Font Size: a A A

Research On Safe And High Efficient Routing Algorithm Of Structured P2P Network

Posted on:2010-12-06Degree:DoctorType:Dissertation
Country:ChinaCandidate:C W LuFull Text:PDF
GTID:1118360302471127Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
Nodes in structured P2P networks must strictly follow a set of logical relations (such as DHT Table) to connect with the neighbor nodes. It will seriously threaten the security and efficiency of P2P routing if the logic relations were destroyed. So, security and efficiency of the routing algorithm in structured P2P networks becomes the focus of advanced research recently.Currently, the research of secure routing algorithm in structured P2P network mainly aimed at some sort of independent attacks, such as worms, botnets, distributed deny of service and identity attack. However, they can just passively respond to certain types of attack, while have no effect with other types of attack. These methods lack of active and broad-spectrum detection and defense mechanisms, and also less routing algorithms to consider the reliability and security overhead. On the other hand, the efficiency of the routing algorithm will sharply decline as the hot-spot resources are downloaded. This will seriously deteriorate the anti-attack performance of P2P network. Therefore, those researches can not really solve the security and efficiency of P2P routing.RAPD (Routing Algorithm based on Positive Detection), a broad-spectrum detection of security routing algorithm is proposed. P2P routing attacks have a few common characteristics: using malicious nodes (including the nodes hijacked by virus) to launch attacks; the main factor affecting attack effect is the number of malicious nodes; there exist three main methods of routing attacks: No forward routing message, wrongly transmit routing message and tamper routing message. RAPD algorithm periodically sends detection packets along each routing path. Detection packet is made up of a random positive integer X which record the number of hops and several sets encrypted in each layer. The relay nodes can use their private key to unlock the corresponding layer to deal with their own information. The destination node removes the encrypted X series in the detection packets, sign and pass back along the original routing path. In the process, if there is a malicious node to launch attacks , RAPD algorithm can check them out by resending detection packets, matching the key, or checking the X value, and then reset the routing path to avoid those malicious nodes, so as to protect the security of the routing path. Because it is designed for common routing attacks, RAPD algorithm can deal with a variety of existing and future possible routing attacks, and has a good performance of broad-spectrum detection to routing security. RAPD method has first pull into the "Rumor" mechanism: Carrying the detection packets in the cyclical message packets which originally maintain the P2P network topology, both avoided changing too much to the original routing algorithm, and prevented malicious nodes to identify the detection packets. Thus greatly reduces the detection cost, and improves the reliability of detection and the compatibles with the original routing algorithm.FTARH (Fault-tolerant Algorithm of Routing based on Hiberarchy), a hierarchical fault-tolerant routing algorithm is proposed. It at first transformed the de Bruijn diagram into a structured non-uniform directed graph to match the actual structure, then split the P2P network topology into two closely connected sub-graph of de Bruijn. One is included by a better performance of the nodes, and the other by the poor performance of the nodes. High-performance node is responsible for the backbone routing and the updates of resources distribution, as well as the maintenance of local network topology. They also provide the routing information or data transfer services to the low performance nodes in order to improve the fault-tolerant of the routing. Besides, spare nodes are specially made to prevent the sudden failure of a high-performance nodes, which will led to a long period of confusion in local network.DAASLB (Load-balance Algorithm based on Dynamic Assignment of Address Space), a load-balanc algorithm based on dynamically assigning addresses is presented. It is not based on "resources distributed in the node uniformly", "nodes distributed uniformly in the address space", but based on the node's actual situation, such as bandwidth, load capacity, quantity of resources, topological location, to assign a continuous, permanent ID address space, and then use the improved virtual server technology to mapping a physical node into a number of virtual nodes, each virtual node occupy an ID address and bear some of the load to act as a true P2P network nodes. Finally, according to the performance of the physical node or the changes of network size to adjust the number of virtual nodes dynamically, in order to adjust load and transfer load intelligently. It show that, DAASLB algorithm are fast converged, less overhead and load balancing, and greatly improved the load capacity and working efficiency of P2P network.TMAGDM (Topology-matching Algorithm based on Geography-distance Measurement), a topology matching algorithm based on geographic position is proposed. It firstly gives an improved global network positioning (GNP) method, to map each physical node into a point in N dimensional space, and for the first time introduce latency and bandwidth as the main parameters in the coordinate formula of geometric point, therefore, the logic distance between the points basically corresponds to the physical distance between nodes. And accordingly to this, proposed a new algorithm for the establishment and updating of the routing table, which greatly improved the topology matching between the logical topology and physical topology, as well as increased the working efficiency of P2P network. So as to effectively prevent the attacks utilized the low efficiency problem caused by the topology mismatching.
Keywords/Search Tags:Secure Routing, Fault-tolerant Routing, Load Balance, Topology Mismatching, Structured Topology, Distributed Hash Table, Global Network Position, Peer-to-Peer Network
PDF Full Text Request
Related items