an enhanced and secured qos for p2p networks.pptx

Describe about the enhanced and secured quality of services for peer to peer network.

1. AN ENHANCED AND SECURE QUALITY OF SERVICES FOR PEER TO PEER OVERLAY NETWORKS SUPERVISED BY Dr.S.N.Sivanandam Ph.D Professor Emeritus Dept of Computer Science and Engineering Karpagam College of Engineering Coimbatore. PRESENTED BY P.K.Sasikumar Research Scholar Anna University JOINT SUPERVISOR Dr. S. SMYS Professor Department of Electronics and Communication Engineering RVS Technical Campus, Coimbatore .

2. Agenda Introduction Related research Research Objective Proposed Approaches A Dynamic Performance of Multi-Tier Clustered Architecture (MCA). A Signaling System for Quality of Service - aware Content Distribution for P2P Overlay Networks. Secure Grouping Management Architecture (SGMA) for Peer To Peer Overlay Networks. Trust Model and Key Management scheme for secure communications in P2P overlay networks. Conclusion Future Work References List of Publications. 2

3. INTRODUCTION Overlay Network An overlay network is a network, which is built on top of an existing network and is always depending on the underlying network for routing and forwarding. The nodes in the overlay can be connected by virtual or logical links in which it corresponds to a path in the underlying network. Peer to Peer (P2P) network is said to be an overlay network when it runs on top of the internet. 3

4. INTRODUCTION Overlay Network Advantages of Overlay Networks Incremental deployment Overlay network does not require any changes in the existing routers in the network . Adaptable Metrics such as latency, bandwidth and security will be utilized to make the routing and forwarding decisions . Robust Even if a path fails in the overlay network it may be redirected to the same destination by means of multiple independent paths which is already created by the network by increasing the nodes. 4

5. INTRODUCTION Overlay Network Overlay is classified into structured and unstructured overlay networks. 5 Structured Overlay Networks Unstructured Overlay Networks

6. P2P NETWORKS P2P is an overlay network and also a decentralized model, which means there is no central coordination or control point to access other peer resources such as bandwidth, processing power and disk storage . Nodes in a P2P network, called peers, play a variety of roles in their interaction with other peers. When accessing information, they become clients and while serving information to other peers, they are servers. When forwarding information for other peers, they are routers. 6

7. RELATED RESEARCH 7 Author Algorithm/ Technique Description Improved Parameters Ferreira et al. (2016) SopCast (Streaming Over P2P cast) always forces the protocol to find the new connections between the peers to establish continuous streaming of video Size of the diameter gets reduced and having a shortest path between peers. Zhou et al. (2016) Data Prefetching Scheme based on Data Relevancy, (DPSDR) The super node will analyze the history access records for the nodes present in its cluster and then compute the interest relevancy between data objects Minimize the delay and higher cache hit ratio. Meng et al.(2016) Preferences-based trust model( preferTrust ) The receiver has to calculate its expectation vector and each responders trust vector in order to choose a trusted one Improve a peer’s trust decision satisfaction and resist malicious attacks. Freitas (2016) Twister : microblogging network architecture a new peer-to-peer microblogging platform with security, scalability, usability and privacy features. Improved security

8. 8 Author Algorithm/ Technique Description Improved Parameters Ayyasamy & Sivanandam (2009) Intelligent Replica Algorithm(IRPA) The requested contents were brought to closer proximity of the requested node which in turn reduces the access latency and traffic in the network Reduces the search latency ,Access latency and traffic. Zhou (2011) Peer Rank and Selection Algorithm(Searching),HDHTR+SDHTR(routing) The super peer stores the location and routing path and convey the information to the source peer group Reduced search ing delay and reduced traffic. Amble et al. (2011) Content-Aware Caching and Traffic Management The main constraints identified are finite cache sizes and the periodicities with which content is refreshed in the caches optimal throughput and yields minimum queues Sedaghat et al. (2016) topology aware resource management uses the scheduler to maintain the topology fixed channel allocation for the dynamic mobile nodes

9. 9 Author Algorithm/ Technique Description Improved Parameters Stefan Kraxberger (2011) Scalable Secure Routing Security properties of protocol can be adjusted through the selection of different security levels Protection from attackers Azharuddin & Annapurna (2012) a novel secure key issuing scheme to provide peer registration and issuing private key service to peers securely using secure key issuing protocol without the requirement of secure channels. Key issuing may be improved but KGC(Key Generate Center) failure leads to high fault tolerant. Cheng et al. (2012) METrust to reduce dishonest evaluation and strategic attack by considering the factors including mutual evaluation, similarity risk, time window, incentive, and punishment mechanism. improved initiative of peers, encourage good behavior

10. Summary The study on QoS for P2P overlay networks reveals the following facts:- Each system has its own advantages over other even though there exists a possibility of increasing the QoS parameters by introducing new algorithms. Whenever Security parameters are implemented it may affect the systems overall performance. To overcome such issues, MCA, SSQACD, SGMA and TMKM have been proposed and are analyzed. 10

11. Aims to reduce the issues occurring due to the link failure, node failure and resilience in P2P overlay Network. Aims to improve the query processing speed and content awareness of every node involved in the transmission in P2P overlay networks. 11 Research Objective

12. Research Objective The main objective of Secure Grouping Management Architecture (SGMA) is to allocate the resources securely and verifying whether the data is malicious attacked or not, before every transmission. Develop a trust model and efficient key management scheme to establish a secure communication in P2P overlay networks and to eliminate the malicious node joining the network during routing or group communication. 12

13. A DYNAMIC PERFORMANCE OF MULTI-TIER CLUSTERED ARCHITECTURE (MCA) Objective A Multi-tier Clustered Architecture (MCA) has been created for overlay networks in which the proposed scheme will create a structured P2P networks with different types of network. A cluster head plays a vital role in maintaining the topology and increase the Quality of Service(QoS). 13

14. METHODOLOGY The Topology Aware Routing (TAR) and Hybrid Push-Pull Topology (HP2T) are two existing techniques which have high robustness and high routing overhead while maintaining the topology in P2P network. Tran et al. (2016) suggested a scheme to increase the QoS by combining the existing topologies TAR and HP2T to generate a hybrid push pull P2P network. While combining the topologies the existing technique will have less maintenance in topology which leads to low delivery rate. To avoid such issues in the existing techniques, a new scheme known as Multi-tier Clustered Architecture (MCA) has been proposed for P2P network. 14

15. PROPOSED METHOD The proposed scheme has three steps to generate a successful architecture as follows- Creation of P2P network Cluster head selection Connecting the network to form the overlay structure After the P2P network has been created, each individual network will have to select its own cluster head based on the individual performance of nodes in the network. By connecting all the cluster head, overlay network will be formed. 15

16. In the proposed work the MCA uses the three different structured P2P networks with three different topologies such as tree topology, mesh topology and star topology . Tree topology, Mesh topology, Star Topology, The overlay network as, The proposed system uses three different topologies and hence, the value of is mention as 3   16

17. Algorithm //3-tier network architecture// If Select rand ( topo ) // select the three random topology Case-1 tree  equation 3.1 // create network by using equation 3.1 Case-2 mesh  equation 3.2// create network by using equation3.2 Case-3 star  equation 3.3// create network by using equation3.3 End case End if // Case-1: Tree topology P2P network architecture// // // parent node as Cluster head End //Case-2: Mesh topology P2P network architecture// // Node carrier high number of neighbour nodes –cluster head End //Case 3: Star Topology P2P network architecture// // –centralized node, – Neighbour nodes // centralized node as Cluster head End // Overlay Formation// // connect all cluster heads Intimate to all nodes End if End   17

18. Experimental Results Simulation Setup Figure 1 Multi-tier Clustered Architecture 18

19. Figure 1 shows the architecture of MCA which modeled the network with the help of access and overlay links. Each peer is connected by its own groups topology structure and all the clusters are logically connected via its cluster heads. For test the network, the ‘ NS-2 ’ simulator is used. ‘ NS-2 ’ is a general-purpose simulation tool that provides discrete event simulation of user defined networks. 19

20. The numbers of nodes are varying from 100 to 500 nodes and the physical medium is wireless medium. The constant bit rate of packet size is 1000 bytes and the interval time between the each packet is 0.01 seconds. The queuing size is 50 cm and the queuing time limit is 50 seconds. The simulation ends within 200 seconds. The proposed system performance is analysed using packet delivery ratio, packet drop ratio and delay. 20

21. Y=Packet Delivery ratio X=(Network Size) TAR HP2T MCA=3 MCA=6 MCA=9 100 0.50099 0.59901 0.7 0.79802 0.979208 200 0.450495 0.548515 0.649505 0.848515 0.961386 300 0.4 0.50099 0.59604 0.89604 0.952475 400 0.50099 0.548515 0.7 0.750495 0.79802 500 0.447525 0.4 0.750495 0.79802 0.845545 The existing Topology Aware Routing (TAR) and Hybrid Push-Pull Topology (HP2T) have reduced delivery rate due to the less topology maintenance. By varying the topology size, the proposed network performance(MCA) is analyzed. When increasing the number of P2P networks with different topologies , the packet delivery ratio of proposed system also increases . 21 SIMULATION RESULTS Figure 2 Packet Delivery Ratio vs Network Size

22. Packet Drop ratio X=(Network Size) TAR HP2T MCA=3 MCA=6 MCA=9 100 0.849758 0.749302 0.403246 0.353023 0.3 200 0.823956 0.673258 0.424886 0.369072 0.327212 300 0.800935 0.597204 0.449307 0.390693 0.351633 400 0.895114 0.847682 0.546286 0.420695 0.40117 500 0.897209 0.799535 0.598605 0.49814 0.445116 22 Figure 3 shows the packet drop rate against the network size and it is evident that the ratio of proposed system is better than that of the existing technologies. Figure 3 Packet Drop Ratio vs Network Size

23. Y=Delay (ms) X=(Network Size) TAR HP2T MCA=3 MCA=6 MCA=9 100 0.5 0.400021 0.3 0.25 0.2 200 0.548611 0.445833 0.306944 0.254167 0.201389 300 0.597222 0.497222 0.313889 0.255556 0.205556 400 0.645833 0.595833 0.345833 0.295833 0.245833 500 0.694465 0.594465 0.383354 0.31391 0.202799 23 Figure 4 indicates that even when increase in network size the delay received by the proposed MCA is less compared to that of the existing systems such as Topology Aware Routing (TAR) and Hybrid Push-Pull Topology (HP2T) method. Figure 4 Delay Vs Network Size

24. SUMMARY The proposed MCA is used to maintain the topology in the overlay network. With the help of cluster head selected in each topology in MCA will increases the network performance by avoiding the resilience and link failure. Simulation results have shown that the performance of the proposed multi-tier clustered architecture increases the network performance compared to the existing Topology Aware Routing (TAR) and Hybrid Push-Pull Topology [HP2T] techniques. 24

25. A SIGNALING SYSTEM FOR QUALITY OF SERVICE - AWARE CONTENT DISTRIBUTION FOR P2P OVERLAY NETWORKS Objective To develop a signaling system for content awareness in which every node involved in the transmission should be aware of its own content and processing speed. In addition, P2P node has an ability to choose its path for transmission. 25

26. FACTORS AFFECTING QUALITY OF SERVICE IN P2P NETWORK Latency Bandwidth Reliability Buffer cache ratio Available capacity CPU speed Memory size 26

27. METHODOLOGY The data flow in the physical layer can be improved by proposing content aware caching and maintaining the traffic inside the network. They consider that the node interested in data transmission should have the awareness of the network. If a node has a multiple path to transmit the data, it will cause loss of time and process cycle. To overcome this issue, a simple signaling system is proposed for content awareness in which every node involved in the transmission should be aware of its own content and processing speed. 27

28. A fixed level of maximum processing speed, bandwidth, catching size is calculated for all the nodes present in a network before the deployment of the network elements. Calculation of Quality Factors To calculate the quality factor, some of the following factors have to be considered. They are Bandwidth Factor . CPU Factor Buffering to Cache Ratio 28

29. Bandwidth Factor(BF) CPU Factor(CF) Buffer to Cache Ratio(BC) where ( The occupied buffer size for caching is then a product of the quantity ( nb ) and the average length of cached descriptions ( lb ). Suppose constant s is the unit size of 1 second descriptor, Csize be the maximum memory size present at a node )   29

30. Quality Factor where are weight factors which are used as normalization constants whose value lies between 0 and 1 . The weights can be adjusted as per user requirements depending on the three metrics stated above. Determining of Quality Factor Signal   30 Quality Factor range Assigning quality factor signal 0.81-1.00 00 0.56-81 01 0.31-0.55 10 0.0-0.30 11

31. The quality of factors can be divided into four levels. For the signaling purpose, two bits signaling systems are selected. 00 denotes the most quality of service is present. 01 denotes a good quality of service is present and data can be easily transmitted with a little delay. 10 denote the quality of service is lesser, were data can be transmitted with a mere delay. 11 show the least quality of service is present, which implies possibility of transmitting data is almost impossible. 31

32. Decision Table Present at Nodes The above table contains three columns. The first contain node id fields which keeps the whole set of node id used to send the acknowledgement packets. The second column of the table contains the corresponding quality factor signal Third column gives the remark to the chosen node. 32 Node <IDs> Quality factor signal Chosen node 548 00 Yes …………… ………………… ……………. …………….. ………………… ………………

33. Procedure of finding the right node and the right quality of service connection Calculate Quality Factor as discussed previously. The node has to determine the quality factor signal. At first, an interested node generates a requesting data packet. In the request packet, one field is reserved for the intermediate groups. This field contains a stack. The stack is used for the path determination. Every group has a group leader. As group leaders have all the node’s ids with them, the interested node searches the data in its group. If the data is not found among the own group, then the request data packet is sent to other groups. When the packet is received by any group, which has not generated data packet; it first checks the flag value. If the flag value is one, it will check the query presents in it. If it is available, then it will check the quality present. If the quality is in the range of the defined range then it sends the acknowledgement packet. The node sends the data packet keeping the flag zero. The acknowledge data packet also contains the QoS level. At the time of sending the acknowledgement packet, the node copied the intermediate nodes <ids> to the intermediate field of the data packet keeping the order same . When a node gets the data packet and does not find the request data in it, then it pushes the id to the stack of the data packet and sends to other nodes whose group id is not on the top. 33

34. When a node receives a data packet and the received data packet’s flag is zero, then it first checks the destination. If the destination id matches with its own group id, the data packet is transmitted to the specific node. The interested node checks the QoS level. It maintains the QoS levels got from different nodes in a table. Then it finds the best way of getting the data and sends a contract to the node. The sender node kept the flag 1 at the time of transmission of contract. If the node finds the data packet’s flag field is zero and the destination field does not match with its own id; then it just pops the stack and sends the data packet to the top value group of the stack. After getting the contract, the node with the data makes a path sends the required data to the requested node. At this time, the flag is zero. The data may be of a single unit or of a stream of data. 34

35. Experimental Results Simulation Setup 35 Figure 5 Topology of P2P Network The system uses the Bit Torrent packet-level simulator for overlay networks . The system uses a simplified topology in the simulations as in the Figure 5. The proposed Signaling System-QoS Aware Content Distribution (SSQACD) architectures and existing Flow level Network Bandwidth Simulator (FLNBS)performance is evaluated in terms of bandwidth utilization, processing delay, signaling overhead and packet delivery ratio. The packet transmission rate is varied as 1, 1.5, 2, 2.5 and 3Mb.

36. Simulation Results 36 X=Rate(Mb) FLNBS Mb/s SSQACD Mb/s 1.0 0.661017 0.988701 1.5 0.977401 1.37288 2.0 1.27119 1.77966 2.5 1.68927 2.25989 3.0 2.23729 2.74011 Figure 6 Rate Vs Bandwidth Utilization Figure 6 depicts the results of bandwidth utilization of SSQACD which is compared with FLNBS. From the graph it is inferred that the SSQACD outperforms the FLNBS by 27 % in bandwidth utilization.

37. X=Rate (Mb) FLNBS Delay (sec) SSQACD Delay (sec) 1.0 0.079944 0.051977 1.5 0.1 0.054802 2.0 0.120904 0.092373 2.5 0.137006 0.096328 3.0 0.139266 0.099153 37 Figure 7 Rate Vs Delay The figure 7 indicates that the delay of SSQACD was reduced by 31 % compared to the delay of FLNBS. Simulation Results

38. SIMULATION RESULTS 38 X=Rate (Mb) FLNBS Overhead (Pkts) SSQACD Overhead (Pkts) 1 437 437 1.5 4444 2113 2 8561 4153 2.5 12678 8160 3 16721 12095 Figure 8 Rate Vs Signaling Overhead Figure 8 inferred that the SSQACD outperforms FLNBS by 52% in signaling overhead .

39. Simulation Results 39 Rate (Mb) FLNBS Delivery Ratio SSQACD Delivery Ratio 1.0 0.417094 0.593162 1.5 0.417094 0.745299 2.0 0.538462 0.835897 2.5 0.57094 0.863248 3.0 0.676923 0.91453 Figure 9 Rate Vs Packet Delivery Ratio From Figure 9 it is incident that the SSQACD performs better than FLNBS as it gains 34 % in delivery ratio.

40. SUMMARY Thus the proposed Signaling System-QoS Aware Content Distribution(SSQACD) outperforms in bandwidth utilization, processing delay, signal overhead ratio as compared to Flow Level Network Bandwidth Simulator(FLNBS). 40

41. Secure Grouping management architecture (SGMA) for peer to peer overlay networks Objective To formulate a Secure Grouping Management Architecture (SGMA) which allocates the resources securely as well as it avoids the unauthorized access of the attackers. 41

42. PROPOSED METHOD The main objective of Secure Grouping Management Architecture (SGMA) is used to allocate the resources without being attacked by attackers. It assigns each node in different group with respective Group ID (GID ). Group Head(GH) can be elected for all the groups. The responsibility of the group head is to allocate the resources and to verify the transmitted data being maliciously attacked before transmission . Entity level security is being installed for end point systems. Malicious node is detected by Group Head (GH) in each topology by using the entity value and the control message including entity level. If the entity level is one, then the transmitted data is not malicious. If it has more than one entity, then transmitted data and its respective node are infiltrated by unauthorized access. Further, the information about the malicious attack will be intimated to the remaining group head in the network. 42

43. Secure Grouping Management Architecture 43 Figure 10 Secure Grouping Management Architecture The nodes in the network are labeled as A, B, C, D, E, F, G, H, I, J, K, L, M and N and GH denote group head. The nodes are further sub divided into three equal columns Group 1, Group 2 and Group 3 and GH is elected based on the comfort of individual group.

44. Group Head Election The group members list the possible number of neighbour nodes in the nearest group. The node which is elected as group head has lesser distance level to its adjacent group header nodes. Figure 10 shows that the D and K has distance level less than that of the other nodes in the group . Hence they are elected as Group Head . 44

45. Algorithm Elect GH GH Assign GID {Gm1, Gm2, … Gmn } // Gm  Group Members GH Send REQ{ Gmi } // i =1,2..n Gmi Ack GH GH check Entity If entity ≤ 1 No intruders affected Else Intruders affected GH discards Gmi 45

46. Experimental Results Simulation Setup Simulation results are analyzed the performance of the proposed architecture in three different stages by varying the group size . If the group size is five that is denoted as Gs =5, if the group size is 10 that is denoted as Gs =10 and the group size is 20 that is denoted as Gs =20 and so on . Proposed work is compared with the existing approach Adaptive Trusted Authorization and Request (ATAR) model and Resource Sharing in P2P network (RSPP). Possible group members 46 Number of nodes Possible Number of Group Members in each group Gs=5 Gs=10 Gs=20 100 20 to 30 10 to 15 5 to 8 200 40 to 50 20 to 25 10 to 12 300 60 to 70 30 to 35 15 to 17 400 70 to 80 40 to 45 20 to 22 500 80 to 90 50 to 55 25 to 27

47. Simulation Results X=(Network Size) Y=Overhead Ratio ATAR RSPP SGMA Gs=5 SGMA Gs=10 SGMA Gs=20 50 0.700483 0.605797 0.396135 0.301449 0.2 150 0.828986 0.801932 0.419807 0.314976 0.348792 250 0.747826 0.700483 0.497585 0.328502 0.304831 350 0.849275 0.825604 0.301449 0.396135 0.348792 450 0.883092 0.852657 0.48744 0.328502 0.22029 47 Figure 11 Overhead Ratio Analysis Figure 11 indicates that the existing ATAR and RSPP has high overhead ratio compared to SGMA since both the existing system uses more control packets.

48. X=(Network Size) Y=Delay (ms) ATAR RSPP SGMA Gs=5 SGMA Gs=10 SGMA Gs=20 50 0.5 0.403158 0.201053 0.154737 0.108421 150 0.601053 0.449474 0.222105 0.205263 0.129474 250 0.693684 0.596842 0.238947 0.302105 0.205263 350 0.798947 0.748421 0.302105 0.352632 0.251579 450 0.874737 0.798947 0.487368 0.335789 0.205263 48 Figure 12 Delay Analysis Figure 12 Shows that the delay of Proposed SGMA is higly reduced as compared with ATAR and RSPP

49. X=(Network Size) Y=Packet Drop Ratio ATAR RSPP SGMA Gs =5 SGMA Gs =10 SGMA Gs =20 50 0.448908 0.401747 0.299563 0.255022 0.20262 150 0.448908 0.401747 0.320524 0.289083 0.252402 250 0.50131 0.454148 0.346725 0.317904 0.281223 350 0.600873 0.553712 0.388646 0.351965 0.302183 450 0.8 0.700437 0.448908 0.401747 0.323144 49 Figure 13 Packet Drop Ratio Vs Network Size Figure 13 infers that the Packet Drop Ratio of proposed SGMA is reduced than that of existing systems.

50. SUMMARY Secure Grouping Management Architecture (SGMA) for P2P overlay networks avoids the intruders and to share the resources to the peers more effectively. The simulation results reveals a fact that if the group size increases, the network performance also increases . The proposed Secure Grouping Management Architecture (SGMA) reduces the delay, drop rate and overhead which are an evident from the simulation results. 50

51. TRUST MODEL AND KEY MANAGEMENT SCHEME FOR SECURE COMMUNICATIONS IN PEER TO PEER OVERLAY NETWORKS Objective To incorporate a Trust Model and efficient Key Management scheme to establish a secure communication in P2P overlay networks . In order to establish a secure peer group, each node is first registered by sending request to Trusted Third Party (TTP). After peer’s registration, TTP issues a shared secret key by verifying the proof of registration. To avoid escrow problem, Key privacy can be achieved by using Key Privacy Authorities (KPA) during key issuing phase. To eliminate the malicious node joining the network during routing or group communication, an improved secure trust model is developed. 51

52. METHODOLOGY AND PROPOSED SCHEME Stefan Kraxberger (2011) proposed a scalable secure routing protocol focuses mainly on providing only scalability for heterogeneous unstructured P2P networks according to the users security needs. It does not maintain any trust relationship between peers. When issuing shared secret key to peer through channel using trusted third party can turn into adversary node that leads to escrow problem. Peer does not submit any proof for registration to join with the network. So it may be possible for replay attack by the malicious node while joining in the group. 52

53. 53 Figure 14 Block Diagram of Trust Model and Key Management Designing of Secure Model Peer Registration to avoid Replay Attack Secure Key Issuing to registered peer Selection of Secret Trust model to provide Scalability during Routing Trust model for secure routing by eliminating malicious node

54. Figure 14 illustrates the block diagram of the proposed scheme TMKM. The proposed secure model is based on four phases- Peer registration All the peers need to join the group have to register in order to avoid the replay attack. secure key issuing Secure key issuing scheme is used to issue private key to all the registered peers in the group secure trust model The trust model will estimate the trust value for each peer in the group and the estimated value is used to eliminate the malicious node either at the time of joining group or at the time of routing . selection of security levels To enhance the scalability of the network during routing, the security properties are adjusted by the selection of different security levels like low, medium and high. 54

55. Peer Registration Before the node joins the group, a peer P must send the request to TTP A simple protocol is described as below for peer registration : A peer generates a request with a random nonce and sends it to TTP . Once TTP receives the request, it issues (IDP, ProofP ) for peer P . The proof of registration ProofP is a message that can confirm whether the peer has been registered or not . Assigning ID by TTP can prevent a peer from selecting its own ID and lessen the Sybil attack in the considered network . The communication between TTP and peer P might be secured by using Shamir’s threshold secret sharing scheme . 55

56. Secret Key Issuing Scheme 56 Figure 15 Implementation of Secret Key Issuing Scheme

57. Request : Peer transmits a request with its proof of registration along with a nonce to get the partial private key. TTP Response : On receiving peer ’s request, checks the proof to verify whether is registered or not, in case the result is positive, responses with a partial private key otherwise it discarded the request. Blind KPA Request : When peer receives the partial private key from , it selects some KPAs and requests them in parallel to give key privacy service by transmitting a request. KPA Response : Each KPA validates the peer and issues a partial private key to it. Key Retrieval : On receiving at least private keys from pieces of KPAs, peer combines them and unbinds to produce the private key.   57

58. SECURE TRUST ROUTING MODEL Each peer is associated with a reputation value. When receiving the request, the pre-neighbour peers verify the trust value of the requester. If the trust value is higher than a threshold, both peers record each other in their neighbour list and the same is established. Hence, peers with higher trust values are connected, which enhances the quality of service, detaches malicious peers and resists collusion efficiently. 58

59. Scalable Secure Routing Protocol In order to provide the scalability, the security properties of protocol are adjusted by selecting the different security levels. Security Levels Admission Control Data Security Secure Routing using Trust Model 59

60. Security Aspects 60 Computing recommendation value and removing malicious node Partial private keys 2 Security levels 1 Security Aspects Peer Registration Secure Private Key Assignment of ID Refreshing Trust Value None None None Admission Control Data Security Secure Trust Model Figure 16 Represent Security Level in Peer Group Concept

61. THE ALGORITHM FOR TRUST MODEL AND KEY MANAGEMENT 61

62. 62

63. Experimental Results Simulation Setup 63 Figure 17 Topology of P2P Overlay Network NS-2 BitTorrent packet-level simulator for P2P networks is used . Each peer is connected with an asymmetric link to its access router . All access routers are connected directly to each other modelling only on an overlay link This enables to simulate different upload and download capacities as well as different end-to-end (e2e) delays between different peers . Figure 6.5 shows the topology consideration for the proposed model Trust Model and Key Management(TMKM).

64. Simulation Results The proposed TMKM scheme is compared with the existing Reputation Exchange Protocol (REP) . Varying the Attackers The number of malicious nodes performing dishonest activities is increased from 2 to 10 among 100 transactions. From figure 18, it is found that the proposed TMKM outperforms REP by obtaining 7% higher delivery ratio. 64 Number of Attackers REP Delivery Ratio TMKM Delivery Ratio 1 0.898058 0.944013 3 0.866019 0.912621 5 0.848544 0.909709 7 0.813592 0.894175 9 0.81068 0.881553 Figure 18 Packet Delivery Ratio Vs Varying Attackers

65. 65 Number of Attackers REP Packets Dropped TMKM Packets Dropped 1 468 288 3 507 344 5 843 488 7 1175 688 9 1319 910 From the figure 19, it can be seen that TMKM has 36% lesser packet drop when compared to REP. Figure 19 Packet Drop by Varying Attackers

66. Number of Attackers REP Misdetection Ratio TMKM Misdetection Ratio 1 0.504505 0.450551 3 0.515015 0.475776 5 0.545145 0.517117 7 0.571772 0.523423 9 0.596997 0.533233 66 Figure 20 Missdetection Ratio by Varying Attackers From the figure 6.8, it can be noted that the proposed TMKM has 8% reduction in misdetection ratio when compared with the existing scheme REP.

67. Varying Transaction 67 Number of Transactions REP Packet Delivery Ratio TMKM Packet Delivery Ratio 50 0.897314 0.943803 150 0.890858 0.943713 250 0.883438 0.93978 350 0.881307 0.933205 Figure 21 Packet Delivery Ratio by Varying Transactions From the figure 21, it is evident that the proposed TMKM has 29% lesser packet drop when compared to the existing REP scheme. Number of Transactions REP Packets Dropped TMKM Packets Dropped 50 464 291 150 492 324 250 492 367 350 528 393 Figure 22 Packet Drop by Varying Transactions From the figure 21, it is inferred that the proposed TMKM outperforms REP by obtaining 6% higher delivery ratio.

68. Number of Transactions REP Missdetection Ratio TMKM Missdetection Ratio 50 0.498848 0.450343 150 0.515253 0.46703 250 0.519071 0.469293 350 0.522606 0.470566 68 Figure 23 Missdetection Ratio for Varying Transactions From the figure 22, it can be seen that TMKM has 9.6% reduced misdetection ratio when compared to REP . SUMMARY In order to establish a secure communication in P2P overlay network a trust model and efficient key management scheme has been proposed . From the Simulation results it is evident that the proposed TMKM have increased packet delivery and reduced packet drop and misdetection ratio when compared with REP.

69. CONCLUSION AND FUTURE WORK Thus the QoS and Security issues in Overlay networks (WSN) has been analyzed with the specified set algorithms and analysis methods. The proposed architecture such as MCA, SSQACD, SGMA and TMKM and their algorithms will provide a better QoS and Secure communication in the P2P overlay network compared to the existing methodologies . Overall the proposed work is well suited for all types of P2P overlay networks. 69

70. SCOPE FOR FUTURE WORK Construction of overlay and optimal routing is still a difficult task in overlay networks. Swarm intelligent algorithms may be used for better optimization in various applications which may be implemented in P2P networks ( Dweepna & Parth 2012 ). Efficient load balancing algorithms were implemented to decrease the delay across the network (Takeda et al. 2015 ). An effective fault tolerant mechanism can be implemented to reduce the fault occurrences in the network ( Amudhavel et al. 2015). The next task is query propagation in P2P networks ( Nicolini et al. 2015). In a P2P networks the node has to choose whether it has to answer the query received or forward it to its neighbouring node. In such a case, an effective query processing methodology can be implemented for better optimal results. 70

71. REFERENCES Alekseev, S & Schäfer , J 2016, A New Algorithm for Construction of a P2P Multicast Hybrid Overlay Tree Based on Topological Distances : In The Seventh International Conference on Networks & Communications Androutsellis-Theotokis , S & Spinellis , D 2004, 'A survey of peer-to-peer content distribution technologies', ACM computing surveys (CSUR), vol. 36, no. 4, pp. 335-371 . Ayyasamy , S & Sivanandam , SN 2009, 'A QoS-Aware Intelligent Replica Management Architecture for Content Distribution in Peer-to-Peer Overlay Networks', International Journal on Computer Science and Engineering, vol. 1, no. 2, pp. 71-77 . Buyukkaya , Eliya , Maha , Abdallah & Gwendal Simon 2015, 'A survey of peer-to-peer overlay approaches for networked virtual environments', Peer-to-peer networking and applications 8, vol. II, pp. 276-300 . Cheng, Xiao-Long Xu & Bing-Zhen Gao 2012, ' METrust : A mutual evaluation-based trust model for P2P networks', International Journal of Automation and Computing, vol. 9, no. 1, pp. 63-71 . Chithra Selvaraj & Sheila Anand 2011, 'A Role Based Trust Model for Peer to Peer Systems Using Credential Trees', International Journal of Computer Theory and Engineering, vol. 3, no. 2 . Conrad, Yi, Nemanja , John & Scott. 2016. Wireless Key Management for Authentication. 20,160,036,594 . Cui, Haitao , Lv , Chao, Guangyong & Caixue 2015, Efficient key management for IOT owner in the cloud : Big Data and Cloud Computing, pp. 56-61 . Han, Hong, Mengyin , Bin, Wu & Shengqiang , Li 2015, 'A topological potential weighted community-based recommendation trust model for P2P networks', Peer-to-Peer Networking and Applications, vol. 8,no. 6, pp. 1048-1058 . 71

72. Jason, A & Shanmugasundaram , R. 2011, 'Design of QoS Aware Dominating set based Semantic Overlay Network (QADSON) for Content Distribution', Journal of Computer Science, vol. 7, no. 10 , pp . 1478-1489 . Liu, Yao, Dehai , Weishan & Leyi 2016, 'Patching by automatically tending to hub nodes based on social trust', Computer Standards & Interfaces, vol. 44, pp. 94-101 . Mahambre , S , Kumar, S & Bellur , U 2007, 'A Taxonomy of QoSAware Adaptive Event- Dissemination Middleware', IEEE Internet Computing, vol. 11, no. 4, pp. 34-44 . Malatras , A 2015, 'State-of-the-art survey on P2P overlay networks in pervasive computing environments', Journal of Network and Computer Applications, vol. 55, pp. 1-23 . Meng , Tianjiao Li & Yu Deng 2016, ' preferTrust : An ordered preferences-based trust model in peer-to-peer networks', Journal of Systems and Software, vol. 113, pp. 309-323 . Moalla , S, Hamdi , S & Defude , B 2010, A new trust management model in P2P systems : In Signal-Image Technology and Internet-Based Systems (SITIS), pp. 241-246 . Moustakas , Hüseyin Akcan , Mema Roussopoulos & Alex Delis 2016, 'Alleviating the topology mismatch problem in distributed overlay networks: A survey', Journal of Systems and Software, vol. 113, pp . 216-245 . Tahta , Sevil & Ahmet Burak 2015, ' GenTrust : A genetic trust management model for peer-to-peer systems', Applied Soft Computing, vol. 34, pp. 693-704 . Takeda, A, Oide , T, Takahashi, A & Suganuma 2015, Efficient Dynamic Load Balancing for Structured P2P Network : Network-Based Information Systems ( NBiS )18th International Conference on. IEEE, pp . 432-437 . Tarkoma , S 2010, Overlay Networks: Toward Information Networking, CRC Press . Zarrin , J, Aguiar , RL & Barraca , JP 2014, A self-organizing and self-configuration algorithm for resource management in service-oriented systems : In 2014 IEEE Symposium on Computers and Communications (ISCC), pp. 1-7. 72

73. LIST OF PUBLICATIONS Sasikumar, PK, Ayyasamy , S & Sivanandam , SN 2016, "A Dynamic Performance of Multi-Tier Clustered Architecture for P2P Overlay Networks." Asian Journal of Information Technology, vol. 15,no.10 pp . 1547-1551. Sasikumar,PK,Ayyasamy,S,Smys,S&Sivanandam,SN2016, " An Analysis of Secure Grouping Management Architecture for P2P Overlay Networks." The International Journal of Advanced Engineering and Technology, vol. 3, no. 2, pp.1008-1011. Sasikumar, PK, Ayyasamy , S & Sivanandam , SN. "A signaling system for Quality of Service Aware content distribution for P2P overlay networks." The International Arab Journal of Information Technology, Accepted for publication. Sasikumar, PK, Ayyasamy , S, Smys , S & Sivanandam , SN 2016. "Trust Model and Key Management Scheme for Secure communications in Peer-to-Peer Overlay Networks" ( communicated). 73

74. 74 Thank you