International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)

Reducing Error in Packet Switched Wide Area Networks Using

a Modified Method
Ogbimi, Voke Francis
Department of Information Technology, University of Debrecen, Debrecen, Hungary
Abstract-- Modelling of discrete systems such as At the network level, it is limited by congestion, packet
communication networks have previously and even today loss, jitter and high latency (the effect of physical distance
receive much attention from different scientific communities between networks) [2]. At the application level,
in both academia and modern industries. Scientific, economic performance is further limited by natural behaviour of
and technical progress and the increasing complexity of these
application protocols especially with latency, packet loss
systems raise more new challenges and new problems for
researchers. In addition, since imperfections in the simulation and congestion at the network level.
process can be negative factors in time, cost and operational
efficiency in the development of a new or existing systems. It II. W IDE AREA NETWORKS (WAN)
is necessary to use effective and suitable methods that aid the A Wide Area Network is an interconnection of
researcher in the simulation process when they allow the
computers and computer related devices to perform many
evaluation of the performance of a new or existing system for
example, the evaluation new possible features in the system given functions, typically using local and long range
that must achieve new goals. telecommunication or networked systems. Wide area
A model was proposed and it solved the problem of error in Networks are typically used to send large data between
packet switched wide area networks by looking at the effect of endpoints and deliver to user electronic mail services, give
the number of erased channels on packet error probability of access to database systems and Internet. Wide Area
the schemes. In this problem the effect on the number of Networks also help with specialized operations in many
erasures on number of messages in relation to packet error fields such as manufacturing, medicine, navigation,
probability was investigated. Messages are always sent and education, entertainment and telecommunication. Because
acknowledged. When messages are sent and there is no
there are so many workstations (computer devices) in a
acknowledgement, it means that error has occurred.
Equations were presented to show the effect of erasures and to wide area network and are spread over large distances, a
minimize the error despite the number of erasures in the mesh design is required to route and transfer data across the
network. network.
All wide area networks are collection of two or more
Keywords: Error, Dynamic Erasure. Packet Switched, block types of equipment: a station and a node. A station is
Modified Method. Wide Area Networks a device that a user communicates with interacts with to
access networks and it contains the software applications
I. INTRODUCTION that allow someone to use the network for a particular
The data communication industry created a framework purpose.
that is capable of delivering cheap bandwidth in high A WAN Architecture
volumes in the 1990s during the expansion of the internet.
Bandwidth became surplus to the extent that even the effect The support structure of a wide area network is a part of
of Meltcalfe’s law (the network growth is inversely the network or network cloud. A network cloud is a
proportional to the number of users square in a given area) collection of nodes and interconnecting telecommunication
were inadequate to consume the available bandwidth for links. The type and number of interconnections between
years [1]. The result of bandwidth rapid expansion was nodes and the way network data is transferred from node to
bandwidth commercialization with decrease in prices and node are the responsibility of the network cloud as being
an environment that actively promoted the fantasy that high two different entities (devices). It helps to think of the wide
bandwidth can address almost any performance problem. area network and the network cloud as being two different
However, application performance in networks is affected entities. The wide area network is the entire system: the
by many factors associated with both network and nodes, the stations, the communications lines, the software,
application logic that must be addressed in order to achieve and perhaps even the users.
satisfactory application performance.

277
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
The network cloud is the virtually physical The results showed that Network coding with FEC was
interconnection of nodes and communications lines that highly profitable and that its throughput improved after
transfer data from between locations. All these components adding FEC to the system and has a higher throughput than
work in concert to create the network. A user at a SW with FEC but fell sharply with increased error rate. It
workstation and running a network application passes his was also found out that SW-NC-FEC rose more sharply
or her data to the network through a station, which passes when packet generation rate increased. The paper also
the data to the cloud. The network cloud gets the data to the found out that SW-NC-FEC produced more throughput as
designated node, which then delivers it to the appropriate vandermonde matrix was enlarged, while SW-NC
designated station. Clearly, there would be no network degraded.
without a network cloud, but it should not matter to the The paper [4] adopted ARQ and BCH Codes in
network what the inside of the network cloud looks like. developing an adaptive Error Control in Sensor Networks
The network cloud is the medium for getting the data from Using Coverage Area Information. The proposed adaptive
sender to destination. error control strategy was informational value. To evaluate
the informational value that uses coverage area
information, a novel approach was proposed. Packets from
sensor nodes in regions with low spatial density had higher
informational value. The adaptive schemes using ARQ and
BCH codes increased the reliability of packets with high
informational value when compared to static error control,
without a great increase in the energy consumption. In
future work they hope to analyse Different FEC and hybrid
FEC/ARQ strategies. The proposed strategies could be
used and adapted to several applications of sensor networks
to increase the reliability of messages considered more
important in the network.
The paper [5] analysed the Feedback Error in Automatic
Repeat request. They proposed a new method of
acknowledging packet delivery for unreliable feedback
channel conditions. The proposed method, dubbed BCF-
SAW, relied on backwards composite acknowledgement
and provided the retransmission protocols with
configurable ultra-reliability. It further provided the
scheduler of the wireless system with new degrees of
Figure 1: Diagram showing WAN Architecture in form of a network freedom to configure the communication link in order to
cloud meet the desirable reliability requirement even in highly-
B. Past Error Control Methods used in Wired and Wireless unstable feedback channel conditions. The presented
Networks numerical analysis showed orders of magnitude increase in
Lots of past research works carried out using Automatic reliability of the retransmission protocols over the practical
Repeat request Method. Among them are [3] which range of target block error rate only at the expense of a
maximized throughput of SW ARQ with Network Coding negligible increase in experienced average packet delay.
through Forward Error Correction. The paper investigated Arq Protocols Supporting Qos In Embedded Systems by
whether adding data recovery technique improved the [6], studied a framework to provide reliable transmission
performance network that used network coding. A using an ARQ scheme in a real-time manner for embedded
comparative analysis of throughput in a Stop and Wait networks. The framework was demonstrated through two
Automatic Repeat Request (SW_ARQ) data transmission different methods over a point to point link. Using EDF
with Network Coding (NC) and Forward Error Correction scheduling, the framework inserts retransmission packets
(FEC) was analysed. Programming Language developed calculating the delay bounds of the original transmission.
three discrete event simulations using Vandermonde Strategies were used to give a best effort on attempt to
matrix,: SW-ARQ without NC, SW-ARQ with NC and SW retransmit erroneous packets, trying to guarantee the
with NC and FEC. deadlines of original and retransmission packets.

278
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
The methods lead to a performance utilization of the The paper proposed a scalable heuristic scheme for
real-time communication by providing retransmissions and selecting a redundant path and classified the resulting
a better quality of service. Basically a lower message error redundant path length and disjointness according to their
rate were provided by adding one retransmission packet per characteristics for various internet like topologies. The
message while the promises of real-time deadlines are kept. simulation showed that for various topologies of the
Calculating the delay bounds of the messages, a Internet, only 10% of the participating nodes were needed
retransmission packet was scheduled in the channel. This for the proposed path redundant selection scheme to
way, a notable reduction of message error rate is achievable effectively find redundant path sharing few links which
with the cost of a reasonable overhead. The possibility and reduced packet loss by 66 percent.
schedulability of the presented methods have been [9] combined Forward Error Correction and Network
evaluated and some simulations were done to verify the Coding in Bufferless Networks using Optical Packet
strength of the methods in accepting real-time channels. Switching as a case study. A combined forward Error
The evaluation shows the performance of the methods in Correction and Network coding scheme was used to
accepting real-time channels in reality, and at last a address performance issues arising in a bufferless network.
comparison was been made between the two methods. A case study for reducing packet loss in optical packet
The paper [7] increased the information throughput switched networks was presented. The Analysis showed
efficiency over a stop and wait protocol by developing an that combining coding scheme has indeed increased
adaptive ARQ Strategy for Packet Switching Data successful decoding probability and reduced packet loss
Communications Networks. This improvement from the rate with orders of magnitude, if the underlying FEC
simulation results was particularly evident for a demonstrated sufficient redundancy. Factoring in realistic
communication system operating in an environment with contention probabilities in OPS networks, high rate, low
high channel Pb. The adaptive features to an existing over head Reed Solomon Codes were used to reduce packet
Software protocol was added and also designed only to be loss and meet both buffering and decoding delay
modified with no hardware requirements. Several different requirements simultaneously.
adaptive arq strategies were presented, and all adapted to The Paper [10] improved the efficiency of Forward Error
the packet length as the channel Pb increased or decreased. Correction coding in reducing packet loss in Internet
The noise or interference analysis or empirical data of the Protocol networks. A method was explored for measuring
specific system which reflected the condition required to the performance of FEC coding thereby the coding method
determine how many levels to use in an adaptive structure combining with interleaving in reducing the packet loss in
for a system. The different strategies were presented to IP networks. The performance of FEC data transferred
provide a system designer with the ability to optimize a from the source to destination was evaluated using FEC
specific system for the particular noise or fading conditions decoder which also voluntarily created the packet loss, and
experienced. recovered lost packets at the destination. The FEC coding
The adaptive software protocol model simulation performance was measured using an analytical method
program used could possibly be modified to simulate other stated in the paper. Single multiplexer network model was
ARQ protocols and also compare with other protocols in used for transmission of the data from multiple sources to
attempt to optimize the system further. destinations. In the paper, the unified method provided an
integrated framework for exploring the compromises
C Past Research works in Forward Error Correction
between the various key parameters i.e. channel coding
Lots of research works using Forward Error control have rates, interleaving depths, block lengths. It provided the
been performed. Among them are [8] which combined Path selection of various optimal coding strategies with various
Diversity with Forward Error Correction (PDF) systems for QOS requirements and system constraints.
Packet Switched Networks. In this research work, packet The Paper [11] used Forward Error Correction Schemes
loss rate was reduced by sending packets on appropriate to Protect Real Time Data Against Packet Loss and also
rates on disjoint paths from single sender to receivers. This recovered the lost packets of real time data. The factor was
work was extended to applications that were sensitive to the time in which data can be recovered as data that is
delay over packet switched networks for which single delayed for too long was regarded as lost. In the paper, the
sender to receiver connection were established using authors proposed and simulated a FEC enabled network
collection of relay nodes. protocol to complement current techniques like Low
Latency Queuing (LLQ) to reduce congestion error on
normal packet switched networks.

279
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
An erasure decoder was developed by combining two RS code as an MDS code, meets the conditions d = n –
erasure decoding techniques. The performances of various k + 1 or alternatively d – 1 = n – k, which means that it can
erasure decoders were compared. These techniques were correct any combination up to n – t erasures according to
also deployed in optical networks to reduce the effect of equation above when it is used as the erasure code for
packet loss arising from contention errors and Redundant reliable end to end communication in packet switched
Array of Inexpensive Disks systems. network. In this scenario, the source node transport layer
produces n transport packets using the RS encoder for
III. ERASURE CODING IN P ACKET S WICHED NETWORKS every k information packets made known from the
Erasure codes are FEC code classes which does not application layer. At the destination node transport layer,
require retransmission when they are employed. An erasure the RS decoder is able to recover the original information
is an error at which its location in the message is known to as long as it can receive any k out of the n packets
the decoder but has an unknown value. An erasure code is transmitted correctly. For this reason, the MDS erasure
designed to recover the erasures, as a substitute for codes are also called the k-out-of-n codes.
correcting errors from the packets correctly received or RS codes for different application requirements can
encoded bits. In a packet based network, an (n, k) erasure provide a large number of MDS codes. Extended RS codes
consists of k information packets and n-k parity packets are also maximum distance separable [13] and [14] by
over a finite field GF(q) (q is a power of a prime). adding one or two overall parity check(s). These attributes
It is assumed that packet-loss in an erasure channel is an have made RS codes an attractive candidate in this field.
independent event introduced by [12], with a fixed constant The decoding complexity of RS codes is achieved using
probability, . Packets are either presented as erasures or the fast Fourier transform technique at the scale of O (n
correctly received as erasures to the decoder in this channel log2 n), [15]. RS codes have been considered for different
model. The Erasure channel Shannon capacity is (1- and applications in the network environment [16], [17], and
[18], [19] and [20]. For example, employing the RS code in
transmission at any rate R< (1( with a random linear
association with the ARQ technique can largely improve
code can be achieved [12]. The number of erasures an
the performance of reliable multicast in the IP network
erasure code can correct, is bounded by d – 1 where d
[18], which is important for ensuring QoS in multimedia
is the minimum distance of the code. Clearly, this same
(video and audio) distribution across the network.
code doubles the number of errors that can be corrected
when it is used solely for error correction. The construction B Bose- Chaudhuri Hocquenghem Code (BCH) Code
of erasure codes can be classified into two categories using They discovered a code called BCH. It could be used to
a simple taxonomy: the maximum-distance separable detect and correct multiple errors code. BCH code is a
(MDS) code approach and sparse graph-code approach. generalized form of Hamming code. For any positive
The construction methods and performance features are integer possibility codes for m (m >3) and t (t < a
some of the summarised useful erasure codes as follows binary BCH code exists with the following parameters:
A Reed Solomon codes Block length n =
Reed Solomon codes are q-uary BCH codes special Number of Parity Check digits = n – k mt
subclass and form the most important class of the MDS Minimum distance d 2t + 1
codes. They have been widely used for error control in Generating polynomial g(x) is generally created as
digital communications and storage systems because of follows:
their powerful correcting capacity of burst error. RS codes Lowest Common Multiple (L.C.M) of {m1(x), m2(x),
can also be applied effective erasure codes to recover errors ………….m2t-1(x)}
of multiple erasures of size q. In general (n,k) RS code can The message m(x) is divided by g(x) and remainder will
be formulated using the generator polynomial. be represented as check bits r(x). Now whole encoded
g(x) = (x – αb)(x – αb+1)………….( for message is represented as:
some 2 and some a 1, where GF(q) for any i
E(x) = m(x) + r(x) (1)
and αn = 1 but αs 1 for any positive s n. The codes
produced have in length n = q – 1 symbols (or packet) with BCH Decoding
each packet containing q bits. The decoding of BCH is performed in three steps:
 Syndrome is computed from the received packet.

280
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
 Error location polynomial is generated from set of D. Simulating the proposed Coding with both RS codes and
equations and derived from the syndrome. BCH Codes
 Error location polynomial is used to identify and We analyze the packet loss rate of the coding schemes
correct the errant bits. with RS codes and BCH codes. For the simplicity of the
C Combined Coding Scheme analysis, RS (n, k) code and BCH codes (n, k) are used.
Both codes correct up to m erroneous messages. The
The data link layer adds authenticity to the physical messages we are talking about now is with respect to RS
layer by adding mechanism to detect and retransmit and BCH codes are in digits. The packet error ratio (PER)
damages or lost packets. It also used a mechanism to is the ratio of altered received packets of data to the total
recognize duplicate packets. Error control is normally number of the received packets. A packet is declared
achieved adding a trailer to the end of the packet. In this incorrect if at least one bit is erroneous. The expectation
coding scheme was designed in detail. value of the PER is called packet error probability, Pp for
End System which data packet length of N digits can be expressed as
= 1 – (1- )N (2)
Link
Where N is the number of packets in the network and in
Intermediate System queue, assuming that the bit errors are independent of each
other.
Using Dynamic Erasure with both RS and BCH codes,
Link Intermediate
up to, m errors can be corrected.
System
Number of packets in a source data block denote the
transmission of the block of source data, the number of
erasures as a throughput. The packet error probability
End System
Intermediate System produced in blocks of data is given by
Dynamic Erasure + RS Code (3)
Link
Dynamic Erasure + BCH Code (4)

IV. RESULTS AND D ISCUSSION

This section shows the table and graphs of Packet error
Figure 2: Hop to Hop Delivery For Error Control Probability produced from number of erasures generated
The combined coding scheme is as follows. First we from the proposed method compared with the combination
calculate the number of packets and find the parity symbol of the proposed method and two other methods.
bit transmitted. The erased channel or servers is k using RS Table I
and BCH codes. Parity digits is added at the end of the Packet Error Probability Of The Schemes At Bit Error Probability
Of 0.1.
packet. Remember that there will still be symbols, so after
RS and BCH coding, there will be RS packets and BCH Erased Dynamic with Dynamic with
packets with the proposed packets. Both RS and BCH Channels Dynamic Reed solomon BCH Code
packets will be decoded with the decoder and parity 2 1 1 0.999983
checking for the symbols will be obtained. If there is
4 0.96955 1 1
erasure in the parity checking, then the bits are discarded.
The remaining packets symbols are erasure free and will be 6 1 1 1
used as a decoding process. When all the source symbols 8 1 0.89288 1
are decoded successfully, then the receiver feeds back an 10 1 1 1
acknowledgement to the sender. The Sender continues to 12 0.996993 1 1
transmit the information until it receives the 14 0.87368 1 1
acknowledgement.
16 1 1 1
18 1 0.9978 1
20 0.999115 1 1

281
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
Table II
Graph showing Packet Error probability at Bit Error Packet Error Probability Of The Schemes At Bit Error Probability Of
probability of 0.1 0.3
1.02 Dynamic Dynamic
Erased with Reed with BCH
1.00 Channels Dynamic solomon Code
Packet Error Probability

0.98 2 1 1 1
0.96 Dynamic
4 1 1 1
0.94
6 1 1 1
Dynamic with
0.92
Reed solomon 8 0.413 1 1
0.90
Dynamic with 10 1 0.999999 1
0.88 BCH Code
12 0.999846 1 1
0.86
0 10 20 30 14 0.16897 1 1
Number of Erased Channels 16 1 0.988549 1
18 1 0.903109 1
Figure 3: Figure showing Packet Error probability of the Schemes at
20 1 1 1
Bit Error probability of 0.1

The graph shows the packet error probability for each of

the scheme at bit error probability of 0.1. The blue graph Graph showing Packet Error probability of the Schemes
at Bit Error probability of 0.3
shows the Dynamic Erasure Coding Method. The wine
colored graph shows the Dynamic Erasure coding with 1.20
Reed Solomon Method while the Lemon green coloured
graph shows the Dynamic Erasure coding with BCH 1.00
Packet Error Probability

Coding method. Dynamic erasure coding graph reduced

0.80
from 1 to 0.96955 and rose back to one, became uniform Dynamic
and then reduced to 0.87368, rose to 1, became uniform 0.60
and sloped down to 0.999115. Dynamic Erasure coding Dynamic with
with Reed Solomon was uniform at 1 the fell to 0.89288 0.40 Reed solomon
and rose back to 1, remained uniform, fell to 0.9978 and Dynamic with
finally rose to 1. Dynamic with BCH code rose from 0.20
BCH Code
0.999983 to 1 and remained uniform for the remaining
0.00
number of erasures. This graph shows that Dynamic
0 10 20 30
Erasure coding produced a lower packet error probability at
Number of Erased Channels
points 2, 6,8,10, 16 and 18. Dynamic Erasure with Reed
Solomon produced a lower packet error probability from 2,
4,6, 10, 12, 14, 16 and 20. Dynamic Erasure with BCH Figure 4: Graph showing Packet Error probability of the Schemes at
coding was uniform apart from point 4, 6, 8 10, 11, 12, 14, Bit Error probability of 0.3
16, 18, 20. Dynamic with BCH have a better retention of The graph shows the packet error probability for each of
packet error probability more than the Dynamic Erasure the scheme at bit error probability of 0.3. The blue graph
with Reed Solomon code and Dynamic Erasure coding. shows the Dynamic Erasure Coding Method. The wine
This is because of the deep production of steeper slope colored graph shows the Dynamic Erasure coding with
from both graphs (Dynamic Erasure Coding and Dynamic Reed Solomon Method while the Lemon green coloured
Erasure coding with Reed Solomon coding.) graph shows the Dynamic Erasure coding with BCH
Coding method.

282
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
Dynamic erasure coding graph reduced from 1 to 0.413 Graph showing Packet Error probability at Bit Error
and rose back to 1, fell and pass through two points and probability of 0.5
then, rose back to 1, became uniform till the end. Dynamic
Erasure coding with Reed Solomon was uniform at 1 and 1.01
sloped down to 0.999999 and rose back to 1, remained 1.00

Packet Error Probability

uniform, fell to 0.903109 and finally rose to 1. Dynamic
with BCH code remained uniform at 1 for althrough the 1.00
Dynamic
number of erasures. This graph shows that Dynamic 0.99
Erasure coding produced a lower packet error probability at
points 2, 4, 6, 8, 10, 16, 18 and 20. Dynamic Erasure with 0.99 Dynamic with
Reed solomon
Reed Solomon produced a lower packet error probability 0.98
from 2, 4, 6, 8, 12, 14, and 20. Dynamic Erasure with BCH Dynamic with
0.98 BCH Code
coding was uniform although the point. Dynamic with
BCH have a better retention of packet error probability 0.97
more than the Dynamic Erasure with Reed Solomon code 0 10 20 30
and Dynamic Erasure coding. This is because of the deep Number of Erased Channel
production of steeper slope from both graphs (Dynamic
Erasure Coding and Dynamic Erasure with Reed Solomon
Figure 5: Graph showing Packet Error probability of the Schemes at
coding). Bit Error probability of 0.5
Table III
Packet Error Probability Of The Schemes At Bit Error Probability Of
The graph shows the packet error probability for each of
0.5 the scheme at bit error probability of 0.3. The blue graph
shows the Dynamic Erasure Coding Method. The wine
Dynamic Dynamic
Erased with Reed with BCH colored graph shows the Dynamic Erasure coding with
Channels Dynamic solomon Code Reed Solomon Method while the Lemon green coloured
graph shows the Dynamic Erasure coding with BCH
2 0.972044 0.984449 0.983857
Coding method. Dynamic erasure coding graph rose from
4 1 1 1 0.983857 and remained uniform from points 4 to 10 and
6 1 0.999988 1 fell to 0.993305 and rose back to 1, became uniform till the
last point. Dynamic Erasure coding with Reed Solomon
8 1 1 1
graph rose from 0.984449 to 1, and sloped down to
10 1 1 1 0.996276, and rose back to 1, remained uniform, passing
12 0.993305 0.996276 0.995857 through 0.999999 and rose back to 1, remained uniform
until the last point. Dynamic with BCH code graph rose
14 0.999999 0.999999 0.999999
from 0.983857 to 1 and remained uniform from points 4 to
16 1 1 1 10, fell to 0.995857, rose back to 1 passing through point
18 1 1 1 0.999999 and remained uniform till the end of the points
for number of erasures. This graph shows that Dynamic
20 1 1 1
Erasure coding produced a lower packet error probability at
points 4, 6, 8, 10, 16, 18 and 20. Dynamic Erasure with
Reed Solomon produced a lower packet error probability
from 4, 8, 10, 16, 18, and 20.

283
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
Dynamic Erasure with BCH coding graph produced a The graph shows the packet error probability for each of
lower packet error probability at points 4, 6, 8, 10, 16, 18, the scheme at bit error probability of 0.7. The blue graph
20. Both Dynamic Erasure Coding and Dynamic Coding shows the Dynamic Erasure Coding Method. The wine
with BCH have a better retention of packet error colored graph shows the Dynamic Erasure coding with
probability more than the Dynamic Erasure with Reed Reed Solomon Method while the Lemon green coloured
Solomon coding. This is because of the deep production of graph shows the Dynamic Erasure coding with BCH
steeper slope from Dynamic Erasure with Reed Solomon Coding method. Dynamic erasure coding graph was
coding. initially uniform from points 2 to 6, sloped down to
Table IV
0.999999, rose to 1 and fell to 0.999827, rose to 1,
Packet Error Probability Of The Schemes At Bit Error Probability remained uniform at points 14 and 16, sloped down to
Of 0.7 0.999755, and finally rose to 1 and remained uniform until
Dynamic Dynamic the last point of number of erasures. Dynamic Erasure
Erased with Reed with BCH coding with Reed Solomon graph was initially uniform at
Channels Dynamic solomon Code 1, from point 4 and sloped down to 0.999997, rose back to
2 1 1 1 1, remained uniform until the last point. Dynamic with
BCH code graph was initially uniform at 1 from points 2 to
4 1 1 1
4, sloped down to 0.999995, rose to 1 passing through
6 1 1 0.999995 0.999959 and remained uniform until the end of points.
8 0.999999 0.999997 0.999959 This graph shows that Dynamic Erasure coding produced a
lower packet error probability at points 2, 4, 6, 10, 14, 16
10 1 1 1 and 20. Dynamic Erasure with Reed Solomon produced a
12 0.999827 1 1 lower packet error probability at points 2, 4, 6, 10, 12, 14,
14 1 1 1
16, 18 and 20. Dynamic Erasure with BCH coding
produced a lower packet error probability at points 2, 4, 6,
16 1 1 1 10, 12, 14, 16, 18, 20. Dynamic with Reed Solomon have a
18 0.999755 1 1 better retention of packet error probability more than the
20 1 1 1
Dynamic Erasure with BCH code and Dynamic Erasure
coding. This is because of the deep production of steeper
slope from both graphs (Dynamic Erasure Coding and
Graph showing Packet Error probability at Bit Error Dynamic Erasure with BCH).
probability of 0.7
Table V
1.00 Packet Error Probability Of The Schemes At Bit Error Probability Of
0.9
1.00 Dynamic Dynamic
Packet Error Probability

Erased with Reed with BCH

1.00 Channels Dynamic solomon Code
Dynamic
1.00 2 1 1 1
4 1 1 1
1.00 Dynamic with
Reed solomon 6 1 1 0.9
1.00
Dynamic with 8 1 1 1
1.00 BCH Code
10 1 1 1

1.00 12 1 1 0.9999
0 10 20 30 14 1 1 1
Number of Erased Channels
16 1 1 1
18 1 1 0.99
Figure 6: Graph showing Packet Error probability of the Schemes at
Bit Error probability of 0.7. 20 1 1 1

284
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
V. CONCLUSION
Graph showing Packet Error probability at Bit Error
probability of 0.9 In this paper, A method was proposed using Dynamic
Erasure Coding Method based on the number of erased
1.02
channels and good channels. Via simulation, It was shown
1.00 that the method reduced Packet Error Probability. With
increase in the number of erasures, the network error
Packet Error Probability

0.98 reduced to a very minimal level and combining Dynamic

Dynamic Erasure with Reed Solomon and Bose Chaudhurin
0.96
Hoqchequeim Coding the Packet Error Probability reduced
0.94 Dynamic with with also increase the number of messages transmitted and
Reed solomon acknowledged.
0.92
Dynamic with REFERENCES
0.90 BCH Code
[1] Bosco, H. L. and Dowden, D.C., (2000), Evolution of Wide Area
Networks, Bellab Technical Journal, Volume 5, Pages 46–72.
0.88 [2] Pitsillides, A. and F.N. Pavlidou, 2001. A survey of wireless ATM
0 10 20 30 handover issues, C. Chrysostomou. Proceedings of the International
Number of Erasured Channels Symposium of 3G Infrastructure and Services, July 2-3, 2001,
Athens, Greece, pp: 34-39.
[3] Aliyu, F. M., Y. Osais, I. Keshta, A. Binajjaj, (2015). Maximizing
Figure 7: Graph showing Packet Error Probability of the schemes at Throughput of SW ARQ with Network Coding through Forward
bit Error Probability of 0.9 Error Correction, International Journal of Advanced Computer
Science and Application, Vol. 6, No. 6.
The graph shows the packet error probability for each of [4] Kleinschmidt, J. H. and W. C. Borelli, (2009), Adaptive Error
the scheme at bit error probability of 0.9. The blue graph Control Using ARQ and BCH Codes in Sensor Networks Using
shows the Dynamic Erasure Coding Method. The wine Coverage Area Information, Conference Profeedings of the IEEE
20th International Symposium on Personal Indoor and Mobile Radio
colored graph shows the Dynamic Erasure coding with
Communications, (PIMRC 2009), 13–16 September 2009, Tokyo,
Reed Solomon Method while the Lemon green coloured Japan.
graph shows the Dynamic Erasure coding with BCH [5] Khosravirad, S.R. and H. Viswanathan (2017), Analysis of Feedback
Coding method. Dynamic erasure coding graph, remained Error in Automatic Repeat Request, Nokia Bells Labs.
uniform at 1 althrough until the last point of number of [6] Mofrad, A. B., (2008), ARQ Protocols Supporting QoS in
erasures. Dynamic Erasure coding with Reed Solomon Embedded Systems, Masters Thesis in Computer Systems
graph remained uniform although at 1 until the last point. Engineering, Technical Report, IDE0609, School of Information
Science, Computer and Electrical Engineering, Halmstad University,
Dynamic with BCH code graph was initially uniform at 1 Turkey.
from points 2 to 4, sloped down to 0.9, rose to 1, a [7] Guth K.J. (1989), An Adaptive ARQ Strategy for Packet Switching
remained uniformed at points 8 and 10, sloped down to Data Communication Networks, Naval Postgraduate School,
0.9999, rose to 1, remained uniform for points 14 and 16, Monterey.
sloped down to 0.99 and finally rose to 1. This graph shows [8] Nguyen T., and A. Zakhor, (2003), Path Diversity with Forward
that Dynamic Erasure coding produced a lower packet error Error Correction (PDF) System for Packet Switched Networks, In
Proceedings of INFOCOMM 2003, San Francisco, California,
probability at points 2, 4, 6, 8, 10, 12, 14, 16, 18 and 20. U.S.A.
Dynamic Erasure with Reed Solomon produced a lower [9] Biczok, G. Y. Chen, K. Kralevska and H. Overby, (2016),
packet error probability at points 2, 4, 6, 8, 10, 12, 14, 16, Combining Forward Error Correction and Network Coding in
18 and 20. Dynamic Erasure with BCH coding produced a Bufferless Networks: A Case Study for Optical Packet Switching,
lower packet error probability at points 2, 4, 8, 10, 14, 16, IEEE 17th International Conference Switching and Routing,
Yokohama, Japan, Pages 61-68.
20. Both Dynamic Erasure coding and Dynamic Erasure
[10] Prasad, K. S. R., M. V. S. S. Nagendranadh, and M. S. Srinivas,
Coding with Reed Solomon have a better retention of (2013), Improving the Efficiency of Forward Error Correction
packet error probability more than the Dynamic Erasure Coding in Reducing the Network Packet Loss; International Journal
with BCH code a. This is because of the deep production of of Advanced Research in Computer and Communication
steeper slope from both graphs (Dynamic Erasure with Engineering, Vol 2, Issue 8, Pages 3306–3314.
BCH). [11] Versfeld, D. J. J., and H. C. Ferreira and A. S. J. Helberg, (2015),
Using Forward Error Correction to Protect Real time data against
Packet loss, Proceedings of Southern African Telecommunicaion
Networks and Applications Conference, Spier Wine Estate,
Stellenbosch, South Africa, 6-8th September, Pages 1 – 8.

285
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 9, Issue 10, October 2019)
[12] Elias, P. (1955), Coding for Noisy Channels, Proceedings of 3rd [27] Berkekamp, E. R., R. E. Peile, and S. P. Pope, (1987), The
London Symposium, Information Theory Pages 61-66. Applications of Error Control to Communications, IEEE
[13] Tanaka, H., and F., Nishida, (1970), A Construction of a Polynomial Communications Magazine, Vol. 25, No. 4, Pages 44 – 57.
Code, Electron Communication, Japan, 53-A, Pages 24-31. [28] Blaum, M., Brandy, J., Bruck, J., and Menon, J., (1996), Evenodd:
[14] Gross, A.J., (1973), Some Augumentations of Bose_Chaudhuri error An Efficient Scheme of tolerating double disk failures in RAID
correcting codes in Srivastava, J. N.(Edition); A survey of Architecures, IEEE Transaction Computation, No. 44, Pages 192-
Combinatorial Theory, North-Holland, Amsterdam. 202.
[15] Justesen, J., (1976), On the Complexity of Decoding Reed Solomon [29] Bose, R.C. and D. K. Ray-Chaudhuri (1960), On a Class of Error
Codes, IEEE Transformational Information Theory, No. 22, Pages Correcting Binary Group Codes at Information and Control 3, Pages
237-238. 68-79.
[16] McAuley, A. J., (1990) Reliable Broadband Communication Using a [30] Chen, H., Fu Song, (2016), Improving Coding Performance and
Burst Erasure Correcting Code, Proceedings ACM SIGCOMM ’90 Energy Efficiency of Erasure Coding Process for Storage Systems –
Philadelphia, P.A., September, Pages 287-306. A Parallel and Scalable Approach, Institute of Electronics and
Electrical Engineers, 9th International Conference on Cloud
[17] Rizzo, L., (1997). Effective Erasure codes for Reliable
Computing, Pages 933-936.
Communication Protocols, ACM Communications Review Vol.27,
No. 2, Pages 24-36. [31] Dai, B., W. Zhao, Jan Yang and Lu Lv, (2014) CODEC: Content
Distribution with (N,K) Erasure code in MaNET, International
[18] Nonnenmacher, J., E.W. Biersasck and D. Towsley, (1998), Parity
Journal of Computer Networks and Communications (IJCNC), Vol.
Based Loss Recovery for Reliable Multicast Transmission, IEEE
6, No. 4, Pages 39–51.
Transactional Communication, 1998, Vol. 6. Pages 349– 361.
[32] Dimakis, A.G. and K. Ramchadran (2008), Network Coding for
[19] Sakakibara, K., and M. Kasahara, (1995), A Multicast Hybrid using
Distributed Storage in Wireless Networks, Networked Sensing
MDS Codes and GMD decoding, IEEE Transaction in
Information and Control, Springer Science + Business Media, LLC,
Communication 43, pp 2933 – 2939.
2008, Pages 115–134.
[20] Rubenstein, D., J. Kurose, and D. Towsley, (1998), Real-Time
[33] Donglas, C., S. B. Toby and R. Bridgehall, (2004), Energy
Reliable Multicast, Using Proactive Forward Error Correction,
Efficiency of CSMA protocols for Wireless Packet Switched
Proceedings NOSDAV, Cambridge, U.K. July, Pages 279-294.
Networks, Proceedings of IEEE Wireless Communication and
[21] Ababneh, J. and O. Almomani, (2014), Survey of Error Correction Networking Conference, Volume 1, Pages 447-452.
Mechanisms for Video Streaming over the Internet, International
[34] Dressler, F., M. B. Li, R. Kapitza, S. Ripperger, C. Eibel, B. Herzog,
Journal of Advanced Research in Computer Science Applications,
T. Honig and W. Schroder-Prekischat. (2016). Monitoring Bats in
West Yorkshire, U.K., Vol. 5, No. 3., Pages 155–161.
the Wild: On Using Erasure Codes for Efficient Wireless Sensor
[22] Abdullah, A.S., M.J. Abbasi and N. Fisal, (2015), Review of Networks, ACM Transaction on Sensor Networks, Vol 12, No 1,
Rateless-Network-Coding Based Packet Protection in Wireless Article 7.
Sensor Networks, Hindawi Publishing Corporation, Mobile
[35] Elfouly, T., M. Saleh and O. M. Malluhi (2008), Efficient Forward
Information Systems, Volume 2015, Article ID 641027, Pages 1–15,
Error Correction for Reliable Transmission in Packet Network,
http://dx.doi.org/10.1155/2015/641027.
Proceedings of 2008 International Conference on Parallel and
[23] Ajutsu, H., K. Ueda, H. Saito, (2017), MEC: Network Optimized Distributed Techniques and Applications, Las Vegas, U.S.A,
Multi-Stage Erasure coding for Scalable Storage Systems, IEEE 22nd Volume 1, Pages 103-109.
Pacific Rim International Symposium on Dependable Computing,
[36] Tovirac, J. (2005), Error Control for Message Multicast over
Christchurch, Canterbury, New Zealand, Pages 292–300.
Wireless Links, PhD Dissertation, Institute for Telecommunications
[24] Arrobo, G. and R. Gitlin, (2014), Minimising Energy Consumption Research, University of Southern Austrailia, Austrailia.
for Cooperative Network and Diversity Coded Sensor Networks,
[37] Towsley D., J.K. Wolf, (1979). On the Statistical Analysis of Queue
2014 Wireless Telecommunications Symposium, Pages 103 – 109.
Length and Waiting times for Statistical Multiplexers with ARQ
[25] Bada, A.B. (2017) Automatic Repeat Request (ARQ) Protocols, The Retransmission schemes, IEEE Transaction on Communications,
International Journal of Engineering and Science (IJES), Vol 6, Issue VOL COM-27, No. 4, March 1979.
5, Pages 64-66.
[38] Tucker, R.S. (2011), Scalability and Energy Consumption of
[26] Balinga, J. (2011), Energy Consumption in Wired and Wireless Optical and Electronic Packet Switching, IEEE Journal of Light
Access Networks, IEEE Communications Magazine, Volume 49, Wave Technology, Vol. 24, No 16, Pages 2410-2421, 445 Hoes
Issue 6, pages 70 – 77, South West, Nimbus Avenue, Portland, Lane, Piscataway, New Jersey 08854-4141, U.S.A.
Oregon 97233, U.S.A

286