Font Size: a A A

Research On Resource Location And Search Model In Content Delivery P2P Network Based On User Requirements

Posted on:2016-12-01Degree:DoctorType:Dissertation
Country:ChinaCandidate:M HeFull Text:PDF
GTID:1108330482457705Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Unstructured content delivery P2P networks established on the Internet are new network architectures, they have been widely used in resource sharing and service distribution. Therefore, resource location is the one of core problems in unstructured content delivery P2P networks. The basic functions such as topological structure, routed protocol and resource search and location are realized in P2P distributed networks. In the meantime, large and dynamic characteristics of P2P distributed network also increased the difficulty of resource location technology. In unstructured content delivery P2P networks locating area, how to provide the valuable resources to users and how to improve the user resources locating accuracy and transferring efficiency has become an urgent problem.The resource location in content delivery P2P network is divided into four types. The center locating model, unstructured content delivery P2P resources locating model, structured content delivery P2P resources locating model and the hybrid P2P resource locating model, the most popular one is the unstructured content delivery P2P resources locating model, this type model is used widely. In unstructured content delivery P2P distributed network, system is dependent on the active assistance of puerperal equipments; each user directly benefits from the other members rather than the server. Users in the system play the role of servers and clients, but they cannot realize each other. Users establish a virtual connection, so that they formed a virtual logical network on the application layer in the whole system.In the resource locating process of unstructured content delivery P2P network based on user requirements, user preferences are in difference. This difference is constructed on the advantages and disadvantages of network performance. Network service operators need to provide more reliable services according to the different requirements. These services include resource search, location, storage, management and dynamic adjustment. For the purpose of improving resource locating efficiency, we adptively choose the super peers in P2P distributed network. With load balancing strategy, we also improve the user query, search, and resource distribution effect.(1) In unstructured content delivery P2P networks, how to locate the resources of user required rapidly and precisely is a hot spot issue in currently, and is also one of core problems faced by P2P application fields. Related unstructured content delivery P2P resources location algorithms cannot be optimized at the same time in respect to precision ratio, recall ration and query cost, which can cause serious network bandwidth burden, and huge index maintenance costs. Therefore, this paper proposes a query strategy called User Requirements Resource Location Strategy (U2RLS). The innovation of this strategy is integrated into user requirements, user preferences and user interest, which based on the original unstructured content delivery P2P network resources location Flooding algorithm. Firstly, we subnet the user resources, and adopt the flooding mechanism and query index mechanism with the information of user required to locate the resources search partial coverage phenomenon, so as to solve the problem that query nodes use relay nodes blindly. The experimental results show that U2RLS has high search success rate, limited network resources consumption and short response time in query process, it also can significantly improve the efficiency for user resources location.(2) Resource Location strategy has become very popular for mass data distribution in complex and dynamic network. Searching for the objects is a fundamental problem to unstructured content delivery P2P networks. In this paper, we propose a resource location strategy for user requirements, which employs the collaborative exchange of information between peers to construct interest communities. This strategy can gather similar peers and disseminate useful information among them. Furthermore, we design an algorithm in unstructured content delivery P2P systems named Appropriate Degree Gossip Search Algorithm (ADGSA). Using this search algorithm, the performance including search success ratio, recall rate and search response time has been improved dramatically. An efficient resource location system with a fast organization of users cluster based on their requirements can be provided, preventing the creation of unlike communities. The simulation results show that our strategy has better search efficiency, short response time and high recall ratio.(3) Rare resource location is an unsolved problem in the area of unstructured content distributed P2P network. For the locating accuracy and efficiency in the rare resource environment, our paper proposes a novel Gossip-based appropriate degree search strategy by two-hop index replication for rare resource (GADSS-TIRS). This strategy employs a periodical probing mechanism (PPM) to judge whether these resources are rare resources. Once the resource types are determined, searching probability for rare ones is increased by two-hop index replication with resource weight (TIR-RW). In addition, we also provide a Gossip-based appropriate degree search algorithm (GADSS), that utilizes the appropriate degree property as heuristic information to guide the search in dynamic environment. Because the appropriate degree node contains more resources, the resources are located successfully according to the appropriate node search path. When the resources are located in the common nodes search path, query message arrivals terminal node and turns back, it selects the second appropriate degree node to search the resources. As compared to the conventional P2P resource search algorithms, GADSS has the ability to reduce the message quantity substatntially. The simulation results show that GADSS has the better search performance results show that GADSS has the better search performance for rare resources, and maintains the lower network overhead.(4) In network detection, routing strategy plays an important role in the mechanism of diversity. However, unstructured content delivery P2P networks turn to be fragile and vulnerable for the mass data distribution and complex dynamic environment. Location strategy is a crucial application to realize the resources search owing to the advantages of its characteristics. In the P2P models and systems, routing plays an important role to maintain this successful P2P technique. However, how to design a robust routing mechanism is still a challenge to unstructured content delivery P2P networks. To address this problem, we propose a novel routing periodical probing strategy (RPPS) to diagnose invalid links and peers, and design a novel routing failure restoration strategy (RFRS) which combines the RPPS to enhance location performance when mobile clients exploit P2P systems to locate resources they required. RPPS mechanism can increase search hit rate and accuracy during the location process, and also improve the probability of discovering the faulty links and peers in the unstructured content delivery P2P networks. As node failure occurs, many nodes receive the unreachable message. RFRS assigns other nodes as neighbors, and scans their routing table, it is able to reconstruct the new links to replace those failure ones even in the complex and dynamic network environment. If the failure type is a single link, RFRS builds monitor events, and finds out the best link to replace the failure one, if they are multiple link failures, these failure links are divided into two groups, which named incoming links set and outgoing links set. Then the incoming nodes set and outgoing nodes set formed, new links are established between these two sets according to the link weights, instead of the failure of mulitple links. Simulation results under realistic operating conditions substantiate that our strategy improve the robustness and stability in unstructured P2P network.
Keywords/Search Tags:Content delivery P2P network, Resource location, Rare resource, Flooding algorithm, Gossip algorithm, User interest community
PDF Full Text Request
Related items