efficient shortest path routing (espr) algorithm for multicasting in
TRANSCRIPT
Efficient Shortest Path Routing (ESPR) Algorithm for Multicasting in
Wireless Mesh Network
1Geetanjali Rathee, 2Ninni Singh, 3Hemraj Saini 1, 2, 3 Department of Computer Science and Engineering
1, 2, 3 Jaypee University of Information Technology, Waknaghat-173234 INDIA
[email protected],[email protected], [email protected]
Abstract
Today’s fast and mobility oriented era of
communication the networking infrastructure
requires efficient communication. SPT and MCT
are the two standard algorithms to provide efficient
communication in wireless networking. SPT
provides communication by recognizing the
shortest path in a tree and propel the packet
throughout this path. Since failures exist in the
network, the SPT algorithm resists the failure by
sending an ICMP messages which increases
congestion ratio and rerouting of the packets while
MCT algorithm forwards the packets on the basis
of least number of transmissions. As the MCT
algorithm transmits by consuming less number of
cost, but during failure of intermediate links, it is
also having the scope of improvement in the
efficiency, i.e. transmission cost, rerouting cost and
complexity of the network. These drawbacks are
overcome by the proposed algorithm which works
efficiently during failure of intermediate links. NS-2
is used to simulate the results and shown the results
with comparative analysis of three algorithms
Keywords: WMN, SPT, MCT, Congestion,
ESPR.
1. Introduction In wired computer network, as the number of users
increase in the network, there will be a demand of
some novel technologies that can oversee the issues
like mobility and as a result wireless computer
network is developed. In wireless network, wired
connection is not needed and because of this
feature, there is an exponential growth of users for
Internet access. As we all know communication in
wireless network runs for substantial purpose,
Wireless Mesh Network (WMN) become a
promising concept to solve the challenges in
today’s scenario in cost effective aspects to the
service provided by adopting three features i.e.
self-structuring, flexibility and self-management [1,
2].
WMN [3], as shown in figure-1, is made up of two
types of nodes, first, wireless mess client and
second, wireless mesh router. The wireless mesh
router has routing capability and not having any
resources like- memory and power constraints
while wireless mesh clients has resources like-
memory and power constraints but not having any
routing capability. Mesh router has less mobility so
it is mostly viewed as a stationary node. The mesh
router which connects the other mesh routers and
clients to the backbone or Internet is called a mesh
gateway. Mesh gateway is also practiced to connect
any two mesh networks [3].
Fig. 1. Wireless Mesh Network
Multicast Routing [5, 9] is a type of
communication where information is delivered to a
set of destination nodes simultaneously as shown in
figure-2. In this, the delivery of messages is
managed at each layer only once. The duplication
of messages is provided only at branch layer where
links are differentiated to the destination node.
Even though Multicast Routing is applied in
different applications like- teleconferencing, video
conferencing and email sending, researchers on
Multicast Routing in WMN is still in its early days.
Shortest Path tree (SPT) [6] and Minimum Cost
Trees (MCT) are the two underlying approaches of
multicast routing.
The aim of SPT algorithm is to create a tree such
that the distance between source and destination
should be minimal. The most commonly used SPT
algorithms are Dijkstra [7] and Bellman- ford [8].
Fig. 2. Multicast Routing
Hemraj Saini et al, Int.J.Computer Technology & Applications,Vol 6 (1),111-115
IJCTA | Jan-Feb 2015 Available [email protected]
111
ISSN:2229-6093
The goal of MCT is to downplay the overall cost of
the tree. On comparing both trees SPT and MCT,
the end to end cost of SPT is lower than that of
MCT. The distance between source and
destinations is maximum in case of MCT that
means the average path length of SPT is lower than
MCT. The complexity of MCT is more than the
SPT. Due to its higher complexity rate, the majority
of routing protocols is based on SPT i.e. distance
vector multicast routing (DVMR) [11], Multicast
Open Shortest Path First (MOSPF) [12]. Rest of the
work is presented in four sections. In section-II,
recent proposed approaches, i.e. SPT [11] and
MCT [6] have discussed with their drawbacks. In
section-III, the proposed approach has provided
with failure cases. The comparative analysis of
three approaches will be done in section-IV.
Finally, the section-V concluded the work.
2. Related Work In this section, the computation cost of SPT and
MCT have been compared. This comparative study
evaluates the drawback of both algorithms during
communicating the data in the network. For this
study, the WMN [4], as shown in figure-3 have
deemed.
.
Fig. 3. Shortest Path Tree
The network consists of a source node S, two
destination node D1, D2 and four forwarding nodes.
Firstly the SPT [11] has calculated and then the
MCT [6].
SPT algorithm, forwards the packet by
constructing a shortest path tree to the destination
node. Let us consider a source node S which wants
to transmit a packet to a destination node D1 and
D2. As shown in figure-3, source node S will
transmit the packet to D1 and D2 by choosing S-C-
E-B-D1 and S-C-E-B-D2 paths, therefore, the cost
of packet transmission from S to D1 is 4 (1+1+1+1)
and from S to D2 is 4 (1+1+1+1) and the overall
cost for transmitting the packet from source S to D1
and D2 is 5(S-C-E-B -D1and B-D2).
Now, another multicast routing algorithm MCT
has been considered for the computation. Instead of
considering the cost, it sends the packet on the
basis of least number of transmissions. The
working of MCT is shown by considering figure-4.
Let again source node S wants to transmit a packet
to multiple destinations D1 and D2. The MCT
evaluates the transmission of the entire network
and sends the packet by choosing the path having
less number of transmissions. By showing in
Figure-4, MCT choose S-A-B-D1 and S-C-E-D2,
so, the cost of transmission from S-D1 is 3(2+2+1)
and from S-D2 is 3(1+1+3). The overall cost of
packet transmission from S-D1 and S-D2 is 10.
Fig. 4. Minimum Cost Tree
The drawback of MCT is its complexity.
Furthermore, in the case of failure [10], both the
algorithms utilize the ICMP message protocol,
which increases the computing power, cost,
rerouting transmission and congestion ratio. Thus,
to overcome all such challenges the proposed
algorithm is provided in section-III, which work in
an efficient manner.
3. Proposed Approach In the previous section, some of the disadvantages
associated with both SPT and MCT algorithms are
identified. To overcome above mentioned
drawbacks we have suggested an algorithm
efficient shortest path routing (ESPR). As shown in
table-1.
Table-1 shows if a node Ni wants to send the
packets to destination node Di and Dj, initially node
Ni will check for the failure case using through an
ack (acknowledgement).
If ack is received within RTT then there is no
failure in the network and node Ni will send the
packet to destination Di and Dj using SPT
algorithm. In other case if ack is not received
within RTT that means failure exist in the network
and the previous node Ni-1 will choose the next
shortest path from the Routing table and send to the
destination nodes Di and Dj.
Fig. 5. When node C-D fails
Hemraj Saini et al, Int.J.Computer Technology & Applications,Vol 6 (1),111-115
IJCTA | Jan-Feb 2015 Available [email protected]
112
ISSN:2229-6093
Table 1. ESPR Algorithm
Terminologies used:
S= source node
Di and Dj= two destination nodes
Ni= current node
Ni-1= previous node
N= Network
RTT= Round Trip Time
1. Select a node Ni from a network N.
2. if ( an ack receives within RTT)
Then
No failure exists
Node Ni sends the packet using SPT
algorithm
3. else
node Ni-1 chooses next another path
Ni-1 send the packet to destination Di, Dj
The proposal utilizes the concept of SPT i.e. source
node S transmits the packet by choosing the
shortest path to the destination node, but in case of
failure instead of sending an ICMP message to the
source Node S, the failure detected node will
choose another shortest path (by considering its
routing table) that leads to destination node D1 and
D2 and starts transmitting the packet through that
path. The failure case is shown in the figure-5.
In normal cases (when no failure exist inside the
network), ESPR works same as SPT i.e. Source
node S sends the packets to the destination node D1
and D2 through S-C-E-B-D1 and S-C-E-B-D2, while
during failure, i.e. if link C-E fails, as node C is not
able to get an acknowledgement within its (Round
Trip Time) RTT, it will immediately choose
another path to send the packets to the destination
node D1 and D2 through S-C-B-D1 and S-C-B-D2
path. The corresponding flowchart of our algorithm
is shown in figure-6.
Fig. 6. Flow chart of ESPR
4. Performance Evaluation So far we have discussed our proposed algorithm
with both previous approaches and shown how the
proposed algorithm works in the failure cases. In
this section we have simulate our work in NS-2
simulator on different network sizes and shown the
corresponding graphs of SPT, MCT and ESPR in
terms of throughput, resiliency.
5 10 15 20 25 305
10
15
20
No. of Nodes
Thro
ughput/
ResilencyS
PT
Throughput
Resilency
Fig.7. Throughput and resiliency of SPT
Furthermore we have construct a table of all three
algorithms using two parameters i.e. throughput (in
term of cost) and resiliency (against node failure).
The above graphs and table clearly shows that PA
works efficiently in failure cases as compare to
SPT and MCT. Finally we conclude our paper in
table 3 by showing the features, pros and cons of
SPT, MCT and PA.
5 10 15 20 25 307
8
9
10
11
12
13
14
15
16
17
No. of Nodes
Thro
ughput/
ResilencyM
CT
Throughput
Resilency
Fig. 8. Throughput and resiliency of MCT
Hemraj Saini et al, Int.J.Computer Technology & Applications,Vol 6 (1),111-115
IJCTA | Jan-Feb 2015 Available [email protected]
113
ISSN:2229-6093
5 10 15 20 25 304
4.5
5
5.5
6
6.5
7
7.5
8
8.5
9
No. of Nodes
Thro
ughput/
ResilencyE
SP
R
Throughput
Resilency
Fig. 9. Throughput and resiliency of ESPR
5 10 15 20 25 304
6
8
10
12
14
16
No. of Nodes
Thro
ughput
(in t
erm
s o
f cost)
SPT
MCT
ESPR
Fig. 10. Comparative throughputs of SPT,
MCT and PA.
5 10 15 20 25 305
10
15
20
No. of Nodes
Resilency (
in t
erm
s o
f link f
ailure
)
SPT
MCT
ESPR
Fig. 11. Comparative resiliencies of SPT,
MCT and ESPR
Table 3. Pros and cons of SPT, MCT and
ESPR
SPT MCT PA
Task: SPT finds
the
shortest
path in a
network.
MCT is
generally
used to find
path having
least
transmission
nodes.
It works
same as
SPT, but it
works
efficiently
in the case
of a node or
link failure.
Pros: Minimize
end to end
delay.
The Overall
cost of the
multicast tree
reduces.
1)
Minimizes
no. of
computation
2) Minimize
congestion
ratio.
3) Minimize
rerouting.
Cons: 1) Use
ICMP
message
in the case
of failure.
2)
Increases
congestion
1) Increases
the
computational
complexity.
2) Increases
congestion
because of
rerouting.
---
Table 2. Throughput and resiliency of SPT,
MCT and ESPR
Parameter
Metric
Approaches 7-
node
15-
node
30-
node
Throughput
(In terms of
cost)
Shortest Path
Tree
5 7 10
Minimum
Cost Tree
6 10 12
PA 4 5 7
Resiliency
(Against
node
failure)
Shortest Path
Tree
8 12 19
Minimum
Cost Tree
7 9 14
PA 5 7 9
5. Conclusion and Future Work
In this work an ESPR multicast algorithm has
proposed which ensures efficient transmission in
WMN during failure. The task was simulated in
NS2 at different network sizes and found that cost
of transmission is less as compared to other
algorithms i.e. SPT and MCT. The proposal has
introduced several advantages over SPT and MCT
i.e. minimizes numbers of computations,
congestion ratio and re-routing of packets. Further
to achieve more precise results of ESPR algorithm,
it will be analyzed in the physical environment and
the results will be revised for other doable
measures.
Hemraj Saini et al, Int.J.Computer Technology & Applications,Vol 6 (1),111-115
IJCTA | Jan-Feb 2015 Available [email protected]
114
ISSN:2229-6093
References [1.] I. F Akyildiz, X .Wang and W. Wang, “A Survey on
Wireless Mesh Networks”, IEEE Radio
Communications, 47(4):445-487, 2005. [2.] H. Saini, “1-2 Skip List Approach for Efficient
Security Checks in Wireless Mesh Networks”,
International Journal of Electronics and
Information Engineering, 1(1):9-15, 2014. [3.] L .Zhou, and Z. Haas, “Securing adhoc
networks”, IEEE Network Magazine, 13:24-
30:1999.
[4.] Raffaele Bruno, M. Conti, and E. Gregori, “Mesh networks: commodity multihop ad hoc
networks”, Communications Magazine,
IEEE, 43(3):123-131, 2005.
[5.] S.Paul, “Multicast on the Internet and Its Applications”, Kluwer Academic Publishers, 1998.
[6.] G. Zeng, B.Wang, Y. Ding, L. Xiao, and M. Mutka,
“Efficient Multicast Algorithms for Multichannel
Wireless Mesh Networks”, IEEE Trans. Parallel Distrib. Syst., 21(1):86-99, 2010.
[7.] L. Ping, and B. Cuimei, “A Improvement of the
Shortest Path Based on the Dijkstra Algorithm [J]”,
Qinghai Normal University, 1(1):79-80, 2008. [8.] W. Zhang, H. Chen, C. Jiang, and L. Zhu,
“Improvement and Experimental Evaluation
Bellman-Ford Algorithm”, International
Conference on Advanced Information and Communication Technology for Education
(ICAICTE 2013), 150-153, 2013.
[9.] J. Sobczak, and P. Zwierzykowski, “Multicast Routing in Wireless Mesh Networks”, In the
proceedings of The Eighth Advanced International
Conference on Telecommunications, 62-68, 2012.
[10.] G. Rathee, A. Mundra, N. Rakesh, and S.P. Ghrera, “Buffered Based Routing and Resiliency Approach
for WMN”, Published in IEEE International
Conference of Human Computer Interaction,
Chennai, India, 1-7, 2013. [11.] B.A Fourozan, “TCP/IP Protocol Suite”, 3rd
Edition, Singapore, McGrawHill, 2004.
[12.] P.G. Ranjitkar, I. M. Suliman, P. Geil, Kuipers and
R. Prasad, “IP multicast implementation based on the multicast extensions to OSPF protocol”
Personal Wireless Communications, IEEE
International Conference on, 484 – 489, 2000.
Hemraj Saini et al, Int.J.Computer Technology & Applications,Vol 6 (1),111-115
IJCTA | Jan-Feb 2015 Available [email protected]
115
ISSN:2229-6093