name CUBIC refers to the function of the congestion window increase which is cubic, when receiving an ACK: cwnd = C (T โ€“ K) ยณ + Maxcwnd When having a packet loss: cwnd = ฮฒ Maxcwnd With : HTCP [10]: uses ฮ” the duration since the last congestion rather than cwnd as information about the product delay bandwidth BDP. The increase parameter of AIMD (additive-increase/multiplicative-decrease) varies depending on ฮ”. It also depends on the value of RTT which has provoked the unfairness among the competitor flows with different RTTs. In case of receiving an ACK: cwnd = cwnd + 2 1โˆ’ฮฒ fฮฑ(ฮ”) cwnd In case of packet loss: cwnd = gฮฒ(B) ร— cwnd With : fฮฑ(ฮ”) = 1 ฮ” โ‰ค ฮ”L max fฮฑ ฮ” Tmin ,1 ฮ” > ฮ”L gฮฒ(B) = 0.5 B K+1 โ€“ B(k) B(K) > ฮ”B min Tmin Tmax ,0.8 otherwise cwnd*:*Current congestion window ฮฑ*:*Increase parameter ฮฒ*:*Decrease parameter ฮ”*:*Duration of time since the last congestion ฮ”L*:*Threshold to change the mode low speed to highspeed Tmin *:*Minimal duration of sending and receiving packets Tmax *:*Maximal duration of sending and receiving packets B K + 1 *:*Maximal throughput reached before a congestion fฮฑ ฮ” *=*1 +10 (ฮ” - ฮ”L) + 0,25 (ฮ” - ฮ”L)ยฒ Cwnd*:*Current congestion window C*:*Constant T*:*Duration of time since the last congestion Maxcwnd *:*Size of the last congestion window K*:* Maxcwnd ฮฒ/C3
  • 3. Performance Evaluation Of High Speed Congestion Control Protocols www.iosrjournals.org 15 | Page Illinois TCP [11] : Similar to standard TCP, Illinois-TCP increases the congestion window by using two parameters ฮฑ and ฮฒ, it is based on packet loss to define the congestion window value, however the values of ฮฑ and ฮฒ are not constant by using the delays. When receiving an ACK: cwnd = cwnd + ฮฑ ๐‘๐‘ค๐‘›๐‘‘ When losing packets: cwnd = (1- ฮฒ) ร— cwnd With : ฮฑ = ๐‘“1(๐‘‘ ๐‘Ž ) = ฮฑ ๐‘š๐‘Ž๐‘ฅ ๐‘‘ ๐‘Ž โ‰ค ๐‘‘1 ๐‘˜1 ๐‘˜2+ ๐‘‘ ๐‘Ž ๐‘œ๐‘กโ„Ž๐‘’๐‘Ÿ๐‘ค๐‘–๐‘ ๐‘’ ฮฒ = ๐‘“2(๐‘‘ ๐‘Ž ) = ฮฒ ๐‘š๐‘–๐‘› ๐‘‘ ๐‘Ž โ‰ค ๐‘‘2 ๐‘˜3 + ๐‘˜4 ๐‘‘ ๐‘Ž ๐‘‘2 < ๐‘‘ ๐‘Ž < ๐‘‘3 ฮฒ ๐‘š๐‘Ž๐‘ฅ ๐‘œ๐‘กโ„Ž๐‘’๐‘Ÿ๐‘ค๐‘–๐‘ ๐‘’ ๐‘˜1 = (๐‘‘ ๐‘š โˆ’ ๐‘‘1) ฮฑ ๐‘š๐‘–๐‘› ฮฑ ๐‘š๐‘Ž๐‘ฅ ฮฑ ๐‘š๐‘Ž๐‘ฅ โˆ’ ฮฑ ๐‘š๐‘–๐‘› * ๐‘˜2 = (๐‘‘ ๐‘š โˆ’ ๐‘‘1) ฮฑ ๐‘š๐‘–๐‘› ฮฑ ๐‘š๐‘Ž๐‘ฅ โˆ’ ฮฑ ๐‘š๐‘–๐‘› - ๐‘‘1 ๐‘˜3 = ฮฒ ๐‘š๐‘–๐‘› ๐‘‘3โˆ’ ฮฒ ๐‘š๐‘Ž๐‘ฅ ๐‘‘2 ๐‘‘3โˆ’ ๐‘‘2 *๐‘˜4 = ฮฒ ๐‘š๐‘Ž๐‘ฅ โˆ’ ฮฒ ๐‘š๐‘–๐‘› ๐‘‘3โˆ’ ๐‘‘2 ฮฑ ๐‘š๐‘–๐‘› = ๐‘˜1 ๐‘˜2+ ๐‘‘ ๐‘š๐‘œ๐‘ฆ , ฮฑ ๐‘š๐‘Ž๐‘ฅ = ๐‘˜1 ๐‘˜2+ ๐‘‘1 ฮฒ ๐‘š๐‘–๐‘› = ๐‘˜3 + ๐‘˜4 ๐‘‘2 , ฮฒ ๐‘š๐‘Ž๐‘ฅ =๐‘˜3 + ๐‘˜4 ๐‘‘3 YeAH TCP [12] the algorithm starts with the Slow Start if cwnd < ssthresh and Scalable TCP once cwnd reaches the ssthresh. This protocol uses two modes, fast mode and slow mode. In the case of the fast one, the congestion window increases aggressively, for the second mode, the protocol behaves similarly to Reno TCP. The mode is chosen according to the queue status. Having RTTBASE the minimal RTT measured by the sender and RTTMIN the minimal RTT estimated from the current window. The estimated delay is: RTTqueue = RTTmin โˆ’ RTTbase Thanks to this value, the number of packet on hold in the queue can be defined as follows: Q = RTTqueue ร— G = RTTqueue ร— cwnd RTTmin And G is the throughput. II. SIMULATION AND PERFORMANCES EVALUATION 2.1. Architecture In order to simulate the high speed congestion control protocols, we have chosen a topology (as shown in figure 1) composed of a sender and a receiver linked together with two routers by a line of 1Gbps of bandwidth, the delay is 1ms. The routers are linked to each other with a line having a bandwidth of 200Mbps, the delay is equal to 92 ms and the queue capacity is exactly 100 packets [13]. The MSS size equals 1460 bytes. The differences among the bandwidths capacities provoke congestion. Figure 1. Basic topology To evaluate the performances, we have used 2, 4, 6, 8, 10, 12, 16 and 24 identical flows (as shown in figure 2) with a link capacity of 1 Gbps and a delay of 1 ms. Figure 2. Topology with multiple flows 2.2. TCP-Friendly To test the TCP Friendly of different protocols, we have used a multiple flows topology with 4 input flows triggered at the same time. The results are shown in the following graphs:
  • 4. Performance Evaluation Of High Speed Congestion Control Protocols www.iosrjournals.org 16 | Page Figure 3.a. HS TCP : Throughput variation for 4 flows * Figure 3.b. Scalable TCP : Throughput variation for 4 flows Figure 3.c. Bic TCP : Throughput variation for 4 flows * Figure 3.d. Hamilton TCP : Throughput variation for 4 flows Figure 3.e. Cubic TCP : Throughput variation for 4 flows * Figure 3.f. Illinois TCP : Throughput variation for 4 flows Figure 3.g. YeAH TCP : Throughput variation for 4 flows
  • 5. Performance Evaluation Of High Speed Congestion Control Protocols www.iosrjournals.org 17 | Page 2.3. Efficiency The efficiency (rate of utilization or performance) of a network is the percentage of utilization [14]. Mathematically, it is the division of the average throughput by the optimal throughput. An efficient network uses the maximum of capacity. The following variable ๐‘ž๐‘– is used as the average throughput of a flow i. The throughput of a source with n flows is: ๐‘„ = 1 ๐‘› ๐‘ž๐‘– ๐‘› ๐‘–=1 We calculate the ratio between the average throughput and the optimal one which is the result of an ideal network performance: ๐ธ = ๐‘„ ๐‘„ ๐‘œ๐‘๐‘ก We have used these two topologies to simulate the seven protocols for 1, 2, 4, 6, 8, 10, 12, 16 and 24 flows. The efficiency calculus results are as mentioned in the following graph: Figure 4. Efficiency for different flow numbers 2.4. Fairness The fairness is the attempt of sharing the network capacities among users in a fair way. For the purpose of measuring the fairness, one method is used in the networks field which is called the Maximin law proposed by Raj Jain [15]. Here is the procedure that allows us to calculate the fairness of a proposed algorithm: Having an algorithm that provides the distribution vi = [x1, x2,โ€ฆ,xn] instead of the optimal distribution vopt = [x1,opt , x2,opt , โ€ฆ , xn,opt ]. We calculate the standardized distribution for every source as follows: ๐‘‹๐‘– = ๐‘ฅ๐‘– ๐‘ฅ ๐‘œ๐‘๐‘ก Thus, the fairness index F equals the sum of distributions squared and divided by the square of sums: ๐น = ( ๐‘‹๐‘– ๐‘› ๐‘–=1 )ยฒ ๐‘› ๐‘‹๐‘–ยฒ๐‘› ๐‘–=1 We have used these two topologies to simulate the seven protocols for 1, 2, 4, 6, 8, 10, 12, 16 and 24 flows. The fairness calculus results are as mentioned in the following graph: Figure 5. Fairness for different flow numbers 2.5. Performance The performance of a congestion control algorithm is the relation between efficiency and fairness: Performance = ฮฑ ร— E + (1 โ€“ ฮฑ) ร— F With ฮฑ = [0,1.. 0,9], E the efficiency and F the fairness. For the different algorithms, we have calculated the performances for a network rather efficient with ฮฑ = 0,8, a network rather fair with ฮฑ = 0,2 and a balanced network when ฮฑ = 0,5. The results for 1, 2, 4, 6, 8, 10, 12, 16 and 24 flows are as follows:
  • 6. Performance Evaluation Of High Speed Congestion Control Protocols www.iosrjournals.org 18 | Page Figure 6.a. Performance for different flow numbers (ฮฑ = 0,8) Figure 6.b. Performance for different flow numbers (ฮฑ = 0,5) Figure 6.c. Performance for different flow numbers (ฮฑ = 0,2) 2.6. Link Utilisation per RTT We varied the RTT for different congestion control algorithms to calculate the link utilization. The used RTT values are: 20 ms, 60 ms, 100 ms, 140 ms, 180 ms and 220 ms. We used the basic topology with only one flow; the results are described in the following graph: Figure 7. Link utilization for different RTTs 2.7. Link utilization for different queue sizes Certainly the queue size has an important impact on the efficiency of the network, but which algorithm benefits the more of this change? We varied the queue size for different congestion control algorithms to calculate the link utilization. The used sizes are: 50, 100, 150, 200, 250 and 300 packets. We used the basic topology with one flow, the results are as follows: Figure 8. Link utilization for different queue sizes III. CONCLUSIONS The high number of congestion control algorithms never satisfied the researchers. Since this field is always having changes and enhancements due to the usersโ€™ needs and the evolution of both of hardware and software in Telecommunications. Indeed, it is difficult to conceive an algorithm that gives satisfying results for all the architectures.
  • 7. Performance Evaluation Of High Speed Congestion Control Protocols www.iosrjournals.org 19 | Page During this research work, we simulated seven high speed congestion control protocols for 1, 2, 4, 6, 8, 10, 12, 16 and 24 flows. We evaluated their performances by calculating the efficiency, the fairness as well as the performance while varying the RTT and the queue size. Basing on the simulation and the high speed protocols performance evaluation, we can conclude that: โ€ข Some protocols perform well in some defined cases, but weak in others. โ€ข The network architecture has got an important impact on the protocols performance. โ€ข For 24 flows, all protocols are unfair. โ€ข For 24 flows, the protocols: Bic TCP, Cubic TCP, Hamilton TCP, Scalable TCP and YeAH TCP are efficient. โ€ข Scalable is not only TCP-friendly, but also selfish, since the first flow gets the most important throughput, and doesnโ€™t leave its place for any other flow such as the case of Cubic TCP, Hamilton TCP and Illinois TCP. โ€ข The protocols Bic TCP, HighSpeed TCP and YeAH TCP are TCP-friendly โ€ข For low values of RTT, all the protocols use the network capacities in a bad way. โ€ข For low values of the queue size, Cubic TCP is the most efficient. โ€ข YeAH TCP and Illinois TCP give the best results while increasing the queue size. โ€ข Although his aggressiveness, YeAH TCP is quite TCP-friendly. It is also quite performant, but it doesnโ€™t suit all the architectures. Further work is required to evaluate the performances of these protocols by using Relentless TCP which will be the topic for a multitude of research works in the near future, particularly to define if the protocols that have more stability and take more time before falling another time in a new congestion such as Cubic TCP and Illinois TCP in order to check if they will well exploit the law that forms the basis of Relentless TCP which is the reduction of the congestion window with the number of lost segments. It will be also interesting to create a model that changes dynamically the congestion control protocols used in terms of flow number to better exploit the network capacities as well as studying the QoS. REFERENCES [1] R. Mbarek, M. T. BenOthman and S. Nasri. โ€œFairness of High-Speed TCP Protocols with Different Flow Capacitiesโ€. Journal Of Networks, Vol. 4, No. 3, pp. 163-169, May 2009. [2] S. Hemminger. โ€œTCP testing โ€“ Preventing global Internet warmingโ€. LinuxConf Europe, Sep. 2007. [3] S. Ha, L. Le, I. Rhee and L. Xu. โ€œImpact of Background Traffic on Performance of High-Speed TCP Variant Protocolsโ€. The International Journal of Computer and Telecommunications Networking, Vol.51, May, 2007. [4] M. Nabeshima and K. Yata. โ€œPerformance Evaluation and Comparison of Transport Protocols for Fast Long-Distance Networksโ€. IEICE Trans. Commun., Vol.E89-B, No.4, pp. 1273-1283, Apr. 2006. [5] M. Weigle, P. Sharma and J. Freeman. โ€œPerformance of Competing High-Speed TCP Flowsโ€. Networking 2006, LNCS 3976, pp. 476-487, 2006. [6] S. Floyd. โ€œHigh-Speed TCP for Large Congestion Windowsโ€. RFC 3649, Experimental, Dec. 2003. [7] T. Kelly. โ€œScalable TCP : Improving Performance in High-Speed Wide Area Networksโ€. In Proceedings of PFLDnet, Geneva, Switzerland, Feb. 2003. [8] L. Xu, K. Harfoush, and I. Rhee. โ€œBinary Increase Congestion Control for Fast, Long Distance Networksโ€. In Proceedings of IEEE INFOCOM, Hong Kong, Mar. 2004. [9] I. Rhee and L. Xu. โ€œCUBIC: A New TCP-Friendly High-Speed TCP Variantโ€. In Proceedings of PFLDnet, Lyon, France, Feb. 2005. [10] R. N. Shorten and D. J. Leith. โ€œH-TCP : TCP for high-speed and long-distance networksโ€. In Proceedings of PFLDnet, Argonne, Feb. 2004. [11] S. Liu, T. Basar and R. Srikant.,"TCP-Illinois : A Loss- and Delay-based Congestion Control Algorithm for High-Speed Networks", Performance Evaluation, 65(6-7) : 417-440, Mar. 2008. [12] A. Baiocchi, A. P. Castellani and F. Vacirca. โ€œYeAH-TCP: Yet Another Highspeed TCPโ€. PFLDnet, Roma, Italy, Feb. 2007. [13] J. Semke, J. Mahdavi, and M. Mathis. "Automatic TCP Buffer Tuning", Computer Communications Review, a publication of ACM SIGCOMM, Vol. 28, No. 4, Oct. 1998. [14] R. Mbarek, M. T. BenOthman and S. Nasri. โ€œPerformance Evaluation of Competing High-Speed TCP Protocolsโ€. IJCSNS International Journal of Computer Science and Network Security, VOL.8 No.6, pp. 99-105, June 2008. [15] R. Jain. โ€œCongestion Control and Traffic Management in ATM Networks : Recent Advances and A Surveyโ€. Computer networks and ISDN Systems, Nov. 1995. Authors M. Mohamed Ali Mani received his M.S degree from ISITCOM, Hammam Sousse, Sousse University Tunisia, in 2011. He is currently a lecturer in the Department of Computer Science in ISTLS in the same University. His research interests lie in the general area of computer networks, including high speed transports protocols, flow control, congestion control in high-speed networks and wireless networks. Dr. Rachid Mbarek received his Ph.D from the College of Engineering (ENIS), Sfax University, Tunisia, received his M.Sc. degree in Computer Engineering in 2003 from the same University. He is currently a Lecturer in the Department of Telecommunication in the ISITCOM at Sousse University, Tunisia. His research interests lie in the general area of computer communication networks, including high-speed transports protocols, flow control and congestion control in high-speed networks, network performance analysis, peer to peer networking.