PAPER ID: IJIFR/V 1/E 2/016

Transcription

1 Dynamic Topology Maintenance in Unstructured Peer to Peer Overlay Networks J. Vijitha Ananthi 1, R.Hari Priya 2, R.Nithya 3, R.Karthikeyan 4 Assistant professor 1, Indus college of Engineering, Coimbatore, India Graduate Student 2, Indus college of Engineering, Coimbatore, India Graduate Student 3, Indus college of Engineering, Coimbatore, India Faculty of ECE 4, Indus college of Engineering, Coimbatore, India Abstract P2P networks are used for instant message communication, distributed computing and files sharing. In the present Internet world, computing communication is very difficult in centralized network compared to distributed system. Topology maintenance in MANET using Peer to Peer algorithm is used for maintaining the topology without any link failure in wireless networks. Here maintaining the topology in mobile ad hoc network via peer to peer overlay nodes. When mobility occurs, localization of node and node movement are the major issues. To overcome these issues proposed a new algorithm which is named as TPP- Topology Maintenance in Partial mesh in P2P overlay nodes will maintain the location of the mobile nodes in a distributed network and also easy to find out the location of node in a defined topology. The performance of this algorithm is compared with the existing algorithm which is analyzed in mobile condition with the speed of 20m/s. Keywords: Hop count, Unstructured overlay, Peer to peer network, QoS parameters 1 INTRODUCTION P2P overlay networks used for many features like easily can select the nearby peers, storage data, searching and finding a location, data permanence or guarantees, hierarchical structure, security and safely transmit the data through authentication. P2P network has existed for quite some time, it has only been popularized recently and will probably be subject to even bigger revolutions in the near future There are two types of peer to peer overlay one is structured and another one is unstructured overlay. Unstructured network do not provide any algorithm for organization or optimization of network connections. Structured peer-to-peer system is the pure network consists solely of equipotent peers. It s having only one routing protocol; it does not prefer nodes for any infrastructure. In structured peer-to-peer networks, connections in the overlay are fixed. In proposed algorithm we are using unstructured overlay, existing algorithm uses structured overlay. To overcome the issues of structured overlay, unstructured peer to peer network is used which is shown in figure1. Mesh topology will form like a full mesh topology or partial mesh topology. In Full Mesh topology each node is connected to every other node in the network. In Partial Mesh topology every node is not connected to each node in the network. Due to many interconnections much of the cable is required for implementation of mesh topology so it is quite expensive which is shown in figure 2. Here one node fails, remaining nodes maintain the connection. It doesn t bother about the node failure. So it s reliable and also it can easily coupled with star or ring topology to form a hybrid topology. Wireless mesh networks were originally developed for military applications and are typical of mesh architectures. Over the past decade 52

2 the size, cost, and power requirements of radios has declined, enabling more radios to be included within each device acting as a mesh node. Unstructured node Figure 1: Unstructured Network Mesh node Figure 2: Partial Mesh Network Peer to peer mobile mesh networks which form like a hybrid topology, in these individual nodes on each network connected to each other devices through radio range. The major problem occur in this are scalability. To reduce that here using partial mesh network. In fully mesh network occurs in optimal path due to this, it will produce scalability. 2 RELATED WORK Moscibroda et al. [1-7] extends the work of Fabrikant et al., examining the effect of the network creation game on P2P topologies. The hop count is replaced by the stretch as the distance metric and overlay links are directional; otherwise the model remains the same. The authors provide an upper and lower bound on the price of anarchy. They go on to prove that there exist cases for which no Nash equilibrium exists, meaning that even in the absence of peer churn, the topology will never stabilize. Chun et al. [8-11] simulate a P2P overlay creation game using Fabrikant et al. s model. The cost to each neighbor may be different (e.g., congested links have a higher cost), and the number of neighbors is bounded. The authors use a heuristic random local search algorithm to calculate the Nash equilibrium graphs. They show that without constraining the node degree, the resulting topologies are near-star configurations, in which a small number of nodes will maintain a large number of connections. This allows other nodes to free ride off of those nodes by connecting to them and achieving both a small number of connections and also a low distance Hassam Hossain et.al [12-15] Suggest the effects of using network coding in a P2P file sharing overlay in a cooperative mobile ad hoc network. Because there is no centralized authority and no infrastructure, a tracker node cannot be used, as is commonly done in P2P networks. Instead we make use of multicasting for efficient communication within the overlay. The algorithm presented, called Deluge, and uses the idea of a server node multicasting block, something that is impractical on the Internet, but is possible in MANETs. Therefore, nodes download files from the closest servers, which reduce overall network energy consumption. 53

3 3 PRELIMINARIES P2P networking has existed for quite some time; it has only been popularized recently and will probably be subject to even bigger revolutions in the near future. Napster was the first P2P application which really took off. The way it worked was quite simple: a server indexed all the files each user had. When a client queried Napster [12] for a file, the central server would answer with a list of all indexed clients who already possessed the file. Napster-like networks are known now as first generation networks. Such networks didn t have a complicated implementation and often relied on a central server (hybrid P2P). The central server model makes sense for many reasons: it is an efficient way to handle searches and allows retaining control over the network. However, it also means there is a single point of failure. When lawyers decided Napster should be shut down, all they had to do was to disconnect the server. a. Existing approach An existing algorithm which is TMP algorithm used to maintain the topology in mobile ad hoc networks. From this algorithm to find the minimum cost between the two peers using this below equation, (1) Where, Cn is the minimum cost between the two peers. E is the energy level of the neighboring peer and D is the stretch distance between the selected peer and initial peer. Distance based on the stretch. Stretch 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 b. TMP Algorithm 1. Connected graph G (V, E) 2. Overlay network graph G 1 (V, E) 3. Divide G (V, E) into G 1 (V, E) & G 11 (V, E) 4. G 1 Distributed of (V, E); G 1 G 5. G 11 Cluster MANET of (V, E); G 11 G 6. From G 11 (V, E), V 1 act as intermediate node 7. From G 1 (V, E) any node sends P to V 1 ; 8. If it receives A, it will form as G 1 (V, E) 9. Else it select V 2 node in G 11 (V, E) TMP Algorithm explains that G (V, E) is a whole network which is our input and G 1 (V, E) is overlay network which is our output. From that G (V, E), we have to split the networks as peer to peer network G 1 (V, E) and mobile ad hoc network G 11 (V, E). The peer to peer overlay, P2P network and MANET used to connect with one intermediate node. We used to select the MANET nodes for P2P overlay through that intermediate node alone. So we have to concentrate more on that intermediate node. They are using this intermediate link repeatedly for selecting the nodes in mobile ad hoc networks. When the intermediate node fails due to energy, we have to change the link to other node. In this paper, we have to find the minimal cost between the two peers, overlay the two networks we have to maintain the energy level of the intermediate nodes and maintain the topology. 54

4 4 PROPOSED WORK In proposed work estimates the new algorithm named as topology maintenance in partial mesh in P2P overlay (TPP) algorithm. Mesh networking has recently become popular in deploying wireless networks. Similarly with the growth of internet browsing, implementation of a P2P overlay for web caching has identified as a major requirement to reduce the latency of the web connect delivery and the network bandwidth. So emerging these two concepts together facilitates to overcome dynamic and mobility issues and deliver advantages to all the stake holders in the TPP algorithm which explains clearly below: c. TPP Algorithm 1. G 1 Unstructured overlay of (V, E); G 1 G 2. G 11 Wireless partial Mesh of (V,E);G 11 G 3. FromG 11 (V,E), node G 11 (E 1 ) as mobile node 4. G 11 (E 1 ) sends P to G 1 (E 1 ) G 1 (V,E) 5. From G 1 (E 1 ) sends A to G 11 (E 1 ) G 11 (V, E). 6. Instead of G 11 (E 1 ) as replace a new node as G 1 (E 1 ). TPP Algorithm explains the peer to peer overlay in wireless partial mesh network (POW). Input is unconnected graph G (V, E) which is divided into two networks one is unstructured overlay G 1 (V, E) another one is wireless partial mesh network G 11 (V, E). Here, both different networks having mobile nodes. The wireless mesh node change starts to move from one place to another place, the link will be failure between the mobile nodes. After a particular time new unstructured mobile node replace instead of mesh node and also it handover the link and starts to transmit the message which is shown in figure 3. Figure 3: Peer to Peer Overlay in Partial Mesh In figure 3 explains the peer to peer overlay in wireless partial mesh network, from this two different networks are bootstrapping each other, and to produce new network. Here, mesh and unstructured peer to peer networks having advantages and disadvantages. To overcome the issues of both network and to make the entire network useful. Here proposed TPP algorithm using unstructured peer to peer network for file 55

5 transferring. The main application is use as a network creation game. It doesn t have connectivity issue and it has a high throughput. 4 EXPERIMENTAL RESULTS The performance of our proposed TPP algorithm, and compare its performance to that of the benchmark algorithms. The heuristic algorithm was implemented in the network simulator ns The underlay topologies resulting from the simulation were then fed into the benchmark algorithms solver, which are written in the C++ programming language, to determine the minimum cost and TMP algorithm. d. PARAMETERS AND RESULTS In our simulations, we use 50 partial mesh nodes, with the number of nodes participating in a P2P overlay varying from 50 to 100 in increments of 10. The network area is 1000m 1000 m, the transmission rate is 54 Mbps, and the communication range is 240 m by default. The two-ray ground radio propagation models along with an omnidirectional antenna are used by all nodes. We use the random waypoint mobility model, with all nodes evenly distributed in the simulation area. Nodal velocities are distributed according to a uniform distribution, with a minimum 1m/s and a maximum 30m/s, and a uniformly distributed pause time with mean 60 s, thus mimicking a moderate walking pace with infrequent stops. The AODV [2] routing protocol is used for unicast routing. Goodput is defined as the ratio of number of data packets received to time. Mostly, goodput is lesser than compared to throughput. Throughput calculates all the messages including control packets but goodput analyze only the data packets. Figure 4: Goodput Figure 4 shows TMP algorithm having little much lesser because of ad hoc network transfer the messages based on hop count, during this delay increases packet delivery rate was reduced. In TPP algorithm having lesser amount packets delivered compared to TMP algorithm, because it supports the access point it reduces the packet delivery rate. 56

6 Figure 5: Packet Delivery Ratio Figure 5 shows the analysis of packet delivery ratio between two algorithms. Packet delivery ratio is defined as the ratio of number of received packets to the number of transmitted packets. In this PDR, when the number of nodes increases packet delivery ratio decreases due to mobility. Compared to peer to peer overlay in mobile ad hoc network, peer to peer overlay in partial mesh network have packet delivery ratio as very less. Initially each algorithm having 100% packet delivery ratio when the node starts to increase the size it will reduce up to 20%. Figure 6: Packet Dropping Ratio Figure 6 shows that the packet drop occurs due to the traffic, congestion or network failure. Here transmitting number of messages at that time topology may change, due to that link failure occur it produce number of packet loss. From this analysis, TPP algorithm has more drop rate compared to TMP algorithm. Initially each algorithm having 40% packet drop rate, size of the network increases, it starts to increase the dropping level. Figure 7: End to End Delay Figure 7 shows the end-to-end delay is the time it takes a packet to travel across the network from source to destination. In peer to peer overlay in mobile ad hoc network algorithm increasing the end to end delay when 57

7 the number of nodes increases. But in peer to peer overlay in unstructured overlay reducing the end to end delay when the number of node increasing. In TMP algorithm having end to end delay as constant, when the number of nodes increases because of MANET used to transmit the packets using hop-hop analysis and also when the mobile nodes moves out of coverage area it uses the marking scheme. Partial mesh network have more delay, because all the nodes in this algorithm are in mobile state. Figure 8: Hop count Figure 8 shows the hop count refers to the number of path through which messages pass between sender and receiver. Each router along the data path constitutes a hop, as the data is moved from one Layer 3 network to another. Hop counts are often useful to find faults in a network, or to discover if routing is indeed correct. In peer to peer overlay in partial mesh network having more hop count needed to reach the destination because of cluster head. In peer to peer overlay in mobile ad hoc network (TMP) algorithm reduced the hop count needs due to MANET act as intermediate node between peer to peer overlay. 5 CONCLUSION These two algorithms provide an examination of overlay construction and thus resilient from the maintenance of intermediate link energy between the peer to peer network and mobile ad hoc network. Mostly peer to peer networks have co-operative nodes in mobile ad hoc network, but in this algorithm some of the nodes act as selfish peers to minimize the energy consumption and also reduce the distance to their destinations. From this identify the QOS parameter of this proposed algorithm. The performance of this algorithm is compared to benchmark algorithm. Peer to peer overlay in Mobile ad hoc network (TMP) algorithm using only less amount of hop count between the source and destination. Compared to Topology Maintenance in P2P overlay in MANET (TMP) algorithm, Topology maintenance in P2P overlay in Partial Mesh network (TPP) algorithm is more stable and feasible in throughput, packet delivery ratio and hop count. This approach uses the link based on minimum cost and high level energy. It is efficient topology maintenance for peer to peer overlay in partial mesh network. It may reduce the total cost network-wide. It reduce the energy consumption and to improve the response time and reliability. 58

8 6 REFERENCES [1] Afzal Mawji, Hossam S. Hassanein. Efficient Multipoint P2P File Sharing in MANETs. GLOBECOM'2009. pp.1~6. [2] Andersen D, Balakrishnan H, Kaashoek F, Morris R. Resilient overlay networks. In: Proceedings of the eighteenth ACM symposium on operating systems principles. ACM Press; P [3] Andrzejak A, Xu Z. Scalable efficient range queries for grid information services. In: Proceedings of second international conference on peer-to-peer computing. IEEE Computer Society; P [4] Chawathe Y, Ratnasamy S, Breslau L, Lanham N, Shenker S. Making Gnutella-like P2P systems scalable. In: Proceedings of the 2003 conference on applications technologies architectures and protocols for computer communications (SIGCOMM). ACM Press; P [5] Cheng A-H, Joung Y -J. Probabilistic file indexing and searching in unstructured peer-to-peer networks. Computer Networks 2006; 50(1): [6] Chung F R K, Graham R L. Erdosongraphs: his legacy of unsolved problems. A K Peters, Ltd.; [7] Ganesan P, Gummadi K, Garcia-Molina H. Canon in G major: designing DHTs with hierarchical structure. In: Proceedings of the twenty-fourth international conference on distributed computing systems. IEEE Computer Society; p [8] Garces-Erice L, Biersack E W, Ross K W, Felber P A, Urvoy Keller G. Hierarchical P2P systems. In: Proceedings of the ninth European conference on parallel processing science (Euro-Par). Lecture notes in computer science, vol Berlin: Springer; P [9] Joung Y-J, WangJ-C.Chord2: A two-layer chord for reducing maintenance overhead via heterogeneity. Computer Networks 2007; 51(3): [10] Kermarrec A-M, Massoulie L,GaneshA J. Probabilistic reliable dissemination in large-scale systems. IEEE Transactions on Parallel and Distributed Systems 2003; P: [11] Kostoulas D, Psaltoulis D, GuptaI, Birman K, Demers A. Decentralized schemes for size estimation in large and dynamic groups. In: Proceeding of the 4th IEEE international symposium on network computing and applications (IEEENCA), July 2005: p [12] Liben-Nowell D, Balakrishnan H, Karger D R. Analysis of the evolution of peer-to- peer systems. In: Proceedings of the twenty-first annual symposium on principles of distributed computing. ACMPress; P [13] Liben Nowell D, Balakrishnan H, Karger D R. Observations on the dynamic evolution of peer-to-peer networks. In: Proceedings of the first international workshop on peer-to-peer systems. Lecture notes in computer science, vol Berlin: Springer; P [14] Loo B T, Huebsch R, StoicaI, Hellerstein J M. The case for a hybrid P2P search infrastructure. In: Proceedings of the 3 rd international workshop on peer-to-peer systems, February 2004: p [15] Lu E J L, Huang Y F, Lu S C. M l chord: A multi-layered p2p resource sharing model. Journal of Network and Computer Applications 2009; 32(3): P:

9 About Authors Ms. J. Vijitha Ananthi received the Master s degree in communication System from Karunya University, India. Currently she is working as an Assistant Professor at Indus College of engineering, Coimbatore, India. Her interests are in overlay in wireless networks. She publishes more research articles in the journals. R. Haripriya is a PG Scholar of Indus College of Engineering, Coimbatore, India. Her research interests are in the field of Networking, VLSI, Signal Processing, Bio- Medical, and Image Processing. R. Nithya is a PG Scholar in Indus College of Engineering, Coimbatore, India. Her research interests are in the field of Networking, VLSI, Signal Processing, Bio-Medical, and Image Processing. R. Karthikeyan is a lecturer of Indus College of Engineering, Coimbatore, India. His research interests are in the field of Networking, VLSI, Signal Processing, Bio-Medical, and Image Processing. 60