Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h, Vo lume 3, Issue 2, February -2012 1

ISS N 2229-5518

A Peer to Peer Overlay Approach for Topology

Maintenance in Wireless Networks

J. Vijitha Ananthi, Jennifer S Raj

Abs tract - Over the Internet today, computing and communications environments are signif icantly more complex and conf usion than classica l distributed systems, lacking any centralized organization or hierarchical control. Heuristic algorithm is used to maintaining a topology in w ireless peer to peer netw orks. Here the topology of the Mobile Ad Hoc Netw ork (MANET) is maintained via peer to peer overlay nodes. During mobility localization a nd node movement are the major issues. To solve these issues a new approach is proposed named as “TMP – Topology Maintenance in MANET using Peer to peer overlay nodes” based on distributed and localized computing, to predict the node position in a topology tree. The perf ormance of the proposed method is analyzed in stable and mobile conditions by analyzing pow er consumption, active nodes, active links, hop stretch, p acket delivery ratio, netw ork lif e time and other Qos parameters.

Inde x Terms - Peer to Peer netw ork, Peer to Peer Overlay, Mobile ad hoc netw ork, Topology Maintenance, Hop stretch, Wireless Netw ork,

—————————— ——————————

1. INTRODUCTION

A Mobile Ad hoc NETw or k (MANET) does not use any existing infrastructur e or centr al administration to or ganize the wir eless devices within it. The nodes communicate in multi– hop, peer –to–peer node. Figur e 1: shows the peer to peer distr ibuted networ k. Over the Internet today, computing and communications envir onments ar e significantly mor e complex and confusion than classical distr ibuted systems, lacking any centralized or ganization or hierar chical contr ol. Ther e has been much inter est in emer ging Peer-to-Peer (P2P) netw or k over lays because they pr ovide a good substrate for cr eating lar ge-scale data shar ing, content distr ibution, and application-level multicast applications.
These P2P overlay netw or ks attempt to pr ovide a long
list of featur es, such as: selection of nearby peers, r edundant stor age, efficient sear ch/location of data items, data permanence or guar antees, hierar chical naming, tr ust

The way it w or ked was quite simple: a ser ver indexed all the files each user had. When a client quer ied Napster for a file, the centr al server w ould answer w ith a list of all indexed clients who alr eady possessed the file. Napster -like networ ks ar e known now as first generation netw or ks. Such networ ks didn’t have a complicated implementation and often r elied on a central server (hybr id P2P). The centr al ser ver model makes sense for many r easons: it is an efficient way to handle sear ches and allows to r etain control over the netw or k. How ever , it also means ther e is a single point of failur e. When lawyers decided Napster should be shut down, all they had to do was to disconnect the server .

Peer

Peer

and authentication, and anonymity. P2P networ king has existed for quite some time, it has only been popular ized r ecently and w ill pr obably be subj ect to even bigger r evolutions in the near futur e. Napster was the fir st P2P application which r eally took off.

Peer

Peer

———— ——— ——— ——— ———

J. Vijitha Ananthi is currently pursuing ma sters degree program

in communica tion systems from Ka runya University, Coimba tore,

India.

Ema il: vijithaana nthi@gma il.com

Jennifer S. Ra j working a s a Assista nt Professor in Dept of Electronics a nd Communica tion Engineering in Ka runya University, Coimba tore, India.

Ema il: jennifer.ra j@gma il.com

Figur e 1: Peer to peer netw or k
Gnutella was the second maj or P2P netw or k. After Napster ’s demise, the cr eator s of Gnutella wanted to cr eate a decentralized networ k, one that could not be shut down by simply tur ning off a server . At first the model did not scale because of bottlenecks cr eated w hilst searching for files. Fast Track solved this pr oblem by r ender ing some nodes mor e capable than others. Such networ ks ar e now
known as second gener ation netw or ks and ar e the most

IJSER © 2012 http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h, Vo lume 3, Issue 2, February -2012 2

ISS N 2229-5518

widely used nowadays. Thir d generation networ ks ar e the new emer ging P2P networ ks. They ar e a r esponse to the legal attention P2P netw or ks have been r eceiving for a few year s and have built-in anonymity featur es. They have not yet r eached the mass usage main second generation networ ks curr ently endur e but this could chan ge shortly.
Fr eenet is a good example of a thir d generation. In
this algorithm, w e ar e using thir d generation P2P netw or k. An over lay netw or k is a virtual netw or k of nodes and logical links that is built on top of an existing netw or k w ith the pur pose to implement a netw or k service that is not available in the existing networ k. Curr ent Inter net P2P applications typically pr ovide locator functions using time- to-live (TTL) contr olled-flooding mechanisms. With this approach, the querying node wr aps the query in a single message and sends it to all known neighbors. The neighbor s then check to see whether they can r eply to the query by matching it to keys in their internal database. If they find a match, they r eply; otherwise, they forwar d the query to their own neighb ors and incr ease the message’s hop count. If the hop count passes the TTL limit, forwar ding stops. The TTL value thus defines a boundary or ‚horizon‛ for the query that contr ols its pr opagation. How ever , flooding-based systems don’t scale w ell b ecause of the bandw idth and pr ocessing r equir ements they place on the networ k, and they pr ovide no guarantees as to lookup times or content accessibility.
Over lay netw or ks can addr ess these issues. Over lay networ ks have a networ k semantics layer above the bas ic transport pr otocol level that or ganizes the networ k
topology accor ding to the nodes’ content. The main applications of over lay netw or k ar e r outing, addr essing, security and multicast. In this paper , peer to peer over lay multicast the information to every node. Due to that multicast, all nodes r espond to the neighbor s whether it’s applicable or not. So it consumes mor e ener gy, to save the ener gy nodes will act as selfish peer s in peer to peer over lay. Selfish peer means it does not forward any data to the neighbors and also using the unicast to tr ansmit the data using AODV pr otocol. Existing appr oaches ar e benchmar k algor ithm and heur istic algor ithm thr ough this algor ithm w e can minimize the cost and also maintain the topology.

2. RELATED WORK

Fabr ikant et al. [1-3] intr oduce a netw or k cr eation game. In the game, nodes need to pay the links to r each the destination. in this distance based on the hop count. Links ar e constant, so topology changed due to some nodes enter or leaves in the netw or k after that changes it comes to its
or iginal position means Nash equilibr ium exists. Nodes ar e selfish, trying to have fewer dir ect neighbor s while keeping the entir e networ k as close as possible. A system par ameter , α, catches the trade-off betw een the cost and benefit.
Moscibr oda et al. [4-6] extends the w or k of Fabr ikant et al., examining the effect of the networ k cr eation game on P2P topologies. The hop count is r eplaced by the str etch as
the distance metr ic and over lay links ar e dir ectional; otherwise the model r emains the same. The authors pr ovide an upper and lower bound on the pr ice of anar chy. They go on to pr ove that ther e exist cases for which no Nash equilibr ium exists, meaning that even in the absence of peer churn, the topology will never stabilize.
Chun et al. [7-10] simulate a P2P over lay cr eation game
using Fabr ikant et al.’s model. The cost to each neighb or may be differ ent (e.g., congested links have a higher cost), and the number of neighbor s is bounded. The authors use a heur istic r andom local search algorithm to calculate the
Nash equilibrium graphs. They show that w ithout constr aining the node degr ee, the r esulting topologies ar e near-star configur ations, in which a small number of nodes will maintain a lar ge number of connections. This allows other nodes to ‘‘fr ee r ide’’ off of those nodes by connecting to them and achieving both a small number of connections and also a low distance metric. This r esults in less r esilient topologies because a small number of node failur es w ill r esult in a lar ge degr ee of disconnection.
Afzal Mawji[11-12] pr oposed Both MANETs and P2P networ ks r equir e nodes to help one another in or der to make the networ k useful. P2P file shar ing netw or ks r equir e
nodes to shar e files, and MANETs r equir e nodes to forward data. Unfortunately, helping others comes at a cost. By shar ing files and forwarding data, mobile nodes ar e using their ener gy, bandwidth, CPU, and other r esources. When r esour ces ar e used by others, ther e is less available for the node itself. Node will act as a fr eeloa ders and selfish to r ise their r esour ces. Fr eeloaders ar e users that download files themselves, but do not shar e anything in r eturn. Selfish nodes do not forward data. Users cooperate with one another based on their r emaining ener gy level.
Afzal Mawji and Hassam S. Hassanein [13] stimulate
bootstr apping the tw o netw or ks using bootstr ap algor ithm. In this Nodes discover overlays by finding peers that ar e alr eady participating in a P2P networ k. This is accomplished by having the node first examine its local cache for pr eviously–known peers. If none ar e found, or the information is stale, the node multicasts a join r equest. When a peer r eceives a r equest, it may multicast a r eply and all nodes in the networ k cache the infor mation. The node wanting to j oin the over lay then selects the best peers based on ener gy.

IJSER © 2012 http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h, Vo lume 3, Issue 2, February -2012 3

ISS N 2229-5518

Afzal Mawj i and Hassam S. Hassanein [14] suggest the effects of using netw or k coding in a P2P file shar ing over lay in a cooperative mobile ad hoc netw or k. Because ther e is no centralized author ity and no infrastructur e, a tracker node cannot be used, as is commonly done in P2P networ ks. Instead we make use of multicasting for efficient communication within the overlay. The algorithm pr esented, called Deluge, uses the idea of a server node multicasting block, something that is impr actical on the Internet, but is possible in MANETs. Clients r equest a certain number of blocks fr om multiple differ ent ser vers, depending on the cost of acquir ing them, r esulting in multipoint–to– multipoint communication. In this cost is measur ed as hop distance. Ther efor e, nodes download files fr om the closest servers, which r educe overall networ k ener gy consumption.
Afzal Mawji, Hassam S. Hassanein, [15] suggest the most P2P over lay topology contr ol algor ithms assume that
peers is cooperative. Unfortunately, peers ar e selfish in many cases. They seek to minimize their own costs, in the case of MANETs, to minimize the number of links to other peers and the distance to all destinations. Several studies investigate the impact of selfish peers on the topology in the context of non-mobile netw or ks. How ever , they study the theor etical bounds or r equir e peers to have global know ledge to constr uct the over lay. No practical over lay topology contr ol algor ithm, even for non-mobile netw or ks, exists. As well, a study of the impact of selfish peers in the context MANETs is needed. He pr oposed the heuristic algor ithm w ith low k metr ic and high degr ee constraint.

3. PRELIMIN AR IES

Existing P2P topology contr ol schemes designed for wir ed networ ks ar e able to accommodate a changing topology due to the expectation that peer s will constantly be j oining and quitting. Most P2P overlay control algor ithms assume that peer s ar e co-operative. Due to that co-oper ative nodes, ener gy consume w ill mor e. Degr ee- constr ained minimum spanning tr ee (DCMST) is a degr ee- constr ained spanning tr ee in with the sum of its vertices has the minimum possible sum. Finding a DCMST is an NP- Har d pr oblem. Mor e Resilience in Nash equilibr ium, after the topology changes due to enter ing some nodes, it’s very difficult to constr uct the topology to its original position.

4. PROPOSED WORK

Heur istic algor ithms that can solve the pr oblem in polynomial time have been pr oposed, including Genetic and Ant-Based Algor ithms. The heuristic algor ithm has a fair ly low k metr ic due to the degr ee constraint. The
heur istic technique is an experienced based technique for pr oblem solving, learning and discovery. This method is used to speed up the pr ocess of finding a good enough solution. A genetic algor ithm (GA) is a search heur istic that mimics the pr ocess of natural evolution. This heur istic is r outinely used to generate useful solutions to optimization and search pr oblems. Genetic algor ithms belong to the lar ger class of evolutionary algor ithms (EA), which generate solutions to optimization pr oblems using techniques inspir ed by natural evolution. In existing approach, they use two algorithms and compar ed those r esults also. The two algor ithms ar e benchmar k algorithm and heur istic algorithm. Her e w e ar e going to pr opose TPP i.e., (Topology maintenance in P2P over lay) algorithm. In this algor ithm to find the minimum cost betw een the two peers using this below equation,

(1)

Wher e, Cn is the minimum cost betw een the two peers. E is the ener gy level of the neighbor ing peer and D is the str etch distance betw een the selected peer and initial peer . Distance based on the str etch. Str etch is defined as the ratio of number of physical hops to the shortest physical distance. The total cost is the sum of all peer costs

(2)

Wher e, T.C r epr esents total cost of the peer to peer over lay networ k, Cn is the cost betw een the two peer s and N r epr esents number of nodes in the peer to peer over lay networ k. Using this above tw o Equations w e can find only the minimal cost but w e cannot maintain the topology because of r esilience. Resilience is nothing but r esuming its or iginal position after the node enter s or leaves in the networ k. So over come that this w e go for heuristic algor ithm. In heur istic appr oach, w e can find the minimum cost and also maintain the topology after the node enter s or leaves in the netw or k. It does not need of global knowledge to transfer the data. It uses the local knowledge to select the best peer s in the neighbors to tr ansfer the data. To find the minimal cost betw een the peers w e can use the same equation of the benchmar k algorithm.
If a node wants to enter into the peer to peer over lay means, it can select the best peer based on the hop count and ener gy level of the mobile ad hoc netw or k. Mostly it selects the best peer alone.

4.1 Algorithm : Peer To Peer Overlay

1. Connected graph G (V, E)
2. Over lay netw or k graph G1 (V, E)
3. Divide G (V, E) into G1 (V, E) & G11 (V, E)

IJSER © 2012 http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h, Vo lume 3, Issue 2, February -2012 4

ISS N 2229-5518

4. G1 Distr ibuted of (V, E); G1 € G
5. G1 1 Cluster MANET of (V, E); G11 € G
6. Fr om G11 (V, E), V1 act as intermediate node
7. Fr om G1 (V, E) any node sends ‘P’ to V 1 ;
8. If it r eceives ‘A’, it will form as G1 (V, E)
9. Else it select V2 node in G11 (V, E)

Algorithm explains that G (V, E) is a whole networ k which is our input and G1 (V, E) is overlay networ k which is our output. Fr om that G (V, E), w e have to split the networ ks as peer to peer networ k G1 (V, E) and mobile ad hoc networ k G11 (V, E). Peer to peer netw or k should be distr ibuted networ k, mobile ad hoc netw or k based on cluster formation. From mobile ad hoc networ k selects the one node act as an intermediate node i.e., V1 . Fr om P2P networ k any one of the node sends packet ‘P’ to the intermediate node. In that packet contain ID and 0, 0 r epr esents r equest. If intermediate node acknowledged ‘A’ to the peer to peer netw or k, it will for m as over lay networ k G1 (V, E). Else it selects the another node fr om G11 (V, E). In these two netw or ks have one intermediate node to send the r equest fr om peer node to MANET node.

MANET

n

n

P2P ove rlay

n

Figur e 2: shows that the peer to peer over lay , P2P networ k and MANET used to connect with one intermediate node.
W e used to select the MANET nodes for P2P over lay through that intermediate node alone. So w e have to concentrate mor e on that intermediate node. They ar e using this intermediate link r epeatedly for selecting the nodes in mobile ad hoc networ ks. When the intermediate node fails due to ener gy, w e have to change the link to other node. In this algorithm, w e have to find the minimal cost betw een the tw o peers, overlay the tw o networ ks w e h ave to maintain the ener gy level of the intermediate nodes and maintain the topology. The main application is use as a networ k cr eation game. It doesn’t have connectivity issue and it has a high thr oughput. It is efficient topology maintenance for peer to peer over lay in mobile ad hoc networ ks. It may r educe the total cost netw or k-wide. It r educe the ener gy consumption and to impr ove the r esponse time and r eliability.

5. EXPERIMENT AL RESULTS

W e now evaluate the per formance of TMP algorithm and compar e its per for mance to heur istic algor ithm and also maintain the ener gy level of intermediate node betw een P2P netw or k and MANETs. These algorithms ar e implemented in the networ k simulator ns-2. In our simulations, w e use 100 MANET nodes , w ith the number of nodes participating in a P2P overlay varying fr om 50 to 100 incr ements of 10. The networ k ar ea is 1500m* 1500m, the transmission r ate is 54 Mbps, and the communication r ange is 240m by default. Her e, using Omni dir ectional antennas by all nodes.

n

P

P

ni

P P

Inte rme diate

node

Figur e 2: peer to peer overlay netw or k
So w e have to maintain the ener gy level of the intermediate node. Inter mediate nodes ar e coming fr om mobile ad hoc networ k. If the ener gy fails in first intermediate node, it will change the link to next node.
Figur e: 3 Ener gy Analyses
Figur e: 3 show the ener gy analysis betw een the time and ener gy. Initially a node having 500 j oules, after a particular time it may r eceive or transmit based on that application; it w ill r educe the par ticular ener gy level. But in

IJSER © 2012 http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h, Vo lume 3, Issue 2, February -2012 5

ISS N 2229-5518

TMP algorithm after the ener gy r educes particular thr eshold level it r eplace the new node instead of old node. After the time of 60 second, heur istic algor ithm (HA) loses ener gy. But TMP it incr ease the ener gy level and also r educes the ener gy consumption. One impor tant aspect of ad-hoc networ ks that was ignor ed by these studies is ener gy-efficiency. In r eal life systems, ener gy consumption is a maj or issue. For many ad-hoc netw or ks, the nodes ar e small and portable, imposing str ingent constr aints on the battery size and pow er . In this study, w e consider ed a networ k wher e each mobile node has a limited initial ener gy.
As a node sends, r eceives or forwar ds packets, the ener gy of the node is decr emented accordingly. Once the ener gy of the node r eaches zer o, the node is shut -down (a
‚node death‛) and is consider ed terminated by the system.
Since the ad-hoc r outing pr otocol deter mines which nodes will forwar d the packets and the amount of r outing
over head each node needs, the type of pr otocol definitely affects the ener gy per formance of the system. The pr otocols affect the ener gy dynamics in tw o ways – fir st, the r outing over head affects the amount of ener gy used for sending and r eceiving the routing packets, and second, the chosen r outes affects which nodes will have a faster decr ease in ener gy.

Figur e 4: Packet Delivery Ratio
Packet delivery ratio, packet delivery time and packet dr op ratio ar e some of the qos parameter s. Packet delivery ratio is obtained by dividing the number of data packets corr ectly r eceived by the destinations by the number of data packets or iginated by the sources. Figur e 4: shows the packet delivery ratio, in this defining sequence number s with the r eceived packets. Packet delivery ratio is defined as the r atio of packets deliver ed to the destination to those generated by the CBR sour ces . Compar ed to heur istic algor ithm, TMP algor ithm incr eases the packet delivery rate.
Figur e 5: shows the cost vs. time. When the time incr eases cost may incr ease or decr ease. Initially cost incr eases because of the distance and ener gy. After the mobility distance may decr ease and the ener gy also decr ease, due to this cost also decr eases. Fr om this time of
50 sec nodes moves dynamically, so distance incr eases fr om
the source node. So cost also incr eases.

Figur e 5: Cost
In this networ k having number of nodes, each node having links betw een the other nodes. Due to this selfish peer , some of the nodes act as inactive links. So we have to plot the gr aph between nodes vs. total number of links and the active links.

Figur e: 6 Nodes vs. Active links
Compar ed to the total number of links to number of active link is very lesser . So w e go for the benchmar k algor ithm. In this benchmar k algor ithm, it will check the connection either the node having very lesser active links means it w ill cr eate the connection and also find the minimum cost. Figur e: 6 show the active links vs. nodes. Initially node having some active links, after a particular time interval some of the nodes w ill act as selfish nodes. Due to these selfish nodes some links may act as inactive.

IJSER © 2012 http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h, Vo lume 3, Issue 2, February -2012 6

ISS N 2229-5518

Then we ar e going see about the existing appr oach. In existing it w ill select the node and also it checks the links in node, its alr eady having number of links higher than the active link means it will find out the cost of the entir e topology, otherw ise it w ill cr eate the connection and also find the cost.
The hop count is a measur e of distance acr oss an IP- based netw or k. It is a count of the number of r outers an IP packet has to pass thr ough in or der to r each its destination. Hop count is usually not used by itself, since any in betw een r outer or cable may have or be subj ect to varying data thr oughput (bandwidth), load (see: quality of service), r eliability (especially of cable), and latency . Hop counts ar e often useful to find faults in a netw or k (see: Time to live), or to discover if r outing is indeed corr ect.

Figur e 7: Compar ison of Hop count
Figur e 7 shows the above graphs plotted the
comparison of hop count betw een the heur istic algorithm
and topology maintenance in P2P algorithm in P2P over lay networ k. In this HA to r each the destination it needs mor e hop count. But in TMP in P2P over lay r educes the hop count thr ough the minimum cost.
The below gr aph shows the compar ison of
throughput betw een mobile nodes and stable nodes. In stable nodes incr eases the thr oughput w ith r espect to time. Compar ed to mobile nodes, stable node incr eases the throughput and also packet delivery rate.

Figur e 8: Compar ison of thr oughput
Neighbor knowledge in mobile ad hoc netw or ks pr ovides important functionality for a number of pr otocols. An appr oach w her e each node acquir es neighbor know ledge by observing not only hello packets but also flooded packets is pr esented. Analysis r esults show that this method offer s significant impr ovement over the or iginal scheme. Figur e: 9 show the contr ol over head vs. time. Fr om these gr aph w e can identify in AODV pr otocol ther e is no congestion due to control overhead. Control messages ar e limited only w ith var ies time. Compar ed to heur istic algor ithm, TMP algor ithm r educes the control over head.

Figur e 9: Contr ol overhead

IJSER © 2012 http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h, Vo lume 3, Issue 2, February -2012 7

ISS N 2229-5518


Finally TMP algor ithm is fair ly stable, r elative to minimum cost algor ithm and when the degr ee-constrain is r elaxed, it minimizes its cost.

7. REFERENC ES

Figur e 10: Normalized r outing overhead
Figur e 10 pr esents the contr ol over head in normalized r outing load. Nor malized r outing load is the ratio of the number of control packets pr opagated by every node in the networ k and the number of data packets r eceived by the destination nodes. This value hence r epr esents the pr otocol efficiency. When ther e is no mobility, it has the least value.
This r esult is expected because AODV pr otocols generate mor e contr ol packets while building multiple r outes. When the mobility incr eases, nor malized r outing over head also incr eases.

6. CONCLUSION

This algor ithm pr ovides an examination of over lay constr uction and thus r esilient fr om the maintenance of intermediate link ener gy betw een the peer to peer networ k and mobile ad hoc netw or k. Mostly peer to peer netw or ks have co-operative nodes in mobile ad hoc netw or k, but in this algor ithm some of the nodes act as selfish peer s to minimize the ener gy consumption and also r educe the distance to their destinations. Fr om this identify the QOS parameter of this pr oposed algor ithm. The per formance of this algor ithm is compar ed to benchmar k algorithm. Compar ed to heur istic algorithm (HA), Topology Maintenance in P2P (TMP) algorithm is mor e stable and feasible in thr oughput, packet delivery ratio and hop count. This appr oach uses the link based on minimum cost and high level ener gy. Thr ough this algor ithm ener gy consumption w ill be r educed due to selfish peers in P2P networ k. In MANET, when the node decr eases in ener gy level it will hand over the link to other node. The r esults w er e obtained and also the gr aphs w er e plotted for throughput, packet delivery ratio, packet dr op, control over head and cost of the ent ir e topology.
[1]. Afzal Mawjia, Hossam Hassanein, Xiangyang Zhang. ‚Peer-to-peer over lay topology contr ol for mob ile ad hoc netw or ks‛ Telecommun ications Resear ch Lab
[2]. Fabr ikant, A. Luthra, E. Maneva, C.H.
Papadimitr iou, S. Shenker , ‚On a networ k cr eation
game,‛ in: ACM Symposium on Principles of

Distributed Computing, 2003, pp. 347–351.

[3]. L.M. Feeney, ‚An ener gy -consumption model for per formance analysis of r outing pr otocols for mob ile ad hoc netw or ks,‛ Mobile Networks and Applications 6 (3) (2001) 239–250.
[4]. S.-T. Par k, A. Khr abr ov, D. Pennock, S. Lawr ence, C. Giles, L. Ungar , ‚Static and dynamic analysis of the Inter net’s susceptib ility to faults and attacks,‛ in: IEEE INFOCOM, vol. 3, 2003, pp. 2144–2154.
[5]. T. Moscibr oda, S. Schmid, R. Wattenhofer , ‚On the
topologies formed by selfish peer s, in:‛ 5th

International Workshop on Peer-to-Peer Systems, Febr uary 2006.

[6]. Hossam S. Hassanein, Yu Yang, Afzal Mawj i. ‚A new appr oach to service discovery in w ir eless mob ile ad hoc networ ks‛. IJSNet, 2007: 135~145
[7]. Yu Yang, Hossam S. Hassanein, Afzal Mawj i. ‚Efficient Ser vice Discovery forW ir eless Mob ile Ad Hoc Netw or ks.‛ AICCSA'2006. pp.571~578
[8]. B.-G. Chun, R. Fonseca, I. Stoica, J. Kubiatowicz,
‚Character izing selfishly constr ucted over lay
r outing networ ks‛, in: IEEE INFOCOM, vol. 2,

2004, pp. 1329–1339.

[9]. Ngai-Pan Chow , Ajit Singh, Afzal Mawj i. Architectur e of a Reconfigurable and Reusable Web -Based Envir onment. Inter national Confer ence on Inter net Computing'2004. pp.42~45
[10]. Afzal Mawji, Aj it Singh. Load Balanced Exhaustive Networ k Data Pr ocessing With a Parallel Cluster . PDPTA'2004. pp.207~213
[11]. H. Hassanein, Y. Yang, and A. Mawji, ‚A new
approach to service discovery in wir eless mobi le
ad hoc networ ks,‛ Int. J. Senso r Networks, vol. 2, no.
1/2, pp. 135–145, 2007
[12]. Afzal Mawj i and Hossam Hassanein ‚Incentives
for P2P File Shar ing in Mobile Ad HocNetwor ks ‛

ICC'2008. pp.2248~2252

IJSER © 2012 http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h, Vo lume 3, Issue 2, February -2012 8

ISS N 2229-5518

[13]. Andr ew Ka-Ho Leung and Yu-Kw ong Kw ok, ‚On Topology Contr ol of Wir eless Peer-to-Peer File Shar ing Netw or ks: Ener gy Efficiency, Fairness and Incentive.‛
[14]. Afzal Mawji, Hossam S. Hassanein.
‛Bootstrapping P2P Overlays in MANETs‛

GLOBECOM'2008, pp.763~767

[15]. Afzal Mawj i, Hossam S. Hassanein. ‛Efficient Multipoint P2P File Shar ing in MANETs.‛ GLOBECOM'2009. pp.1~6 Computing, Queen’s Univer sity, Kingston, Ontar io, Canada, K7L 3N6
Department of Computer Science, King Saud
Univer sity, Saudi Arabia

Author ’s B iogr aphy


Ms. J. Vij itha Ananthi r eceived B.E degr ee in Electr onics and communication fr om Anna Univer sity, in 2010. Since July 2011, she has been a student of M.Tech in Communication systems in Karunya
Univer sity, Coimbator e. Her r esear ch inter est includes
over lay in w ir eless netw or ks.

Mrs. Jennifer S Raj r eceived the Masters degr ee in communication System fr om SRM University, India. Curr ently she is pur suing Ph.D in Information and Communication Engineer ing at Anna Univer sity of Technology, Coimbator e, India and
wor king as a Assistant Pr ofessor at Kar unya University,
Coimbator e, India. Her inter ests ar e in w ir eless netw or ks with self or ganization and topology contr ol str uctur es. She is a life member of ISTE, India. She is book r eview er for Tata Mc Gr aw hill publication and publishes mor e than fifteen r esearch articles in the j ournals and IEEE
confer ences.

IJSER © 2012 http :// www.ijser.org