Construction Of P2P Overlay Network In Mobile Environments: Key Technologies

Recently, with the advance in both of the mobile computing and wireless communication technologies, there comes such a phenomena that the computing power of mobile devices is becoming more and more powerful, meanwhile the communication capacity of computing devices is also getting stronger and stronger. Such phenomena then bring a bright prospect for introducing currently quite popular P2P (Peer-to-Peer) computing into the mobile computing environment. Faced with the opportunities, both the academic research and industrial development have shown their great enthusiasm and interests for constructing P2P overlay network in the mobile computing environment, and usually treated it as a new research area, namely MP2P (Mobile Peer-to-Peer).However, the special characteristics of mobile environments have also brought new challenges for research in mobile P2P computing. The challenges mainly lie in three aspects: MP2P architecture, MP2P resource search and retrieval and MP2P security. Furthermore, the challenges may differ with different mobile environments. This thesis firstly divides the mobile environments into two types, conventional infrastructure-based mobile network and mobile ad-hoc network (MANET), and then studies the MP2P architecture and MP2P protocol for each of them. This thesis also studies the security issue for MP2P, and focuses on the trust management for MP2P. The main contributions of the thesis can be summarized as follows:Firstly, this thesis studies the construction of MP2P overlay network in the MANET environment. The studies focus on the two key technologies, MP2P architecture and MP2P resource search and retrieval scheme. Here, we firstly propose a kind of MP2P architecture, Geographic-information-based Hierarchical Hash Index structure, abbreviated as GH2I, referring to the thinking of DHT (Distributed Hash Table) and considering the special characteristics of MANET environment. Based on this MP2P architecture, we then design a MP2P resource search and retrieval algorithm, called CAR (Content Addressable and auto-Retrievable). In this part of work, we managed to construct a scalable, efficient and mobility adaptive MP2P overlay network, which is evaluated and proved by analysis and simulations.Secondly, we study the construction of MP2P overlay network in the conventional environment, based on the application scene of maritime mobile communication network. By referring to the advanced wireless technologies such as WiFi, WiMAX, Inmarsat 3/4 and etc., we firstly propose a Next Generation Maritime Data Communication Network (NGMDCN) conceptual structure. Based on NGMDCN, we then propose a kind of hierarchical MP2P architecture. At last, based on the proposed architecture, we design a marine MP2P system and name it as eSeagull. In fact, this part of work involves another of our researches on the network throughput performance of Automatic Identification System (AIS). AIS is a kind of special Spatial Time Division Multiple Access (STDMA) Ad-hoc network.Finally, the MP2P security issue is discussed. The security issue for MP2P is relatively rare studied. However, the MP2P security issue is inevitably one of the key technologies of MP2P research. MP2P security issue may cover trust management, attack detection, access control, anonymous communication, reputaion and etc. Due to space and time limitations, we cannot do a comprehensive and deep research on all of these security issues, but propose a reputation based trust management model for MP2P. This trust model can can quickly detect the misbehavior peers and effectively limit the impacts of them in a MP2P file-sharing system.This model evaluates and ranks the MP2P peers' reputation according to their providing files' QoS (Quality of Service). Based on the reputation, we then design a kind of rating trust model and also propose the method of trust updation and aggregation. This model mainly aims at isolating the malicious peers and alleviating the malicious peers' impacts. The model also takes the impersonal factors caused by the network itself to considerations when deciding whether a peer is malicious, so that the decisions made could be more fair-and-square. Finally, the model's validity is proved by the theoretical analysis and simulations.