International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248

Volume: 3 Issue: 10 141 – 146

_______________________________________________________________________________________________

Trust Based Node Recovery and Checkpointing Techniques in Manets

Rupali Mittal Dr. Jaswinder Singh

M.Tech Research Scholar Assistant Professor
Department of Computer Engineering Department of Computer Engineering
Punjabi University, Patiala, India Punjabi University, Patiala, India
[email protected] [email protected]

Abstract: Checkpointing is a process of determining the vulnerability of node in case of any attack occurs in the network. It depends on the
cluster change count value of the node. If the measure of the hop exchanges required to reach the destination node from the current node, is
above the previously specified value, the node under consideration is unsafe and safe points must be implemented in between the path and
different subnetworks within that network must have their own implemented safe points. The message must commits to the safe points as it
reaches the respective sub networks.
The message in the networks evolve over the certain subnetworks. The each subnetwork has the checkpoint node, that serves the purpose for
communication between different subnetworks, or between the hops in different subnetworks. This phenomenon supports the system
efficiency and preserves the robustness. The process retrieval methods, therefore, should be implemented with the use of the safe points to
prevent system degradation. In this research paper, an efficient recovery protocol is designed for distributed transactions in MANETs so that
failures can be minimised. Dynamic analysis has also been done and it is compared with other existing protocols to validate the attained result.

Keywords: Manet
__________________________________________________*****_________________________________________________

I. INTRODUCTION and „redo‟ operations are applied to the processes. In case of

A MANET is self- configuring network connected without deferred updates, if the process has committed to some
links. The nodes that have to participate in MANETS do not checkpoint, then the redo operation is implemented on the
need any extraneous knowledge about the network aborted process. If the process aborts before committing to
parameters before joining the network. In mobile ad-hoc any safe point then the process is undone and all the updates
network, every mobile node is free to move in any direction made are withdrawn and process is moved to its initial state.
and will therefore transform its links to other devices The process retrieval operations or techniques i.e. the log-
repeatedly. The main objective in creating a MANET based recovery protocol or the two-phase commit protocol
network is to retain the information required to properly necessitate some extra details to be added for the operation
route the traffic. They have one or numerous and dissimilar to be done. The rollback recovery technique also imposes
transceivers between nodes. This results in a highly certain disadvantages. This technique implement various
dynamic, autonomous topology. Distributed systems safe points in between the initial state and the final state. Is
nowadays are used in each and every field and provide some process in execution has not committed to any of the
many applications like Client-Server systems, transaction safe points, and gets terminated, then all of its updates will
processing, World Wide Web and many more. When these be undone and process needs to be executed from the crash.
systems are exposed to attacks and their computation of Other issue is that, the system has to maintain the large
recovery become a huge problem. Hence, various database to keep track of the updates that have been done for
methodologies have been presented to enhance their the various processes. When the process gets aborted after
reliability and reduce the risk of failures that involve committing to specific safe point, then redo operation is
rollback recovery, transaction and group communication. applied and system needs to keep track of the updates which
The retrieval by pushing back all the updates, view the needs to be recovered after the process is re-executed. When
system as the batch of tasks transmitted over the local area the particular process gets aborted, then effect propagates to
networks. The various actions involved refers to a large some other processes in the system. The process having
static database which provides more robustness to the their execution connected to that specific process, also
system and prevents the system degradation. If the update terminate their execution or needs to be re-executed.The
received is the deferred or partial update, that is, the update process recovery techniques based on the use of various safe
has not been made completely, then the process will return points between the initial and final points requires empty
or retrieved back to its initial stage, referring to the static space or the database to store the updates related
database, which has the initial information stored for all the information. As more and more transaction updates are
processes involved by the system. Hence, protects the being made by the process in execution, the overhead of the
process to go out of state. For the purpose, the database push system sometimes increases in order to maintain the record
all the partial updates and recover the process to the of all the updates. Also when the process aborts in between
previous saved point defined in the database. The „undo‟ the two checkpoints, then the system needs to decide, which
141
IJFRCSCE | October 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 10 141 – 146
_______________________________________________________________________________________________
updates needs to be preserved and which need not to be. The on rollback recovery have been discussed. Additionally, a
updates which are not made completely or are partial, multi-check pointing protocol has been proposed which
required to be removed when the process re-executes. reduces overall overhead incurred while check pointing.
Hence, to keep track of all the things, the automatic memory
management techniques need to be implemented to serve the Tuli Kumar and P.K Jaggi[2013] discussed the technique
purpose. The rollback based recovery techniques is being for the wireless communication networks, where the
used in many real-time based applications. The advantages communication is generally carried between the two or more
of this technique includes the applicability of the robustness hosts on different networks. The safe points or the check
for the system. In case of certain process gets aborted in points are defined for the processes within those systems,
between two checkpoints or the process has committed which are geographically distributed over the large
some safe point, then all the updates that have been done clustering regions. The safe points are defined between the
after the previously committed safepoint/checkpoint need to different hosts/nodes for communicating the message. If the
redone. The process need not to be executed from the crash. process/message has been committed to certain safe point,
The rollback process recovery protocol is basically of two then the information is updated at the dedicated database for
types. One with the uncoordinated safepoints and other with the purpose located at the base station. If the certain host
the coordinated safepoints. In case of technique with network not able to communicate the message to the
coordinated safepoints, the various safepoints between the intended receiver network, followed by the failure to
initial and the terminated state are defined in the start before commit to the safe point, then the message is retrieved back
the execution of the process. The process after undergoing to the pre-final or the initial state. In this case, the redo
the specified update criteria, commits to the next safe point operation is carried out and the message is re-transmitted
from the present state and updates made till that point are using different host network. The process is called the hand-
preserved for the future. In the other technique with the over and the checkpoints used here are process non-
uncoordinated safepoints, the process can create the safe blocking checkpoints.
point at the run time when required. The disadvantage of the
uncoordinated safepoints is the creation of the inappropriate Doug Hakkarinen and Zizhong Chen [2013] defined the
safepoints with the impotent updates that leads to the system technique in which the numerous checkpoints are defined.
overhead due to utilization of excessive resources. The The process based on the updates made by its execution,
technique with the coordinated safepoints has the minus commits to or assigned to the particular checkpoint. If some
point that if the process has not committed to any safe point or the checkpoint is never been assigned to any
checkpoint before it gets aborted, the process needs to be re- process in execution then it is of no use. Hence, need to be
executed from the starting of time, this phenomenon is removed from the system database to free some space to be
called as domino effect. The process in this case has to utilized for the other purposes. To serve the purpose, the
secure the minimum required updates to achieve the stable process needs to keep track/record of various checkpoints.
state. Hence, the rollback technique with the coordinated This technique also minimize the overhead to the system as
and uncoordinated checkpoints is well grounded but impose the inappropriate checkpoints are deleted from the database.
few restrictions. The rollback technique for the process This technique of the process recovery based on checkpoints
recovery is recommended for the systems with the large provided better results when compared with the existing
resources available to maintain a database in order to keep techniques. The technique is utilized for the wireless mobile
track/record of the updates made by the several processes in ad-hoc networks.
execution. Also this technique requires the replications of
the same data to be stored at the different locations to store Jaggi Singh and A.K [2011] proposed algorithm by using
different purposes. These includes, retrieving the process Self Stabilizing Tree. The person behind this research work
back to the particular pre-final state, to decide which of the described an algorithm for recording steady global picture of
updates made by the process need to be retained when it gets dynamic MANET network. In order to minimize the
aborted. snapshot related message as spanning tree, all other cluster
heads systematize themselves into a self stabilizing spanning
II. RELATED WORK tree. The result from tree will always provide result in
Neeraj Sharma and Ravneet Kaur[2015] in this paper shortest possible path. The result indicates that if number of
Dynamic Node Recovery approach have been presented. cluster is increased the number of control message
This approach is employed genetic algorithmic operations to decreased significantly. Furthermore, it can be concluded
ensure optimal recovery of checkpoints in case of node that proposed algorithm may efficiently works with multiple
failures. Refinement of some of the aspects of the existing initiators and dynamic topology.
base approach reduces the recovery time considerably,
thereby, improving the throughput of the network. It also A.K .Singh and P. K. Jaggi [2011] discussed a technique in
enhances the network lifetime as the proposed approach which the process itself needs to act jointly with the
leads to lower energy level drops in the nodes. checkpoints. The various checkpoints are stored in the
database, the process needs to commit certain safepoint
Poonam Saini and Shefali Aggarwal[2015] presented based on the updates made by the execution of the process.
Coordinated and Uncoordinated Check pointing in MANET. In this case, the process is not stick to the phenomenon that
The proposed checking point approach is based on it needs to commit to the first received checkpoint, as in the
movement of node. In this paper various techniques based case where the checkpoints are statically defined between
142
IJFRCSCE | October 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 10 141 – 146
_______________________________________________________________________________________________
the initial and final points. Here the criteria is followed, incoming and outgoing messages for the systems located at
where the updates are made by the process and based on the the different geographic locations. This technique basically
updates the safe point or pre-final state is achieved implements system in which the process itself creates the
dynamically. This type of techniques helps the system to checkpoints depending upon the requirement. This
cope up with the failures more efficiently, and the aborted technique prevents the inefficient utilization of the system
process can be easily retrieved to its pre-final state. resources. The message is sufficiently processed for making
Overhead of the system also get minimized. This technique the required updates and then it is committed for the next
is particularly designed for the mobile ad-hoc networks. safe point and all the updates made are recorded to the
dedicated database maintained for the purpose by the
Suparna Biswas et. al. [2011] discussed the rollback various systems at different locations. This technique is
process recovery technique and checkpoints for the mobile effective and prevents the system degradation.
ad-hoc networks. In this technique the sending mobile hosts
communicate the message with the two or more receiver Tong- Tony –Chang et al. [2007 ] discussed a new solution
hosts. The safe points are defined on the networks between to crash recovery. Processor will start from its most current
the sender host and the various receiver hosts. When the saved state in case of any failure. The result indicates
process/message commit to some safe point on the particular improved result compared to existing approach.
host to host or node to node network line, then the
update/record is made to the database/routing table of that Masakazu Ono and Hiroaki Higaki et al. [2007] presented
particular receiving host/node. The system is highly efficient a checking point approach by using flooding method. In this
as if the message can‟t be transmitted over one network, it is scheme, mobile host can able to communicate without
diverted to some other network with less traffic or overhead enough bandwidth and stable approach. By using flooding
based on the information stored in the routing tables of the method, checkingpoint request is being sent each mobile
receiving nodes. host of a node save the information of a node. In case, when
any node suffers from any lost information and then this lost
Qiangfeng Jiyang et al. [2008] discussed the process message/information is stored by its intermediate nodes.
retrieval technique based on the offline processing of the
Table 1: Comparison table

S.NO. NAME OF THE AUTHOR APPROACH USED CONCLUSION

1. Neeraj Sharma and Ravneet kaur Dynamic Node Recovery approach Enhances the network
[2015] lifetime
2. Tuli Kumar and P K Jaggi. The minimum number of nodes that the Reduces the energy
[2013] process needs to travel before committing the consumption and recovery
first safe point in the subnetwork. latency
S.NO. NAME OF THE AUTHOR APPROACH USED CONCLUSION

3. Hakkarinen Doug and Zizhong Multilevel diskless checkpointing approach This method improves
Chen [2013] expected execution time
4. Suparna Biswas, Priyanka Dey et A hybrid model of secure checkpointing The efficiency of the system
al. [2013] improves, the minimum
number of the hopes that the
process needs to travel for
committing safe point
reduces
5. A. K .Singh and P. K. Jaggi Manually creates the check points in the Increase he robustness of the
[2011] system, where the process has the probability system by reducing the
to get failed multiple failures in the
system.
6. Jaggi singh and A.K. [2011] The process updates routing tables of all the Efficient approach,
nodes in the subnetwork when the process decreasing the number of
commits to the safe point defined by the control messages
specific subnetwork
7. Suparna Biswas et. al, [2011] Mobility based checkpointing approach and The process retrieval is
trust based rollback recovery affordable and the ensures
the system robustness,
prevents system degradation
8. Jiyang Qiangfeng et al. [2008] The process retrieval and the message recovery Improved result and
process using safe points within the system Minimize the network
contention
143
IJFRCSCE | October 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 10 141 – 146
_______________________________________________________________________________________________
9. Tong- Tony –Chang et al. [2007] Rollback recovery approach in conjunction Improved result compared to
with checkpointing existing approach.

10. Masakazu Ono, Hiroaki Higaki Checkingpoint approach by using flooding Improving throughput of
et al. [2007] method. network

III. PROPOSED METHOD 3: 𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝐼𝑛𝑖𝑡𝑖𝑎𝑙 𝐶𝑙𝑢𝑠𝑡𝑒𝑟 𝑡𝑟𝑢𝑠𝑡

This paper proposes a novel method of checkpointing based
on trust value of nodes in MANET. confidence value of a 4: 𝑓𝑜𝑟 𝑖 𝑖𝑛 0 𝑡𝑜 𝑛, 𝑤𝑕𝑒𝑟𝑒 𝑛 𝑖𝑠 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑛𝑜𝑑𝑒𝑠
process depends on the subnetwork in which the process lies
at the specified time, the number of checkpoints committed 5:
by process moving from the current node to the destination 𝑡←
𝑁𝑐𝑡 −𝑃 𝑐𝑡
node, and the predefined required value of the confidence. , 𝑤𝑕𝑒𝑟𝑒 𝑡 𝑖𝑠 𝑡𝑟𝑢𝑠𝑡 𝑜𝑓 𝑛𝑜𝑑𝑒 𝑎𝑛𝑑 𝑃𝑐𝑡 𝑖𝑠 𝑡𝑕𝑒 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠
𝑃 𝑐𝑡

Confidence value of the network depends upon the count of 𝑐𝑙𝑢𝑠𝑡𝑒𝑟 𝑡𝑟𝑢𝑠𝑡 𝑎𝑛𝑑 𝑁𝑐𝑡 𝑖𝑠 𝑡𝑕𝑒 𝑛𝑒𝑥𝑡 𝑜𝑟 𝑡𝑎𝑟𝑔𝑒𝑡 𝑐𝑙𝑢𝑠𝑡𝑒𝑟 𝑡𝑟𝑢𝑠𝑡 𝑣𝑎𝑙𝑢𝑒
processes in the network having the confidence value higher
than the predefined, specified value i.e. threshold.

𝐶𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑝𝑟𝑜𝑐𝑒𝑠𝑠 𝑖𝑛 𝑛𝑒𝑡𝑤𝑜𝑟𝑘

6:
∑ 𝐶𝑕𝑒𝑐𝑘𝑝𝑜𝑖𝑛𝑡𝑠 𝑐𝑜𝑚𝑚𝑖𝑡𝑡𝑒𝑑 𝑏𝑦 𝑡𝑕𝑒 𝑝𝑟𝑜𝑐𝑒𝑠𝑠
= 𝐶𝐶 ←
𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑕𝑒𝑐𝑘𝑝𝑜𝑖𝑛𝑡𝑠 𝑖𝑛 𝑡𝑕𝑒 𝑛𝑒𝑡𝑤𝑜𝑟𝑘 1 − (𝑁𝑐𝑡 − 𝑃𝑐𝑡 ) , 𝑤𝑕𝑒𝑟𝑒 𝐶𝐶 𝑖𝑠 𝑐𝑙𝑢𝑠𝑡𝑒𝑟 𝑐𝑎𝑕𝑛𝑔𝑒 𝑐𝑜𝑢𝑛𝑡
The confidence of the process is the proportion of the safe 7: 𝑒𝑛𝑑 𝑓𝑜𝑟
points committed by the process to the total safe points in
the whole network. Where, each subnetwork has its own 8: 𝑖𝑓 𝐶𝐶 > 𝑡𝑕 𝑟𝑒𝑠𝑕 𝑜𝑙𝑑
defined safe point.
9:
The value of the confidence is calculated in the equation
above specified. The confidence value is basically the 𝑓𝑜𝑟 𝑖 𝑖𝑛 0 𝑡𝑜 𝑚, 𝑤𝑕𝑒𝑟𝑒 𝑚 𝑖𝑠 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑣𝑢𝑙𝑛𝑒𝑟𝑎𝑏𝑙𝑒 𝑛𝑜𝑑𝑒𝑠
proportion of the safe points that have been committed by
the process to the total safe points in the network. Where, 10:
each subnetwork has its associated check point. The
summated average of confidence values for the processes in 𝐶𝑎𝑙𝑐 (𝑚𝑠 , 𝑚𝑑 ), 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑢𝑠𝑖𝑛𝑔 𝑠𝑜𝑚 𝑎𝑙𝑔𝑜𝑟𝑖𝑡𝑕𝑚
subnetworks signifies the confidence of whole subnetwork.
The measure of confidence of the various subnetworks in 𝑚𝑠 𝑖𝑠 𝑟𝑒𝑐𝑜𝑣𝑒𝑟𝑦 𝑛𝑜𝑑𝑒 𝑎𝑛𝑑 𝑚𝑑 𝑖𝑠 𝑐𝑕𝑒𝑐𝑘𝑝𝑜𝑖𝑛𝑡𝑖𝑛𝑔 𝑛𝑜𝑑𝑒
network supports the efficiency of the whole system. If the
process with the higher confidence value moves within the 11: 𝑒𝑛𝑑 𝑓𝑜𝑟
networks then the overall system „support‟ increases. When
some process is found to be unsafe then that process 12: 𝑒𝑛𝑑 𝑖𝑓
committed to the first safe point between the source and
destination. The routing table of the hop associated with the 13: 𝑒𝑛𝑑
committed safe point within the subnetwork is updated.
When the process moves from the source to destination hop B. Performance Parameters
within the network, to find the optimum path, we must keep 1: Recovery Probability: Node recovery after failure is
in track the safe points to which the processes have defined as the probability of recovery. It depends on the
committed for the most. When the process has committed trust value of the node which needs to be recovered and
some safe point, it should be routed towards the another safe cluster change count.
point(in different subnetwork), which has been mostly
committed by the system processes. The whole criteria is 2: Residual Energy: The energy remaining at each node after
followed in the recursive manner, the operation is applied the transmission and reception cycle is termed as residual
for the different safe points in the different subnetworks, energy of the node. It is directly related to the network
until we obtain the feasible solution. lifetime of the node.

IV. RESULT AND DISCUSSION

A. PSEUDO CODE
In figure 1– 2 residual energy is compared with respect to
the simulation time for different number of nodes i.e. 50 and
1: 𝑆𝑡𝑎𝑟𝑡 20 respectively and in figure 3-4 probability of recovery is
compared with respect to the simulation time for different
2: 𝐼𝑛𝑖𝑡𝑖𝑎𝑙𝑖𝑧𝑒 𝑡𝑕𝑒 𝑛𝑜𝑑𝑒 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟𝑠 𝑎𝑛𝑑 𝑡𝑟𝑢𝑠𝑡 𝑣𝑎𝑙𝑢𝑒
number of nodes i.e. 50 and 20 respectively. In figure 5-6
144
IJFRCSCE | October 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 10 141 – 146
_______________________________________________________________________________________________
packet delivery delay is compared with respect to the
simulation time for different number nodes i.e. 50 and 20
respectively and in figure 7-8 packet delivery ratios is
compared with respect to the simulation time for different
number of nodes i.e. 50 and 20 respectively.

Fig 4: Probability of Recovery vs Simulation Time (20

nodes)

Fig 1: Residual Energy Vs Simulation Time (50 nodes)

Fig 5: Packet Delivery Delay vs Simulation Time (50 nodes)

Fig 2: Residual Energy vs Simulation Time (20 nodes)

Fig 6: Packet Delivery Delay vs Simulation Time (20 nodes)

Fig 3: Probability of Recovery vs Simulation Time (50

nodes)

145
IJFRCSCE | October 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 10 141 – 146
_______________________________________________________________________________________________
REFERENCES

1. A. K. Singh and P. K. Jaggi, "Staggered checkpointing and

recovery in cluster based mobile ad hoc networks."
Advances in Parallel Distributed Computing, pp 122-134,
2011.
2. Chih-Cheng Tseng and Kwang-Cheng Chen, "Organizing
an optimal cluster-based ad hoc network architecture by the
modified quine-mccluskey algorithm." IEEE
Communications Letters, Volume 11, Issue 1, January
2007.
3. Doug Hakkarinen and Z.C, "Multilevel diskless
checkpointing." IEEE Transactions on Computers, Volume
62, Issue 4, pp 772-783, April 2013.
4. Masakazu Ono and H.H, "Consistent Checkpoint Protocol
for Wireless Ad-hoc Networks." In PDPTA, pp 1041-1046,
Fig 7: Packet Delivery Ratio vs Simulation Time (50 nodes) 2007.
5. P.K Jaggi and A.K , "Message efficient global snapshot
recording using a self stabilizing spanning tree in a
MANET." International Journal of Communication
Networks and Information Security, Volume 3, Issue 3 , pp
247, December 2011
6. Qiangfeng Jiang, Yi Luo, and D. Manivannan, "An
optimistic checkpointing and message logging approach for
consistent global checkpoint collection in distributed
systems." Journal of Parallel and Distributed Computing,
Volume 68, Issue 12 , pp 1575-1589, December 2008.
7. Ravneet Kaur and Neeraj Sharma, "Dynamic node recovery
for improved throughput in MANET." In 1st International
Conference on Next Generation Computing Technologies
(NGCT), pp 325-330, IEEE, September 2015.
8. R Tuli and P.K , "Minimum process coordinated
Checkpointing scheme for ad hoc Networks." International
Journal on AdHoc Networking Systems (IJANS), Vol. 1,
Fig 8: Packet Delivery Ratio vs Simulation Time (20 nodes) No. 2 ,pp 51-63,November 2013.
9. Shefali Aggarwal and Poonam Saini, "Coordinated and
uncoordinated checkpointing in mobile ad hoc networks."
V. CONCLUSION In 2015 International Conference on Computing,
Communication & Automation (ICCCA), pp 611-615,
The mechanism of network node recovery is a topic of IEEE, May 2015.
concern and new techniques have been evaluated along with 10. Suparna Biswas, S. N “A Handoff based Checkpointing
the existing ones. Various checkpointing and node recovery and Failure Recovery Scheme in Mobile Computing.”
techniques are compared in the present work and their International Conference on Information Networking,
performance on various parameters like packet delivery September 2011.
ratio, throughput of the network. The nodes present in the 11. Suparna Biswas, Priyanka Dey and Sarmistha Neogy,
network are likely to be attacked and save their "Secure checkpointing-recovery using trusted nodes in
checkpointing data to the host cluster head. A node in MANET. "In 4th International Conference on Computer
and Communication Technology (ICCCT), pp 174-180,
mobile environment can pass through diverse clusters in its
IEEE, September 2013.
lifetime towards various attacks. The secure route selection 12. Taesoon Park and H.Y, "An asynchronous recovery scheme
in the network must solve this purpose of increasing based on optimistic message logging for mobile computing
overheads. The selection of the recovery node and the systems." In Proceedings. 20th International Conference
checkpointing node must also be selected in terms of the on Distributed Computing Systems, pp 436-443, IEEE,
available resources on them. In the methodology proposed, 2000.
the trust is increased according to the opinion dynamics rule.
Another important aspect is to find out the better
combination of both the algorithms (Firefly and GA). So
these aspects must be covered in the future scope and can be
compared with the existing results of our algorithm. This
work has also concluded that MANET has to handle number
of issues like stable storage, battery consumption, different
overheads, topological changes and traffic load with the
clusters. Moreover, we propose a multi-checkpointing
movement based trust model for checkpointing which
reduces overall overhead incurred while checkpointing.

146
IJFRCSCE | October 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________