Comparing the Impact of Bandwidth and Congestion on · PDF fileComparing the Impact of...
Transcript of Comparing the Impact of Bandwidth and Congestion on · PDF fileComparing the Impact of...
Comparing the Impact of Bandwidth and Congestion on Selection of
Heterogeneous Lines Applying Fuzzy-Genetic
Ebrahim saboktakin Rizi 1, MohammadR.Reshadinezhad2, *, Naser Nematbakhsh3
1Student of Master Degree, Department of Computer Engineering, Islamic Azad University, Najaf Abad Branch, Najaf Abad, Iran.
2Assistant professor, Department of Computer Engineering, University of Isfahan, Isfahan
8174673440, Iran. 3Assistant professor, Department of Computer Engineering, University of Isfahan, Isfahan
8174673440, Iran.
Abstract
A major challenge to determine the suitable
amount of submitted data for each line in
heterogeneous lines is considering the important
features of each line, including bandwidth and
congestion. According to the indeterminate
character of each line in sending data, selection of
the best sending lines in SCTP_CMT, where the
possibility of failure in less could be considered as
a significant step in improving the service quality
and speed of data transmission. The objective of this article is to review the bandwidth and
congestion among heterogeneous lines in
SCTP_CMT to evaluate the effects of these features
on the amount of the selected data per line. In here,
a fuzzy-genetic analysis method is designated by
applying the designed data set. The effects of
congestion and bandwidth characteristics in
SCTP_CMT are evaluated. The obtained results
are applied to provide intelligent models in
selecting SCTP_CMT lines and speed up data
transfer.
1. Introduction Communications and communication networks
are considered as one of the major concerns in the
digital systems, performance. Communication
protocols, and their performance may have a
substantial impact on the communication systems,
performance. In the recent years, a variety of
protocols such as TCP / IP [1] and UDP [2] has
been introduced, each with its advantages and disadvantages. Among the newly protocols for
communication systems, one can refer to SCTP
protocol [3] which proposed the idea of selecting
data from several alternate routes, [4] with
additional properties such as the multi-house
characteristic and CMT [5] which can send data to
multiple paths simultaneously. In SCTP_CMT,
when there are heterogeneous data lines with
different characteristics, including the bandwidth
and congestion, challenges such as how to select
the best lines in a data transfer comes to mind [6].
So far, there are formal methods in selecting the
best lines in SCTP_CMT heterogeneous lines, and
most of the studies in this field try to use random
choice and the rotational lines methods for data transmission at the same time on multi paths. For
better understanding of data transmission from
multiple heterogeneous SCTP_CMT paths, Figure
1 presents an example on data transferring trough
the mentioned protocol in [7].
As seen in figure1, line types include Wireless
and Mobile with different properties. In this article
understanding the effects of the important features
and reviewing the most important features that can
select a line to transmit data simultaneously from
multiple directions are studied. To review and
assess the characteristics of the lines in
SCTP_CMT with heterogeneous lines, the attempt
is made by the authors to design a genetic fuzzy
analysis method. In this regard; here adapting the
analytical designed method and using the derived
random data sets to evaluate the characteristics of broadband and congestion properties of lines, the
analysis and calculation of the effects of each of the
listed properties are made. The results obtained can
provide a method for non-homologous SCTP_CMT
lines for enhancing the speed and quality of
SCTP_CMT protocol services.
2. Related Works
Several studies have investigated the challenges
of SCTP_ CMT, including [9-8]. Few studies have
examined the selected lines problem.
Classifications of the studies conducted on lines
selection challenges in SCTP_CMT protocol with heterogeneous lines are presented in table 1.
Mohammad R Reshadinezhad et al, Int.J.Computer Technology & Applications,Vol 5 (3),1161-1167
IJCTA | May-June 2014 Available [email protected]
1161
ISSN:2229-6093
Fuzzy Logic was first introduced by Prof.
Zadeh in 1968 and has already undergone
enormous changes [12].The most important idea in
using fuzzy logic, is designing a fuzzy system,
which was introduced in 1978 by Mamdany and
Sugeno for the first time [13]. In order to design the
parameters of the fuzzy system in the computer
engineering field, various methods have been
proposed, in Table (2).
Figure 1. Simultaneous data transferring at
SCTP_CMT
3. Description of the analytical method of
fuzzy-genetic
In this section, the analysis the characteristics
that influence the choice of lines to transmit data
from multiple SCTP_CMT paths with
heterogeneous lines, is presented based on the
analysis of fuzzy logic and genetic algorithms.
Section 3.1 reviews the concepts of fuzzy logic and
fuzzy systems. in Section 3.2, a summary of the
application of genetic algorithms is expressed. The
final section outlines the research methodology
proposed for analysis of fuzzy systems using fuzzy
logic and genetic algorithms.
Table 1. The most important investigations in line selection issue
Researcher
The issue
Qiao[10]
Effect of broadband lines factors on
heterogeneous wireless networks to
transmit in SCTP_CMT
Dreibhoz[11] CMT Evaluation of simultaneously data
transferring from a heterogeneous multi-
path
Adhari[7] Evaluation of Challenges of CMT in
heterogeneous lines
Table 2. The most important proposed methods for fuzzy systems
Provider
Design Method
Pham [14] Optimization by Genetic
Algorithm
Reddy [15] Extraction of fuzzy
parameters by PSO
Jang [16] Fuzzy parameters
optimization using fuzzy-
neural
Bontoux [17] Fuzzy parameters
optimization using ant colony
algorithm
3.1. Fuzzy Logic One of the new technologies which is
widespread in all fields, of science and engineering,
and the humanities, is fuzzy logic. The most
important application of fuzzy logic is in cases
where there is uncertainty for decision making and
optimization become clearer. In the recent years,
the adoptions of fuzzy logic grow in engineering
sciences, with a wide usage. The fuzzy systems are
among the most important applications of fuzzy
logic, developed by Mamdany and Sugeno (2008). They used flexibility in decision making and
management control systems, which made an
important step in the quality and efficiency in the
engineering fields. Fuzzy system is composed of
several components, the most important of which is
the inference engine using the fuzzy rules that
make decisions for control systems. The most
important flow chart of the components in a fuzzy
decision-making control systems are presented in
Figure2 [18].
Mohammad R Reshadinezhad et al, Int.J.Computer Technology & Applications,Vol 5 (3),1161-1167
IJCTA | May-June 2014 Available [email protected]
1162
ISSN:2229-6093
Figure 2. The main components of the
fuzzy system designed by SUGENO
For designing the optimal rules in fuzzy systems,
several models are proposed. in the related work
section.
3.2. Genetic Algorithm Over the past few years, researchers have
investigated the nature of the decision-making
mechanisms in order to understand and optimize
living creature's behaviour. This investigation was
conducted to provide heuristic and meta-heuristic
methods such as the genetics algorithm [19], ant
colony algorithm and colonial competition. Genetic
algorithm and its application in digital systems was
first proposed by John Holland in 1992 and since
then it has undergone extensive changes [20].
The genetic algorithm uses the idea of reproduction
and creates a better generation in living creatures,
acting on optimization and decision making
problems and challenges in science and
engineering, especially the digital systems. In order
to use the genetic algorithms in the optimization, several steps, including the initial generation, the
design of cost function and mutations should be
considered. One of the important points in efficient
genetic algorithm design is the selection of cost
function with genetic algorithm to solve the
problem. To solve the optimization problems a
flowchart is presented the below to designing the
genetic [21].
Figure 3. Flowchart of the genetic algorithm
3.3. Fuzzy- genetic analysis To study the impact of network congestion and
bandwidth characteristics of the selected lines in
SCTP_CMT with heterogeneous lines, in this article, using fuzzy-genetic analysis, the
specifications of [22] are examined. In order to
design fuzzy systems for analysis, the optimal rules
for fuzzy systems are required. Here the genetic
algorithm is used to optimize the rules. A better
understanding on the process of streamlining the
rules used in fuzzy systems is presented in figure 4
[23].
Figure 4. Optimization process rules
Using the evaluation data set presented in section
4.1 for genetic algorithms, the attempt is made to
derive the cost function and other parameters of the
genetic algorithm, including mutations and
enhancement functions through trial-and-error
method for extracting the most optimized rules
The inference
engine
Input
variable
s
Output
variable
s
F
u
z
z
y
f
a
i
r
D
e
f
u
z
z
y
f
a
i
r
Database
Database
Rules
Database
Mohammad R Reshadinezhad et al, Int.J.Computer Technology & Applications,Vol 5 (3),1161-1167
IJCTA | May-June 2014 Available [email protected]
1163
ISSN:2229-6093
required in this proposed fuzzy system and to
analyze the bandwidth and delay characteristics of
SCTP_CMT. The main parameters in the genetic
algorithm used in this study are presented in Table
3.
Table 3. The genetic algorithm settings used in this study
value parameter
Gaussian(0.1,0.1) Mutation
16 Population size
stochastic uniform Selection
5 Crossover fraction
2 Elite count
25 Generations
FIS Production Fitness
In order to become familiar with the way to convert
fuzzy system rules with chromosomes and genes
required in the form of a genetic algorithm Figure 5
of [23] is presented.
Figure 5. Mapping Fuzzy Control rules to chromosomes and genes in the genetic
algorithm
As shown in Figure 5, each one of the rules are
represented as a number and finally, every rule is
extracted in the form of chromosomes. By using
the genetic algorithms and parameters required in
designing these algorithms, here the process of selecting the best generations (rules) in the research
of genetic algorithm are presented in Figure 6 to
optimize the proposed fuzzy system rules .After
extracting the optimal rules by using a genetic
algorithm, the designing a fuzzy controller and
fuzzy-genetic analysis, take place. This proposed
fuzzy system design is performed in MATLAB
software [24]. Except the required rules where the
optimal values are calculated through a genetic
algorithm, other parameters required for the
proposed fuzzy system design are calculated
through the results of previous studies and trial-
and-error method.
Figure 6. Selection of the best rules for the
generation of a genetic algorithm
To understand this proposed fuzzy system, and the
fuzzy analysis of genetic in this study, the main
parameters used in this proposed fuzzy system are presented in Table (4). The optimal rules derived in
the inference engine of this proposed fuzzy system
by a genetic algorithm with fuzzy terms for every
time variable is discussed in [25] Figure 7.
Table 4. The parameters used in this
proposed fuzzy system
4. Evaluation of results
To review the methods of fuzzy-genetic
analysis, the required data sets in the evaluation are
discussed in section 4.1. In section 4.2, the
influence of the broadband factor on the choice of
lines in SCTP_CMT with heterogeneous lines is
discussed and in Section 4.3, this effect is
investigated with respect to congestion factor.
Finally, in Section 4.4 comparison of specifications
of each investigated factories is made.
Lin
gu
isti
c
va
ria
ble
s
Nu
mb
er
of
me
mb
ers
hip
fu
nc
tio
ns
Nu
mb
er
of
rule
s
Ty
pe
o
f m
em
be
rsh
ip
fun
cti
on
s
Ty
pe
d
es
ign
ed
co
ntr
oll
er
Bandwidth 3 27 GAUSS
MF
SUGENO
[26]
Line
congestion 3 27 GAUSS
MF
SUGENO
[26]
Mohammad R Reshadinezhad et al, Int.J.Computer Technology & Applications,Vol 5 (3),1161-1167
IJCTA | May-June 2014 Available [email protected]
1164
ISSN:2229-6093
Figure 7. Inference engine designed using optimized rules
4.1 Sets of evaluation data To analyse the characteristics listed in Section
3, the data sets for the evaluation and analysis of
fuzzy-genetic algorithm are designed and extracted.
The data sets presented here are used to calculate
the cost function in the genetic algorithm, and
examine the impact of congestion and bandwidth
characteristics of the SCTP_CMT with
heterogeneous lines. In order to extract the required
data sets data transferring must be calculated
through equation 4.1.
Z (X\Y)*0.01(D+F) (4.1)
Table 5. Conditions considered to extract
evaluating data sets
Where x is the bandwidth, y is the congestion, d is
the data and f is the data fail. This formula is based
on the amount of features from each line and should be used as the initial data set for designing a
cost function and fuzzy-genetic analysis. The initial
data collection for extracting the evaluation data
sets are presented in Table 5. In this study, the data
sets and fuzzy control of the proposed fuzzy-
genetic analysis are applied for designing the
system rules. The following sections examine the
results of the analysis for fuzzy-genetic
characteristics affecting the choice of transmitted
data on each line in SCTP_CMT with heterogeneous lines.
Then, by using the formula depicted in 4.1 buffer
amounts required per line minimum lost are
extracted and the obtained results are summarized
in Table 6.
4.2 Bandwidth Properties
The most important performance characteristics
of a data transferring in a multi-path system, is line
free bandwidth [6]. Higher bandwidth for each line indicates the ability to send more data through that
line without repetition again. Following the design
and analysis of genetic fuzzy provided in this
study, the amount of data needed to evaluate the
characteristics of bandwidth lines in SCTP_CMT
with heterogeneous lines are discussed. Here after
the simulation of this proposed fuzzy system and
design of optimize required rules of fuzzy systems
using a genetic algorithm in MATLAB software,
the influence of factors on the amount of bandwidth
required for each line are investigated and shown in
Figure 8.
Table 6. The amount of data required for every line by using equation 4.1
Free
bandwidth
line
congestion
Amount of
required data
0.02 0.5 0.04
0.04 0.07 0.57
0.02 0.3 0.06
0.03 0.9 0.03
0.07 0.25 0.28
0.2 0.03 6.66
0.45 0.09 5
0.3 0.02 15
0.8 0.4 2
0.02 0.02 1
0.9 0.07 12.86
Pe
rce
nt
of
the
a
ctu
al
am
ou
nt
of
req
uir
ed
da
ta
Pe
rce
nta
ge
of
co
ng
es
tio
n l
ine
Pe
rce
nta
ge
of
fre
e b
an
dw
idth
of
lin
e
Ty
pe
of
lin
es
he
tero
ge
ne
ou
s
lin
es
0.4 0.25 0.10 Wireless Line 1
2.33 0.15 0.35 ADSL Line 2
0.75 0.20 0.15 Mobile Line 3
0.17 0.30 0.05 Phone Line 4
Mohammad R Reshadinezhad et al, Int.J.Computer Technology & Applications,Vol 5 (3),1161-1167
IJCTA | May-June 2014 Available [email protected]
1165
ISSN:2229-6093
As seen in this figure, with an increase in the
bandwidth, the amount of data to be transmitted
over the data lines in SCTP_CMT are increased.
By using the results obtained in this section, the
average bandwidth impact factor on the amount of
data needed to transmit in each line in SCTP_CMT
is calculated (section 4.4).
Figure 8. The impact of bandwidth characteristics on required amount of data
per line
4.3 Characteristics of line congestion
Another important feature in this calculation is
the amount of data that each line can transmit with
respect to the congestion characteristics of the line.
Line congestion has almost a linear dependence on
bandwidth and with an increase and a decrease in
bandwidth the amount of delay will fluctuate [6].
The simulation is carried out using MATLAB
software to evaluate the characteristics of line
congestion for investigating the effect of this factor
on the amount of data required for each line to be transmitted in SCTP_CMT with heterogeneous
lines. The amount of transferred data in
SCTP_CMT lines in assessment data set is
proposed by considering the characteristics of line
congestion. As seen with an increase in line
congestion, in different time intervals in
simultaneous data transfer from multiple
heterogeneous paths in SCTP_CMT, the amount of
data required for each of the lines will have a
notable decrease in their slopes, and this fact
reflects the strong influence of this factor on the
amount of data required by each line in the
SCTP_CMT (figure 9).
Figure 9. Impact of line congestion characteristics on the amount of required data per line
4.4 Comparison of the features
According to the analytical results obtained in the
previous sections, here the average effects of each
of the features are extracted, and the importance of
features are evaluated. The average effect derived
in this section is based on the measured data sets in (4.1). The average impact factors extracted for
bandwidth and congestion in line on the amount of
data transmission in SCTP_CMT with
heterogeneous lines is presented. In table 7 and
graph 1.
Table 7. Comparing the effects of means bandwidth and congestion
Factor Average of impact
bandwidth 0.52
Network delay 0.38
Figure 10. Comparison of average effect of congestion and bandwidth features
As observed in table 7 and figure 10, after the
fuzzy-genetic analysis is designed in the article, in
order to review the most important features
influencing the data transfer in SCTP_CMT with
heterogeneous lines, bandwidth on average has a
25% greater impact on the amount of data required
for each line with respect to the congestion factor
of the line. The results obtained in here can be
assigned to provide parametric models of the
selection lines used in SCTP_CMT with
heterogeneous lines which would eventually lead to
Mohammad R Reshadinezhad et al, Int.J.Computer Technology & Applications,Vol 5 (3),1161-1167
IJCTA | May-June 2014 Available [email protected]
1166
ISSN:2229-6093
an increase in performance and the QOS in this
presented Protocol.
5. Conclusions and recommendations
One of the major facing the researchers in
SCTP_CMT protocol performance is the selection
of a line and the amount of data to be transmitted
for each line in SCTP_CMT. In order to present
models for selection of transferring data on each
heterogeneous line in SCTP_CMT, identifying the
most important factors influencing data transfer is
required for each line. Therefore, in this study by
using a fuzzy system design, the effects of
bandwidth and congestion line on the amount of
required data in SCTP_CMT with heterogeneous
lines are determined. For extracting fuzzy system parameters used in this study, the genetic
algorithms are used along with, the fuzzy-genetic
expression. After extracting the raw data set and
design of Fuzzy Systems, the assessments of
mentioned criteria are discussed. The findings here
indicate a higher impact of bandwidth on the
amount of required data in selecting each line in
SCTP_CMT with heterogeneous lines. Using the
results of this study in selecting parametric models
based on obtained factors mentioned here can
contribute to future work.
6. References [1] Andrew, L., Marcondes, C., Floyd, S., Dunn, L.,
Guillier, R., Gang, W., Eggert, L, Ha, S., Rhee, I. ,
"Towards a common TCP evaluation suite", In
PFLDnet,March 2008. [2] Stewart, R., Metz, C., "SCTP: new transport protocol
for TCP/IP", Internet Computing, IEEE,Vol. 6, pp. 64-
69, 2001. [3] Rajamani, R., Kumar, S. And Gupta, N., "SCTP
versus TCP: Comparing the performance of transport
protocols for web trac", Technical report, University of
Wisconsin-Madison,May 2002. [4] Natarajan, P ,.Ekiz, N., Amer, P., Stewart., R. ,
"CMT using SCTP multihoming: transmission policies
using a potentially-failed destination state," TR
2007/338, CIS Dept, U of Delaware,2007. [5] Natarajan, P., Iyengar, J., Amer, P.D., Stewart, R,
"Concurrent multipath transfer using transport layer
multihoming", Performance under network failures. In:
MILCOM, Washington, DC, USA, 2006.
[6] Biller, T., "The 10 factors for guaranteed network
marketing success", 2009.
[7] Adhari, H., Dreibholz, T., Becke, M., Rathgeb, E. P. ,
"Evaluation of concurrent multipath transfer over
dissimilar paths ,". [8] Malekpour, A., Jabalameli,H., Djamshid, T.,
"Concurrent multipath communication SCTP a novel
method for Multi-Path data transmission", I2TS´2010 -
Rio de Janeiro, Brazil pp. 13-15, Dec. 2010.
[9] Abd, A., Saadawi, T. And Lee, M. , "LS-SCTP: A
bandwidth aggregation technique for stream control
transmission protocol", Computer Communications, Special issue on Protocol Engineering for Wired and
Wireless Networks,Vol. 27, No. 10, June 2004.
[10] Qiao, Y., Et. Al., "Path selection of SCTP fast
retransmission in multi-homed wireless environments†",
2009.
[11] Dreibholz, T., Becke, M., Rathgeb, E.P., Txen, M,
"On the use of concurrent multipath transfer over
asymmetric paths", GLOBECOM , IEEE Global Telecommunications Conference,Vol .6, No. 10, pp. 1-6,
Dec. 2010.
[12] Zadeh, L. A., "Fuzzy algorithms", Info.
&Ctl.,Vol.12, pp.94-102, 1968.
[13] Mamdani, E. H. a. S. A., "An experiment in
linguistic synthesiswith a fuzzy logic controller",
International Journal of Man-Machine Studies,Vol. 7,
No. 1, pp. 1-13, 1975.
[14] Pham, D. T. a. K., D., "Optimum design of fuzzy
logic controllers using genetic algorithm", Journal of
Systems Engineering,Vol.1, pp.114-11 8 ,1991.
[15] Prasad Reddy, P. V., "Particle swarm optimization
in the fine-tuning of fuzzy software cost estimation
models", International Journal of Software Engineering
(IJSE),Vol.1, No.2, 2009.
[16] Jang, J. S. R., "ANFIS: Adaptive-network-
basedfuzzy inference systems", IEEE Transactions on Systems, Man, and Cybernetics,Vol. 23, No. 3, pp. 665-
685, May 1993. [17] Bontoux, B. a. F., D. , "Ant colony optimization for
the traveling purchaser problem", Computers &
Operations Research, In Press ,Corrected Proof, May
2006.
[18] Sugeno, M., "Industrial applications of fuzzy
control", Elsevier Science Pub. Co.,1985.
[19] Mukhopadhyay, D. M., Balitanas, M. O., Alisherov
F. A., Jeon, S. H. And Bhattacharyya, D., "Genetic algorithm: a tutorial review", International Journal of of
Grid and Distributed Computing,Vol.2, No.3, pp.25-32,
September 2009.
[20] Holland, J. H., Adaptation in natural and artificial systems, The University of Michigan Press,
ISBN.0262581116 1992 [21] Cord&Ona, O., Gomideb, F., Herreraa, F.,
Homannc, F. And Magdalenad, L. , "Ten years of genetic fuzzy systems: current framework and new
trends", Elsevier B.V,Vol.141, pp. 5-31,
Doi.10.1016/S0165-0114(03)00111-8, 2004.
[22] Luciano S´Anchez, O. C. O., Arnaud Quirin ,and
Krzysztof Trawinski, "Introducing a Genetic Fuzzy
Linguistic Combination Method for Bagging Fuzzy Rule-Based Multiclassification Systems", Fourth
Internationa Workshop on Genetic and Evolutioary
Fuuzy System,2010. [23] Herrera, F., "Genetic fuzzysystems: taxonomy, current research trends and prospects", Springer,Vol.1,
pp.27–46, Doi.10.1007/s12065-007-0001-5, 2008. [24] Higham, D. J., And Higham, N. J. , MATLAB Guide, To be published by SIAM, ISBN.0898715784,
2000.
[25] Mathworks, Fuzzy logic toolbox for use with
MATLAB, The MathWorks, Inc, 2002.
[26] Takagi, T., Sugeno, M., "Fuzzy identification of
systems and its applications to modelling and control",
IEEE Trans. Syst. Man Cybern,Vol.15, No.1, pp.116-
132, Doi. 10.1109/TSMC.1985.63133 99 , ISSN. 0018-
9472, 1985.
Mohammad R Reshadinezhad et al, Int.J.Computer Technology & Applications,Vol 5 (3),1161-1167
IJCTA | May-June 2014 Available [email protected]
1167
ISSN:2229-6093