Dissertation > Excellent graduate degree dissertation topics show

Research on P2P Resource Optimization and Several Security Issues

Author: LiuGuoLe
Tutor: XuGuoAi
School: Beijing University of Posts and Telecommunications
Course: Information Security
Keywords: complex networks P2P topology search algorithms trust mechanism traffic identification
CLC: TP393.02
Type: PhD thesis
Year: 2013
Downloads: 43
Quote: 0
Read: Download Dissertation

Abstract


Peer-to-Peer Network(P2P) is a logical overlay network built on the top of the Internet, and its characteristic is very significant, including network node decentralization, scalability, robustness, high cost performance, load balancing, privacy protection, freedom of user behavior and rich information content, etc. With the wide popularity of the Internet, the dramatic increment in network bandwidth and the rapid enhancement of computing capacity of Internet termi-nal systems, P2P network application mode has gradually risen and become the IT industrial and academic applications and research hot spot. P2P has been listed by Fortune Magazine as one of the four technologies which will affect the future of the Internet. At present, the application of P2P networks brings many conveniences while also brings some problems. P2P network traffic has accounted for more than70%Internet network congestion, and its nearly sym-metric network traffic model occupies the limited network bandwidth which leads to the network congestion. A more serious problem is that network attack by P2P application easily causes the Internet network paralysis. The research of P2P network is very broad, including network topology, network search, securi-ty and reliability, distributed data storage, massively parallel computing, appli-cation layer multicast and so on.With the in-depth study of P2P network, people have found that P2P net-work presents the features of complex network such as small world character-istics, scale-free characteristics and power-law distribution. These discoveries make complex network be the tools to study P2P network. Complex network has relations with many fields such as mathematics, physics, computer science, nonlinear science, biology and other subjects, and now it has become one of the current very popular international subjects. Complex network is mainly applied in these areas:empirical analysis and modeling, network structure and behavior mechanisms simulation, and network control mechanism.In this paper, the research content takes P2P networks as the main line, introduces the theory of complex network, and studies the topology structure, search algorithm, trust model, and content identification detection from a differ-ent perspective. The main innovation works of the thesis are briefly summarized as follows:(1) From the perspective of complex network, this paper focuses on the rela-tions between the underlying physical structure and the business of the ap-plication layer. Using complex network theory and introducing the concept of time-to-live(TTL) to the physical layer, this paper studies the influence of the physical network topology structure on the P2P network resource utilization of the application layer. It discusses the optimization of rating network topology based on the complete tree structure, analyzes the net-work performance qualitatively, and puts forward the cross-layer(C-L) scheme and the same-layer(S-L) scheme. Finally, take a five-layer rating local area network (LAN) as an example to give the experimental which verify the effectiveness of the proposed scheme and the previous theoreti-cal analysis.(2) Based on complex network theory, two search algorithms with memory function have been put forward e.g. memory function maximum degree algorithm (MD) and memory function preference degree algorithm (PD). Their performances in unstructured P2P networks (e.g. BA scale-free net-work, ER random graph and WS small-world networks) are tested. Experi-ments show that compared with BFS-based search algorithm, the proposed algorithm(MD) shows excellent performance in the search success rate of network resources, and at the same time reduces the search redundancy information. Compared with DFS-based search algorithm, the proposed algorithm(MD) can effectively reduce the number of search messages and can ensure network resources search success rate at the same time. (3) Introducing the interpersonal relationships of complex networks to the P2P network trust evaluation system, a trust and reputation management scheme-IRTrust is proposed. Firstly, in terms of trusted computing, start-ing from the interpersonal relationships of social network, the weight of network node with more familiar relationship should be greater, and the nearer experience should have the greater weight in the process of trust computing. Secondly, in terms of network nodes credibility, quality of ser-vice and quality of recommendation of network nodes are distinguished. For a specific resource sharing trade, network request nodes need to pro-vide the service quality evaluation of network service nodes, at the same time also provide the recommendation quality evaluation of network rec-ommendation nodes. Finally, in order to prevent and describe the unpre-dictable and uncertain malicious attacks behavior from network nodes, the network nodes need to consider risk factors including computing trust val-ue of network service node, context content, accumulation rate of trust and credibility before trading.(4) Combined with the advantages of deep packet inspection(DPI) and deep flow inspection(DFI) techniques, a hybrid recognition system-D2PFI2is proposed. After data flows perform DPI inspection identification, most of different types of P2P traffic congestions have been identified, and the unknown and encrypted P2P data flows are identified by DFI technique. The proposed scheme can not only ensure the accuracy of identification, but also can identify the pay load encryption data flow. In addition, in the DFI module, the modified C4.5machine learning algorithm is adopted, at the same time the detection results of DPI technique are taken as samples to input the DFI module training in order to improve the detection accuracy. Finally, the simulation experiments confirm that the proposed scheme can effectively synthesize the advantages of these two technologies to identify P2P traffic congestions.

Related Dissertations

  1. The Influence of Magnetic Field Topology on Electron Motion in Hall Thrusters,V439.2
  2. Topology Measurement and Security Analysis on Gnutella and eMule Network,TP393.08
  3. Task Partition of Network Simulation under Large-Scale Computing,TP393.01
  4. Flexible,rigid,Synthesis and characterization of mixed ligand complexes,O621.1
  5. Research on Peer-to-Peer Traffic Identification Algorithm Based on Cluster Analysis,TP393.02
  6. Feature based on complex network propagation characteristics of SNS social networking sites,G206
  7. Intelligent video detection technology based on the traffic lights control,TM923.5
  8. Multi-attribute undirected weighted graph clustering method,O157.5
  9. On Data Scheduling Strategies of Serving Peers in P2P VoD Systems,TN948.64
  10. Graph-based similarity search scientific literature key technology research,TP391.3
  11. Research and Implementation on JXTA-based Virtual Campus Community,TP393.09
  12. Traffic Classification Using Stream Data Mining Algorithm,TP393.06
  13. Research and Implementation of Application Traffic Classification & Restoring,TP393.08
  14. Research and Implementation of Replication Management Related Problems Based on P2P System,TP393.02
  15. The Structure Design and Toplogical Optimal Analysis for the Key Components of Tunnel Type Washer,TM925.33
  16. A Study of Multi-Path Routing Algorithm of Wireless Sensor Networks Based on Energy Balance,TP212.9
  17. The Design and Implementation of the Live Server in Audio and Video Broadcast Distance Education System,TP311.52
  18. Syntheses, Structures and Properties of Metal-organic Coordination Polymers Based on Long-Chain Aromaticmulticarboxylate,O631.3
  19. Assembly, Structures and Properties of Metal-Organic Frameworks Based on Flexible Aromatic Carboxylic Acid,O621.13
  20. Nontrivial Solutions of Nonlinear Singular n-order m-point Boundary Value Problems,O175.8
  21. Syntheses, Structures, and Property Studies of Coordination Polymers Assembled from 2, 4-Dibenzoylisophthalic Acid,O634

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Computer network > General issues > Computer network architecture and design
© 2012 www.DissertationTopic.Net  Mobile