对等网络路由优化及相关技术研究
|
摘要
随着互联网的飞速发展，教育资源日益丰富。为了更好地共享教育资源，并提高系统的可扩展性以及健壮性，将结构化P2P技术引入教育资源共享系统具有重要的意义。但传统结构化P2P技术无法满足教育资源共享系统对复杂查询的需求，因此，本文围绕如何适应复杂查询功能的需求并进一步提高系统查询效率和性能，针对教育资源特点，从基本路由机制、负载均衡、抗Churn性等方面寻求改进结构化P2P网络的方法和途径。经过近几年的研究，在阅读大量文献的基础上，取得了一定的研究成果。本文提出了一种支持多属性、多关键词模糊匹配查询的结构化P2P路由机制，在实现教育资源复杂查询的同时，提高了有限跳数内的查全率；在分析DHT负载不均的内在成因的基础上，提出了包含相似相近推移机制和动态转发均衡机制在内的综合负载均衡机制，在提高支持复杂查询的结构化P2P网络的负载均衡效果、兼顾系统复杂查询效率的同时，减少了节点频繁移动所带来的额外负载；针对支持复杂查询的结构化P2P网络的查询“热点”问题，提出了具有自适应性的热点集合反向缓存机制，加快了系统负载均衡速度，同时减少了统计访问信息所带来的额外负载；分析了P2P网络的Churn问题及其解决方法，提出了具有自适应性的层次DHT网络模型，并研究了该模型的路由优化，提出了查询优先级限定机制、查询请求处理负载均衡机制和双变换负载均衡机制，提高了系统的抗Churn性，有利于促进系统整体性能的提升。
经过近几年的研究，达到了预期目的，本文的主要创新点如下。
（1）提出了一种支持多属性、多关键词模糊匹配查询的结构化P2P资源信息检索模型，并优化了其路由机制。
为了更好地适应教育资源共享系统中资源属性值的特点，在提供复杂信息检索服务的同时，提高系统的可扩展性和适应性以及对用户请求的响应速度，本文以Chord技术为基础提出了一种支持多属性、多关键词模糊匹配查询的扩展资源信息检索模型MAKF_Chord。该模型改进了资源键值的生成方式，提出了双向指纹机制；改进了资源信息的映射方式，提出了双向映射机制；改进了Chord的基本路由，提出了适合于复杂查询的动态转发机制；并利用“预判断”过程进一步提高了系统查询效率。实验结果表明MAKF_Chord具有较高的查全率和负载均衡性。
（2）分析了结构化P2P网络负载不均衡的内在成因，提出了包含相似相近推移机制和动态转发均衡机制在内的综合负载均衡机制。
为了进一步提高支持复杂查询的结构化P2P资源共享系统的负载均衡效果，在减少节点频繁移动所带来额外负载、维护系统稳定性的同时，兼顾系统查询效率，本文提出了相似相近推移机制SCT和动态转发均衡机制DFBM，并将这两部分机制应用到本文所提出的负载均衡模型SP_ALBM中。其中，相似相近推移机制SCT将重负载节点上、推移效益最大的相似资源信息集合推移到最大距离限定范围内、负载较轻的邻近节点，在提高负载均衡效果的同时，减少了节点的频繁移动。SCT所提出的资源信息推移请求批处理机制实现了单位相对负载推移效益和的最大化；“条件概率检索”和“条件转发”机制则减少了不必要的查询处理和查询请求转发，优化了SCT推移机制；部分节点非周性移动机制对键值空间的分配进行非周期性调整，进一步提高了SCT负载均衡性能。动态转发均衡机制DFBM根据节点的周期性负载统计信息和当前成功转发请求速率的变化实时调整转发负载调节因子的大小；并依据此因子值，主动限制重负载节点的查询请求转发量；同时，通过查询范围的负载感知划分和转发节点筛选，将更多的请求转发给轻负载节点，从而提高了系统负载的均衡性。
（3）针对支持复杂查询的结构化P2P网络的查询“热点”问题，提出了具有自适应性的热点集合反向缓存机制。
针对支持复杂查询的结构化P2P网络因“热点”问题而引起的负载不均，为了能够加快系统负载均衡速度，同时减少统计访问信息所带来的额外负载，本文改进了以单个文件为基本单位的传统访问信息统计模式以及缓存管理机制，提出了具有自适应性的热点集合反向缓存机制HSRC。该机制可以根据节点所存储的资源信息数目，自适应地确定访问信息统计的基本单位，并进行相应热源信息统计，同时将依此确定的包含热点资源信息及其相似资源信息在内的热点集合同步缓存于命中本节点次数最多的反向节点，在减少了所需要维护的访问信息量的同时，加快了节点负载均衡速度；并针对结构化P2P网络中三类常见的复杂查询资源信息映射机制，给出了具体的热点集合自适应统计算法；同时，改进了节点缓存管理机制，提出了基于缓存平衡因子的周期性最不经常命中算法PLFH和分布式热点集合缓存更新算法，进一步提高了系统负载均衡效果。
（4）分析了P2P网络的Churn问题及其解决方法，提出了具有自适应性的层次DHT网络模型，并进一步研究了该模型的路由优化，提出了查询优先级限定机制和针对性负载均衡机制。
本文分析了现有层次结构化P2P网络在系统维护以及负载均衡等方面存在的问题，并提出了改进方案——具有自适应性的层次DHT网络模型AH_DHT。该模型改进了节点分类和加入/退出方式，使节点能根据系统当前加入/退出频率进行自动调整，在减少系统维护负载的同时，增强了层次网络的自适应性和抗Churn性；改进了下层节点的网络连接方式和资源信息查找算法，在减弱单个超级节点失效对下层节点的影响的同时，提高了下层节点的查询效率；同时，提出了查询请求处理负载均衡机制和双变换负载均衡机制，充分利用下层节点的操作、资源和自适应性，进一步均衡了系统负载；并提出了基于转发跳数限制和有限抢占截止时间优先队列管理的查询优先级限定机制，对不同优先级的节点提供不同质量的资源共享服务，有利于激励下层节点、促进网络系统整体性能的提高。
Educational resources are increased significantly with the rapid development of the Internet. Itis significant to introduce the structured P2P technology into the educational resource sharingsystem in order to share the educational resources better and improve the scalability androbustness of the system. But the traditional structured P2P technology can not meet the demandof the complex query. Therefore, this paper focuses on how to meet the demand of the complexquery and further improve the query efficiency and the performance of the system, and exploresthe ways to improve the structured P2P network from the four aspects (the basic routing, the loadbalancing, the Churn resistance and the incentive mechanism) according to the characteristics ofthe educational resources. Some research results are obtained after several years’ research. Thispaper presents a structured P2P routing mechanism which can support the multiple-attributemultiple-keyword fuzzing-matching complex query and improve the recall ratio in the limitedhops. It analyzes the internal causes of the DHT load imbalance and presents an integrated loadbalancing mechanism which includes two new mechanisms: the similar-close-transpositionmechanism and the dynamic-forwarding mechanism. The mechanism can improve the loadbalancing of the structured P2P system which supports the complex query, and can reduce theadditional load caused by the frequent mobility of the nodes while giving consideration to thequery efficiency of the system. It also presents an adaptive hot-set-reverse-cache mechanism forthe “hot-spot” issue of the structured P2P networks which support complex queries. Thehot-set-reverse-cache mechanism can accelerate the speed of the system load balancing, andreduce the additional load caused by the access information statistics. Finally, the paper analyzesthe causes and the solving methods of the Churn problem in the P2P network, and presents anadaptive hierarchical DHT network model. It discusses the routing optimization of the model andpresents three sub-mechanisms: the query-priority-limit mechanism, the query processing loadbalancing mechanism and the double-transform load balancing mechanism, which improves theChurn resistance of the system and is conducive to the overall performance upgrade of thesystem.
The main innovation points of this paper are as follows.
(1) This paper presents a structured P2P resource information retrieval model which cansupport the multiple-attribute multiple-keyword fuzzing-matching complex query, andoptimizes the routing mechanism of it.
It proposes a structured P2P resource information retrieval model (MAKF_Chord) which isbased on Chord and supports the multiple-attribute multiple-keyword fuzzing-matching complexquery, in order to better meet the characteristics of the educational resource attribute values andimprove the scalability and adaptability of the system and the response speed of the user request.This model improves the generation mode of the resource key-value and puts forward thebidirectional-fingerprint mechanism. It also improves the mapping mode of the resource information and proposes a bidirectional-mapping mechanism. Then, it improves the basicrouting of Chord and proposes a dynamic forwarding mechanism for the complex query. It usesthe prediction process to further improve the query efficiency of the system. The experimentalresults show that the model has higher recall ratio and better load balancing.
(2) This paper analyzes the load imbalancing causes of the structured P2P network andproposes an integrated load balancing mechanism which includes thesimilar-close-transposition mechanism and the dynamic-forwarding-balancing mechanism.
It proposes the similar-close-transposition mechanism and the dynamic-forwarding-balancingmechanism to further improve the load balancing of the complex-query sharing system andreduce the additional load caused by the nodes’ frequent moving while giving consideration tothe query efficiency of the system. It also applies the two mechanisms to the load balancingmodel (SP_ALBM) which is proposed in the paper. The similar-close-transposition mechanismselects the similar-resource-information set which has the greatest transposition benefit in theheavy-load-node, transfers it to the adjacent light-load-node in a maximum limited range, andreduces the mobility of the nodes while improving the load balancing effect. This mechanismrealizes the maximization of the total transposition benefit of the relative load using the batchmechanism of the request processing. The similar-close-transposition mechanism reduces theadditional load of the resource transposition and optimizes the transposition through the“conditional probability retrieval” mechanism and the “conditional forwarding” mechanism. Andits “aperiodic moving” mechanism can aperiodically adjust the distribution of the key space andfurther improve the performance of the load balancing. The dynamic-forwarding-balancingmechanism adjusts the forwarding-load regulation-factor of the node according to the periodicload statistics and the change of the successful forwarding rate, and limits the forwarding of theheavy-load-nodes according to the factor. It applies the load sensing division of the query rangeand the forwarding node selection to forwarding more requests to the light-load-nodes, so as toimprove the routing load balancing.
(3) This paper proposes an adaptive hot-set-reverse-cache mechanism for the “hot-spot”issue of the structured P2P networks which support complex queries.
It improves the cache management mechanism and the traditional access information statisticswhose basic unit is the single file in order to betterly solve the load imbalancing caused by the“hot-spot” issue in the structured P2P networks which support complex queries. And it proposesan adaptive hot-set-reverse-cache mechanism which is called HSRC. The mechanism canadaptively determine the basic unit of the access information statistics according to the numberof the resource information records which are stored on the node, adds up the hot-sourceinformation and caches the “hot-spot” resource information records and the similar ones into thereverse node which hits the current node most. The mechanism reduces the access informationneeded to maintenance, and speeds up the load balancing. This paper gives the adaptive hot-setstatistics algorithms for three kinds of common resource information mapping of the structuredP2P networks which support complex queries. It also improves the cache managementmechanism, and proposes the cache-balancing-factor-based periodic least frequently hitalgorithm (PLFH) and the distributed hot-set-cache update algorithm which further improve the load balancing effect of the system.
(4) This paper analyzes the Churn problem of P2P networks and the solving methods,and presents an adaptive hierarchical DHT network model. It also further studies therouting optimization and proposes the query-priority-limit mechanism and the pertinentload balancing mechanism.
It analyzes the problems of the existing hierarchical DHT network in the aspects of the systemmaintance and the load balancing, and proposes an adaptive hierarchical DHT network model.This model improves the joining/leaving methods of nodes so that the nodes can adjust theirmethod according to the current joining/leaving frequency of the system, which reduces the
maintenance load and improves the Churn resistance of the system. It improves the networkconnections and the resource information search algorithm of the nodes in the lower layer,which improves the query efficiency while weakening the influence of super nodes’ failure onthe nodes in the lower layer. This paper proposes the query-processing load-balancingmechanism and double-transform load-balancing mechanism to further balance the system loadby using the operation, the resource and the adaptability of the node in the lower layer. It alsoproposes the query-priority-limit mechanism which is based on the forwarding-hop-limit andthe limited-preemption deadline-priority queue management. The mechanism is conducive toencouraging the lower nodes and promoting the overall performance of the network system.
经过近几年的研究，达到了预期目的，本文的主要创新点如下。
（1）提出了一种支持多属性、多关键词模糊匹配查询的结构化P2P资源信息检索模型，并优化了其路由机制。
为了更好地适应教育资源共享系统中资源属性值的特点，在提供复杂信息检索服务的同时，提高系统的可扩展性和适应性以及对用户请求的响应速度，本文以Chord技术为基础提出了一种支持多属性、多关键词模糊匹配查询的扩展资源信息检索模型MAKF_Chord。该模型改进了资源键值的生成方式，提出了双向指纹机制；改进了资源信息的映射方式，提出了双向映射机制；改进了Chord的基本路由，提出了适合于复杂查询的动态转发机制；并利用“预判断”过程进一步提高了系统查询效率。实验结果表明MAKF_Chord具有较高的查全率和负载均衡性。
（2）分析了结构化P2P网络负载不均衡的内在成因，提出了包含相似相近推移机制和动态转发均衡机制在内的综合负载均衡机制。
为了进一步提高支持复杂查询的结构化P2P资源共享系统的负载均衡效果，在减少节点频繁移动所带来额外负载、维护系统稳定性的同时，兼顾系统查询效率，本文提出了相似相近推移机制SCT和动态转发均衡机制DFBM，并将这两部分机制应用到本文所提出的负载均衡模型SP_ALBM中。其中，相似相近推移机制SCT将重负载节点上、推移效益最大的相似资源信息集合推移到最大距离限定范围内、负载较轻的邻近节点，在提高负载均衡效果的同时，减少了节点的频繁移动。SCT所提出的资源信息推移请求批处理机制实现了单位相对负载推移效益和的最大化；“条件概率检索”和“条件转发”机制则减少了不必要的查询处理和查询请求转发，优化了SCT推移机制；部分节点非周性移动机制对键值空间的分配进行非周期性调整，进一步提高了SCT负载均衡性能。动态转发均衡机制DFBM根据节点的周期性负载统计信息和当前成功转发请求速率的变化实时调整转发负载调节因子的大小；并依据此因子值，主动限制重负载节点的查询请求转发量；同时，通过查询范围的负载感知划分和转发节点筛选，将更多的请求转发给轻负载节点，从而提高了系统负载的均衡性。
（3）针对支持复杂查询的结构化P2P网络的查询“热点”问题，提出了具有自适应性的热点集合反向缓存机制。
针对支持复杂查询的结构化P2P网络因“热点”问题而引起的负载不均，为了能够加快系统负载均衡速度，同时减少统计访问信息所带来的额外负载，本文改进了以单个文件为基本单位的传统访问信息统计模式以及缓存管理机制，提出了具有自适应性的热点集合反向缓存机制HSRC。该机制可以根据节点所存储的资源信息数目，自适应地确定访问信息统计的基本单位，并进行相应热源信息统计，同时将依此确定的包含热点资源信息及其相似资源信息在内的热点集合同步缓存于命中本节点次数最多的反向节点，在减少了所需要维护的访问信息量的同时，加快了节点负载均衡速度；并针对结构化P2P网络中三类常见的复杂查询资源信息映射机制，给出了具体的热点集合自适应统计算法；同时，改进了节点缓存管理机制，提出了基于缓存平衡因子的周期性最不经常命中算法PLFH和分布式热点集合缓存更新算法，进一步提高了系统负载均衡效果。
（4）分析了P2P网络的Churn问题及其解决方法，提出了具有自适应性的层次DHT网络模型，并进一步研究了该模型的路由优化，提出了查询优先级限定机制和针对性负载均衡机制。
本文分析了现有层次结构化P2P网络在系统维护以及负载均衡等方面存在的问题，并提出了改进方案——具有自适应性的层次DHT网络模型AH_DHT。该模型改进了节点分类和加入/退出方式，使节点能根据系统当前加入/退出频率进行自动调整，在减少系统维护负载的同时，增强了层次网络的自适应性和抗Churn性；改进了下层节点的网络连接方式和资源信息查找算法，在减弱单个超级节点失效对下层节点的影响的同时，提高了下层节点的查询效率；同时，提出了查询请求处理负载均衡机制和双变换负载均衡机制，充分利用下层节点的操作、资源和自适应性，进一步均衡了系统负载；并提出了基于转发跳数限制和有限抢占截止时间优先队列管理的查询优先级限定机制，对不同优先级的节点提供不同质量的资源共享服务，有利于激励下层节点、促进网络系统整体性能的提高。
Educational resources are increased significantly with the rapid development of the Internet. Itis significant to introduce the structured P2P technology into the educational resource sharingsystem in order to share the educational resources better and improve the scalability androbustness of the system. But the traditional structured P2P technology can not meet the demandof the complex query. Therefore, this paper focuses on how to meet the demand of the complexquery and further improve the query efficiency and the performance of the system, and exploresthe ways to improve the structured P2P network from the four aspects (the basic routing, the loadbalancing, the Churn resistance and the incentive mechanism) according to the characteristics ofthe educational resources. Some research results are obtained after several years’ research. Thispaper presents a structured P2P routing mechanism which can support the multiple-attributemultiple-keyword fuzzing-matching complex query and improve the recall ratio in the limitedhops. It analyzes the internal causes of the DHT load imbalance and presents an integrated loadbalancing mechanism which includes two new mechanisms: the similar-close-transpositionmechanism and the dynamic-forwarding mechanism. The mechanism can improve the loadbalancing of the structured P2P system which supports the complex query, and can reduce theadditional load caused by the frequent mobility of the nodes while giving consideration to thequery efficiency of the system. It also presents an adaptive hot-set-reverse-cache mechanism forthe “hot-spot” issue of the structured P2P networks which support complex queries. Thehot-set-reverse-cache mechanism can accelerate the speed of the system load balancing, andreduce the additional load caused by the access information statistics. Finally, the paper analyzesthe causes and the solving methods of the Churn problem in the P2P network, and presents anadaptive hierarchical DHT network model. It discusses the routing optimization of the model andpresents three sub-mechanisms: the query-priority-limit mechanism, the query processing loadbalancing mechanism and the double-transform load balancing mechanism, which improves theChurn resistance of the system and is conducive to the overall performance upgrade of thesystem.
The main innovation points of this paper are as follows.
(1) This paper presents a structured P2P resource information retrieval model which cansupport the multiple-attribute multiple-keyword fuzzing-matching complex query, andoptimizes the routing mechanism of it.
It proposes a structured P2P resource information retrieval model (MAKF_Chord) which isbased on Chord and supports the multiple-attribute multiple-keyword fuzzing-matching complexquery, in order to better meet the characteristics of the educational resource attribute values andimprove the scalability and adaptability of the system and the response speed of the user request.This model improves the generation mode of the resource key-value and puts forward thebidirectional-fingerprint mechanism. It also improves the mapping mode of the resource information and proposes a bidirectional-mapping mechanism. Then, it improves the basicrouting of Chord and proposes a dynamic forwarding mechanism for the complex query. It usesthe prediction process to further improve the query efficiency of the system. The experimentalresults show that the model has higher recall ratio and better load balancing.
(2) This paper analyzes the load imbalancing causes of the structured P2P network andproposes an integrated load balancing mechanism which includes thesimilar-close-transposition mechanism and the dynamic-forwarding-balancing mechanism.
It proposes the similar-close-transposition mechanism and the dynamic-forwarding-balancingmechanism to further improve the load balancing of the complex-query sharing system andreduce the additional load caused by the nodes’ frequent moving while giving consideration tothe query efficiency of the system. It also applies the two mechanisms to the load balancingmodel (SP_ALBM) which is proposed in the paper. The similar-close-transposition mechanismselects the similar-resource-information set which has the greatest transposition benefit in theheavy-load-node, transfers it to the adjacent light-load-node in a maximum limited range, andreduces the mobility of the nodes while improving the load balancing effect. This mechanismrealizes the maximization of the total transposition benefit of the relative load using the batchmechanism of the request processing. The similar-close-transposition mechanism reduces theadditional load of the resource transposition and optimizes the transposition through the“conditional probability retrieval” mechanism and the “conditional forwarding” mechanism. Andits “aperiodic moving” mechanism can aperiodically adjust the distribution of the key space andfurther improve the performance of the load balancing. The dynamic-forwarding-balancingmechanism adjusts the forwarding-load regulation-factor of the node according to the periodicload statistics and the change of the successful forwarding rate, and limits the forwarding of theheavy-load-nodes according to the factor. It applies the load sensing division of the query rangeand the forwarding node selection to forwarding more requests to the light-load-nodes, so as toimprove the routing load balancing.
(3) This paper proposes an adaptive hot-set-reverse-cache mechanism for the “hot-spot”issue of the structured P2P networks which support complex queries.
It improves the cache management mechanism and the traditional access information statisticswhose basic unit is the single file in order to betterly solve the load imbalancing caused by the“hot-spot” issue in the structured P2P networks which support complex queries. And it proposesan adaptive hot-set-reverse-cache mechanism which is called HSRC. The mechanism canadaptively determine the basic unit of the access information statistics according to the numberof the resource information records which are stored on the node, adds up the hot-sourceinformation and caches the “hot-spot” resource information records and the similar ones into thereverse node which hits the current node most. The mechanism reduces the access informationneeded to maintenance, and speeds up the load balancing. This paper gives the adaptive hot-setstatistics algorithms for three kinds of common resource information mapping of the structuredP2P networks which support complex queries. It also improves the cache managementmechanism, and proposes the cache-balancing-factor-based periodic least frequently hitalgorithm (PLFH) and the distributed hot-set-cache update algorithm which further improve the load balancing effect of the system.
(4) This paper analyzes the Churn problem of P2P networks and the solving methods,and presents an adaptive hierarchical DHT network model. It also further studies therouting optimization and proposes the query-priority-limit mechanism and the pertinentload balancing mechanism.
It analyzes the problems of the existing hierarchical DHT network in the aspects of the systemmaintance and the load balancing, and proposes an adaptive hierarchical DHT network model.This model improves the joining/leaving methods of nodes so that the nodes can adjust theirmethod according to the current joining/leaving frequency of the system, which reduces the
maintenance load and improves the Churn resistance of the system. It improves the networkconnections and the resource information search algorithm of the nodes in the lower layer,which improves the query efficiency while weakening the influence of super nodes’ failure onthe nodes in the lower layer. This paper proposes the query-processing load-balancingmechanism and double-transform load-balancing mechanism to further balance the system loadby using the operation, the resource and the adaptability of the node in the lower layer. It alsoproposes the query-priority-limit mechanism which is based on the forwarding-hop-limit andthe limited-preemption deadline-priority queue management. The mechanism is conducive toencouraging the lower nodes and promoting the overall performance of the network system.
引文
[1] Stoica I, Morris R, Karger D, et al. Chord: A scalable peer-to-peer lookup service for internetapplications[J]. ACM SIGCOMM Computer Communication Review,2001,31(4):149-160.
[2] Stoica I, Morris R, Liben-Nowell D, et al. Chord: a scalable peer-to-peer lookup protocol for internetapplications[J]. IEEE/ACM Transactions on Networking,2003,11(1):17-32.
[3]杨天路，刘宇宏等．P2P网络技术原理与系统开发案例[M]．人民邮电出版社，2007：313．
[4] Saroiu S, Gummadi K P, Gribble S D. Measuring and analyzing the characteristics of Napster andGnutella hosts[J]. Multimedia systems,2003,9(2):170-184.
[5] Ripeanu M. Peer-to-peer architecture case study: Gnutella network[C]. Proceedings of the FirstInternational Conference on Peer-to-Peer Computing. IEEE,2001:99-100.
[6]钱宁，吴国新．无结构化P2P网络资源搜索机制研究综述[J]．计算机科学，2010，37(4)：7-10．
[7]熊伟．结构化对等网络路由机制关键技术研究[D]．长沙：湖南大学，2008：
[8]徐明．对等网络中的查询路由与重构机制[D]．上海：复旦大学，2010：
[9] El-Ansary S, Alima L, Brand P, et al. Efficient broadcast in structured P2P networks [J]. Peer-to-PeerSystems II,2003:304-314.
[10] Merz P, Gorunova K. Efficient broadcast in P2P grids[C]. Proceedings of the IEEE InternationalSymposium on CCGrid2005. IEEE,2005,1:237-242.
[11] Cox R, Muthitacharoen A, Morris R. Serving DNS using a peer-to-peer lookup service[J].Peer-to-Peer Systems,2002:155-165.
[12] Dabek F, Kaashoek M F, Karger D, et al. Wide-area cooperative storage with CFS[J]. ACM SIGOPSOperating Systems Review,2001,35(5):202-215.
[13] Ajmani S, Clarke D, Moh C H, et al. ConChord: Cooperative SDSI certificate storage and nameresolution[J]. Peer-to-Peer Systems,2002:141-154.
[14] Muthitacharoen A, Morris R, Gil T M, et al. Ivy: A read/write peer-to-peer file system[J]. ACMSIGOPS Operating Systems Review,2002,36(SI):31-44.
[15] Rowstron A, Druschel P. Pastry: Scalable, decentralized object location, and routing for large-scalepeer-to-peer systems[C]. Middleware2001. Springer Berlin/Heidelberg,2001:329-350.
[16] Rowstron A, Druschel P. Storage management and caching in PAST, a large-scale, persistentpeer-to-peer storage utility[C]. ACM SIGOPS Operating Systems Review. ACM,2001,35(5):188-201.
[17] Castro M, Druschel P, Kermarrec A M, et al. SCRIBE: A large-scale and decentralizedapplication-level multicast infrastructure[J]. IEEE Journal on Selected Areas in Communications,2002,20(8):1489-1499.
[18] Ratnasamy S, Francis P, Handley M, et al. A scalable content-addressable network[M]. ACM,2001.
[19] Tsoumakos D, Roussopoulos N. A comparison of peer-to-peer search methods[C]. Proceedings of theSixth International Workshop on the Web and Databases.2003.
[20] Zhao B Y, Huang L, Stribling J, et al. Tapestry: A resilient global-scale overlay for servicedeployment[J]. IEEE Journal on Selected Areas in Communications,2004,22(1):41-53.
[21] Kubiatowicz J, Bindel D, Chen Y, et al. Oceanstore: An architecture for global-scale persistentstorage[J]. ACM Sigplan Notices,2000,35(11):190-201.
[22] Dhara K, Guo Y, Kolberg M, et al. Overview of Structured Peer-to-Peer Overlay Algorithms[J].Handbook of Peer-to-Peer Networking,2010:223-256.
[23] Good N S, Krekelberg A. Usability and privacy: a study of Kazaa P2P file-sharing[C]. Proceedings ofthe SIGCHI conference on Human factors in computing systems. ACM,2003:137-144.
[24] Baset S A, Schulzrinne H. An analysis of the skype peer-to-peer internet telephony protocol[J]. arXivpreprint cs/0412017,2004.
[25] Liang J, Kumar R, Ross K W. The FastTrack overlay: A measurement study[J]. Computer Networks,2006,50(6):842-858.
[26] Zhao B, Duan Y, Huang L, et al. Brocade: Landmark routing on overlay networks[J]. Peer-to-PeerSystems,2002:34-44.
[27]龚海刚，刘明，毛莺池等．P2P流媒体关键技术的研究进展[J]．计算机研究与发展，2006，42(12)：2033-2040．
[28]刘亚杰．P2P流媒体内容分发关键技术研究[D]．长沙：国防科学技术大学，2005．
[29]杨戈，廖建新，朱晓民等．流媒体分发系统关键技术综述[J]．电子学报，2009，37(1)：137-145．
[30]秦丰林，刘琚．P2P网络流媒体关键技术[J]．电子学报，2011，39(4)：919-927．
[31]龚海刚，刘明，毛莺池等．P2P流媒体关键技术的研究进展[J]．计算机研究与发展，2006，42(12)：2033-2040．
[32] Foster I, Zhao Y, Raicu I, et al. Cloud computing and grid computing360-degree compared[C]. GridComputing Environments Workshop,2008. GCE'08. IEEE,2008:1-10.
[33] Armbrust M, Fox A, Griffith R, et al. A view of cloud computing[J]. Communications of the ACM,2010,53(4):50-58.
[34] Wang L, Von Laszewski G, Younge A, et al. Cloud computing: a perspective study[J]. NewGeneration Computing,2010,28(2):137-146.
[35] Geelan J. Twenty one experts define cloud computing[J]. Cloud Computing Journal,2009,4:1-5.
[36] Shankar S. Amazon elastic compute cloud[J].2009.
[37] Zahariev A. Google app engine[C]. TKK T-110.5190Seminar on Internetworking.2009:1-5.
[38] Xia T, Li Z, Yu N. Research on cloud computing based on deep analysis to typical platforms[J]. CloudComputing,2009:601-608.
[39] Joung Y J, Fang C T, Yang L W. Keyword Search in DHT-Based Peer-to-Peer Networks[C].Proceedings of the25th IEEE International Conference on Distributed Computing Systems, Columbus,OH,10June2005. Pages:339-348.
[40] Fu X, Feng B. Structured P2P search method to support complex query[J]. Mini-Micro Systems,2006,27(3):401-406.
[41] Jiang S, Guo L, Zhang X, et al. LightFlood: Minimizing redundant messages and maximizing scope ofPeer-to-Peer search[J]. IEEE Transactions on Parallel and Distributed Systems,2008,19(5):601-614.
[42] Chen H, Jin H, Liu Y, et al. Difficulty-aware hybrid search in peer-to-peer networks[J]. IEEETransactions on Parallel and Distributed Systems,2009,20(1):71-82.
[43] Gkantsidis C, Mihail M, Saberi A. Random walks in peer-to-peer networks[C]. Proceedings of theTwenty-third AnnualJoint Conference of the IEEE Computer and Communications Societies. NewYork: IEEE press,2004：120-130.
[44] Chen G, Low C P, Yang Z. Enhancing search performance in unstructured P2P networks based onusers' common interest[J]. IEEE Transactions on Parallel and Distributed Systems,2008,19(6):821-836.
[45] Khan S, Tokarchuk L N. Interest-based self organization in group-structured P2P networks[C].Proceedings of the6th Consumer Communications and Networking Conference. IEEE,2009:1-5.
[46] Kumar A, Xu J, Zegura E W. Efficient and scalable query routing for unstructured peer-to-peernetworks[C]. Proceedings of the24th Annual Joint Conference of the IEEE Computer andCommunications Societies. IEEE,2005,2:1162-1173.
[47]张一鸣，卢锡城，郑倩冰等．一种面向大规模P2P系统的快速搜索算法[J]．软件学报，2008，19(6)：1473-1480．
[48] Harren M, Hellerstein J, Huebsch R, et al. Complex queries in DHT-based peer-to-peer networks[J].Peer-to-Peer Systems,2002:242-250.
[49]钱宁，吴国新，赵生慧．基于贝叶斯网络的无结构化P2P资源搜索方法[J]．计算机研究与发展，2009，46(6)：889-897．
[50]邱志欢，肖明忠，代亚非．一种P2P环境下基于用户行为的语义检索方案[J]．软件学报，2007，18(9)：2216-2225．
[51] Zhou F, Zhuang L, Zhao B, et al. Approximate object location and spam filtering on peer-to-peersystems[C]. Middleware2003. Springer Berlin/Heidelberg,2003:997-998.
[52] Cai M, Frank M, Chen J, et al. Maan: A multi-attribute addressable network for grid informationservices[J]. Journal of Grid Computing,2004,2(1):3-14.
[53] Bharambe A R, Agrawal M, Seshan S. Mercury: supporting scalable multi-attribute range queries[C].Proceedings of SIGCOMM’04,2004:353–366.
[54] Liu B, Lee W C, Lee D L. Supporting complex multi-dimensional queries in P2P systems[C].ICDCS’05,2005:155–164.
[55] Shu Y, Ooi B C, Tan K L, et al. Supporting multi-dimensional range queries in peer-to-peersystems[C]. P2P’05,2005:173–180.
[56] Banaei-Kashani F, Shahabi C. SWAM: a family of access methods for similarity-search in peer-to-peerdata networks[C]. Proceedings of CIKM’04,2004:304–313.
[57] Novak D, Zezula P. M-Chord: a scalable distributed similarity search structure[C]. Proceedings ofInfoScale’06,2006:19.104
[58] Bawa M, Condie T, Ganesan P. LSH forest: self-tuning indexes for similarity search[C]. Proceedingsof WWW’05,2005:651–660.
[59]聂晓文，卢显良，李梁等．DHT负载均衡的必要性[J]．计算机科学，2009，36(9)：92-95．
[60] King V, Saia J. Choosing a random peer[C]. Proceedings of the twenty-third annual ACM symposiumon Principles of distributed computing. ACM,2004:125-130.
[61] Raab M, Steger A.“Balls into Bins”—A Simple and Tight Analysis[J]. Randomization andApproximation Techniques in Computer Science,1998:159-170.
[62] Handurukande SB, Kermarrec AM, Le Fessant F, et a1. Peer sharing behavior in the eDonkey network,and implications for the design of server—less file sharing systems[J]. SIGOPS Oper. Syst. Rev.,2006,40(4):359-371.
[63] Breslau L, Cao P, Fan L, et a1. Web caching and Zipf-like distributions: Evidence and implications[C].Proceedings of the IEEE INFOCOM’99. Los Alamitos: IEEE Press,1999,126-134.
[64] Jung J, Sit E, Balakrishnan H, et a1. DNS performance and the effectiveness of caching[J].IEEE/ACM Transactions on Networking,2002,10(5):589-603.
[65] Karger D, Ruhl M. Simple efficient load balancing algorithms for peer-to-peer systems. In: VoelkermGM, Shenker S, eds. LNCS.Berlin: Springer-Verlag,2005.131-140.
[66] Karthik A R, Lakshminarayanan K, Surana S, Karp R, Stoica I. Load balancing in structured P2Psystems. In: Kaashoek MF, Stoica I, eds. LNCS. Berlin: Springer-Verlag,2003.68-79[7]
[67] Li ZN, Xie GG. A load balancing algorithm for DHT-Based P2P systems[J]. Journal of ComputerResearch and Development,2006,43(9):1579-1585.
[68] Xu Z, Bhuyan L. Effective load balancing in P2P systems[C]. Proceedings of the6th IEEE Int’l Symp.on Cluster Computing and the Grid. Washington: IEEE Computer Society,2006.81-88.
[69]宋宝燕，高楠等．DLRD：一种支持动态负载均衡的P2P网格资源发现机制[J]．通信学报，2008，29(8)：94-99．
[70]刘德辉等．QFMA：一种支持负载均衡的多属性资源定位方法[J]．计算机学报，2008，31(8)：1376-1382．
[71]熊伟，谢冬青等．一种结构化P2P协议中的自适应负载均衡方法[J]．软件学报，2009，20(3)：660-670．
[72]司加全，杨武．一种结构化P2P存储系统的负载均衡方法[J]．南京理工大学学报（自然科学版），2011，35(1)：38-41．
[73]廖小波．结构化对等网络中基于访问热点的负载均衡算法研究[D]．成都：电子科技大学，2009：
[74]弭伟，张春红等．基于结构化P2P系统的SDYA负载均衡算法[J]．北京邮电大学学报，2010，33(6)：116-121．
[75] Zhu Y, Hu Y. Efficient Proximity-aware load balancing for DHT-based P2P systems[J]. IEEE Trans onParallel and Distributed Systems,2005,16(4):349-36.
[76]付志鹏，王怀民等．对等网络的抖动特性研究综述[J]．计算机学报，2011，34(9)：1563-1579．
[77] Rhea S C, Geels D, Roscoe T, et al. Handling churn in a DHT[M]. Computer Science Division,University of California,2003.
[78] Godfrey P, Shenker S, Stoica I. Minimizing churn in distributed systems[M]. ACM,2006.
[79] Yao Z, Wang X, Leonard D, et al. Node isolation model and age-based neighbor selection inunstructured P2P networks[J]. IEEE/ACM Transactions on Networking,2009,17(1):144-157.
[80] Legtchenko S, Monnet S, Sens P, et al. Churn-resilient replication strategy for peer-to-peer distributedhash-tables[J]. Stabilization, Safety, and Security of Distributed Systems,2009:485-499.
[81] Tan G, Jarvis S A. Stochastic Analysis and Improvement of the Reliability of DHT-basedMulticast[C]. Proceedings of the26th IEEE International Conference on Computer Communications.IEEE,2007:2198-2206.
[82] Anwitaman D, Roman S, Karl A. Query-Load balancing in structured overlays[C]. Proceedings of the7th IEEE Int’l Symp. on Cluster Computing and the Grid.2007.http://lsirpeople.epfl.ch/rschmidt/papers/Datta2007QueryLoad.pdf.
[83] Zhu C, Liu Z, Zhang WM, Xiao WD, Yang DS, Xu ZN. Load balance based on rumor mongering instructured P2P networks[J]. Journal of China Institute of Communications,2004,25(4):31-40.
[84]肖卓程，荆金华．层次式Chord：物理拓扑感知的结构化对等网[J]．计算机科学，2006，33(7)：25-28．
[85] Ganesan P, Gummadi K, Garcia-Molina H. Canon in G major: designing DHTs with hierarchicalstructure[C]. Proceedings of the24th International Conference on Distributed Computing Systems.IEEE,2004:263-272.
[86] Joung Y J, Wang J C. Chord2: A two-layer Chord for reducing maintenance overhead viaheterogeneity[J]. Computer Networks,2007,51(3):712-73.
[87] Xu Z, Min R, Hu Y. HIERAS: a DHT based hierarchical P2P routing algorithm[C]. Proceedings of the2003International Conference on Parallel Processing. IEEE,2003:187-194.
[88] Martinez-Yelmo I, Cuevas R, Guerrero C, et al. Routing performance in a hierarchical dht-basedoverlay network[C]. Proceedings of the16th Euromicro Conference on Parallel, Distributed andNetwork-Based Processing. IEEE,2008:508-515.
[89] Garces-Erice L, Biersack E W, Ross K W, Felber P A,Urvoy-Keller G. Hierarchical peer-to-peersystems[J]. Parallel Processing Letters,2003,13(4):643-657.
[90] Artigas M S, Lopez P G, Ahullo J P, et al. Cyclone: A novel design schema for hierarchical dhts[C].Proceedings of the Fifth IEEE International Conference on Peer-to-Peer Computing. IEEE,2005:49-56.
[91] Zoels S, Despotovic Z, Kellerer W. Cost-based analysis of hierarchical dht design[C]. Proceedings ofthe Sixth IEEE International Conference on Peer-to-Peer Computing. IEEE,2006:233-239.
[92] Srebro J M. Content-based addressing in hierarchical distributed hash tables[D]. Rochester Institute ofTechnology,2009.
[93] Koskela T, Kassinen O, Ou Z, et al. Improving community management performance with two-levelhierarchical DHT overlays[J]. Journal of Internet Technology, Special Issue on Internet ResourceSharing and Discovery,2010,11(2):167-181.
[94] Fotiou N, Katsaros K V, Vasilakos X, et al. H-Pastry: An adaptive multi-level overlay internetwork[J].Athens University of Economics and Business, Tech. Rep,2011.
[95] Gupta R, Somani A K. Pricing strategy for incentivizing selfish nodes to share resources inpeer-to-peer (p2p) networks[C]. Proceedings of the12th IEEE International Conference on Networks.IEEE,2004,2:624-629.
[96] Friedman E J, Halpern J Y, Kash I. Efficiency and nash equilibria in a scrip system for p2pnetworks[C]. Proceedings of the7th ACM conference on Electronic commerce. ACM,2006:140-149.
[97]彭冬生，林闯，刘卫东．利用无私节点改善基于支付机制P2P应用的性能[J]．计算机学报，2008，31(6)：953-959．
[98] Ma R T B, Lee S, Lui J, et al. Incentive and service differentiation in P2P networks: a game theoreticapproach[J]. IEEE/ACM Transactions on Networking (TON),2006,14(5):978-991.
[99]黄保华．对等系统的安全与激励机制研究[D]．武汉：华中科技大学，2006：
[100] Parkes D C, Shneidman J. Distributed implementations of vickrey-clarke-groves mechanisms[C].Proceedings of the Third International Joint Conference on Autonomous Agents and MultiagentSystems-Volume1. IEEE Computer Society,2004:261-268.
[101] Feldman M, Chuang J. Overcoming free-riding behavior in peer-to-peer systems[J]. ACM SIGecomExchanges,2005,5(4):41-50.
[102] Montresor A, Jelasity M. PeerSim: A scalable P2P simulator[C]. Proceedings of the IEEE NinthInternational Conference on Peer-to-Peer Computing. IEEE,2009:99-100.
[103] Gian P J. PeerSim HOWTO: Build a New Protocol for the PeerSim1.0Simulator[J].2005-12-24).http://peersim. sourceforge. net/tutorial1/tutorial1. pdf.
[104]李勇，彭宇行，陈福接．大规模视频点播磁盘cache替换算法[J]．计算机研究与发展，2000，37(2)：207-212．
[105]牛新征，佘堃，秦科．移动P2P网络的协作缓存优化策略[J]．计算机研究与发展，2008，18(4)：656-665．
[106]沈磊，罗惠琼．基于用户行为特征的P2P代理缓存的研究[J]．计算机应用研究，2007，24(11)：106-108．
[107] Gupta A, Liskov B, Rodrigues R. Efficient routing for peer-to-peer overlays[C]. Proceedings of theFirst symp. on Networked Systems Design and Implementation (NSDI).2004:113-126.
[108] Chu J, Labonte K, Bissias G, et al. A trace-driven evaluation of chord[R]. Technical Report04-38,Dept. of Computer Science, Univ. of Massachusetts, Amherst, MA1003,2004.
[109]张浩．基于结构化对等网络的高性能计算平台关键技术研究[D]．武汉：华中科技大学，2009：
[110]杨冬，董平，张宏科．对等网络Churn问题评估模型与分析[J]．通信学报，2007，28(6)：39-47．
[111]李伟，徐正全，杨铸．利用虚拟节点的动态结构化P2P网络性能研究[J]．武汉大学学报（信息科学版），2011，36(12)：1504-1507．
[112]张宇祥，杨冬，张宏科．P2P网络中Churn问题研究[J]．软件学报，2009，20(5)：1362-1374．
[113] Li J, Stribling J, Morris R, et al. Bandwidth-Efficient management of DHT routing tables[C].Proceedings of the2nd Conference on Symp. on Networked Systems Design&Implementation106(NSDI2005). Berkeley: USENI Press,2005.99114.
[114] Herrera O, Znati T. Static resiliency vs. churn-resistance capability of DHT-protocols[C].Proceedings of the18th International Conference on Parallel and Distributed Computing Systems(PDCS2005). Phoenix: IASTED/ACTA Press,2005.141-147.
[115] Kersch P, Szabo R, Cheng L, Jean K, Galis A. Stochastic maintenance of overlays in structured P2Psystems[J]. Computer Communications,2008,31(3):603619.
[116] Wu D, Tian Y, Ng KW. Analytical study on improving DHT lookup performance under churn[C].Proceedings of the6th IEEE International Conf. on Peer-to-Peer Computing (P2P2006). Piscataway:IEEE Press,2006.249-258.
[117] Stutzbach D, Rejaie R. Improving lookup performance over a widely-deployed DHT[C].Proceedings of the IEEE INFOCOM.Washington: IEEE Press,2006.1-12.
[118] Saroiu S, Gummadi P K, Gribble S D. A measurement study of peer-to-peer File sharing systems[C].Proceedings of the Multimedia Computing and Networking. San Jose, USA,2002:156-17.
[119]黄敏，蔡志刚．缓存替换算法研究综述[J]．计算机科学，2006，33(12)：191-192．
[120]张宇祥，张宏科．一种层次结构化P2P网络中的负载均衡方法[J]．计算机学报，2010，33(9)：1580-1590．
[121]魏文红，向菲等．一种结构化P2P系统的负载平衡算法[J]．计算机科学，2010，37(4)：82-85．
[122]孙海燕，数据网格副本管理关键技术研究[D]．长沙：国防科学技术大学，2005：
[123] Zoels S, Despotovic Z, Kellerer W. On hierarchical DHT systems–An analytical approach foroptimal designs[J]. Computer Communications,2008,31(3):576-590.
[124] Zoels S, Despotovic Z, Kellerer W. Load balancing in a hierarchical DHT-based P2P system[C].Proceedings of the2007International Conference on Collaborative Computing: Networking,Applications and Worksharing. IEEE,2007:353-361.
[125]宾雪莲，杨玉海，金士尧．一种有限优先级的静态优先级分配算法[J]．软件学报，2004，15(6)：815-822．
[126] ButtazzoG C. Rate Monotonic vs．EDF：Judgment Day [J]．Real-Time Systems(S0922—6443)，2005，29(1)：5-26.
[127] Li W, Kavi K, Akl R. A non-preemptive scheduling algorithm for soft real-time systems[J].Computers&Electrical Engineering,2007,33(1):12-29.
[128]巴巍，张大波，李琦等．一种新的组优先级动态实时调度算法[J]．计算机科学，2009，36(7)：230-233．
[129] Khataniar G, Goswami D. Avoidance of churn rate through temporal centralization in Chord[J].Peer-to-Peer Networking and Applications,2011,4(3):251-258.
[130]王丹．P2P系统资源查询机制研究综述[J]．计算机科学，2004，31(9)：57-59．
[131]海沫．分布式系统中的资源发现机制综述[J]．计算机科学，2010，37(12)：12-17．
[132] Peng Z, Ting-lei H, Cai-xia L, et al. Research of P2P architecture based on cloud computing[C].Proceedings of the2010International Conference on Intelligent Computing and Integrated Systems(ICISS). IEEE,2010:652-655.
[133] Liang C C, Lin H F. Proactive Peer-to-Peer Architecture for Sharing Files on a Cloud Network[J].Advanced Science Letters,2011,4(11-12):11-13.
[2] Stoica I, Morris R, Liben-Nowell D, et al. Chord: a scalable peer-to-peer lookup protocol for internetapplications[J]. IEEE/ACM Transactions on Networking,2003,11(1):17-32.
[3]杨天路，刘宇宏等．P2P网络技术原理与系统开发案例[M]．人民邮电出版社，2007：313．
[4] Saroiu S, Gummadi K P, Gribble S D. Measuring and analyzing the characteristics of Napster andGnutella hosts[J]. Multimedia systems,2003,9(2):170-184.
[5] Ripeanu M. Peer-to-peer architecture case study: Gnutella network[C]. Proceedings of the FirstInternational Conference on Peer-to-Peer Computing. IEEE,2001:99-100.
[6]钱宁，吴国新．无结构化P2P网络资源搜索机制研究综述[J]．计算机科学，2010，37(4)：7-10．
[7]熊伟．结构化对等网络路由机制关键技术研究[D]．长沙：湖南大学，2008：
[8]徐明．对等网络中的查询路由与重构机制[D]．上海：复旦大学，2010：
[9] El-Ansary S, Alima L, Brand P, et al. Efficient broadcast in structured P2P networks [J]. Peer-to-PeerSystems II,2003:304-314.
[10] Merz P, Gorunova K. Efficient broadcast in P2P grids[C]. Proceedings of the IEEE InternationalSymposium on CCGrid2005. IEEE,2005,1:237-242.
[11] Cox R, Muthitacharoen A, Morris R. Serving DNS using a peer-to-peer lookup service[J].Peer-to-Peer Systems,2002:155-165.
[12] Dabek F, Kaashoek M F, Karger D, et al. Wide-area cooperative storage with CFS[J]. ACM SIGOPSOperating Systems Review,2001,35(5):202-215.
[13] Ajmani S, Clarke D, Moh C H, et al. ConChord: Cooperative SDSI certificate storage and nameresolution[J]. Peer-to-Peer Systems,2002:141-154.
[14] Muthitacharoen A, Morris R, Gil T M, et al. Ivy: A read/write peer-to-peer file system[J]. ACMSIGOPS Operating Systems Review,2002,36(SI):31-44.
[15] Rowstron A, Druschel P. Pastry: Scalable, decentralized object location, and routing for large-scalepeer-to-peer systems[C]. Middleware2001. Springer Berlin/Heidelberg,2001:329-350.
[16] Rowstron A, Druschel P. Storage management and caching in PAST, a large-scale, persistentpeer-to-peer storage utility[C]. ACM SIGOPS Operating Systems Review. ACM,2001,35(5):188-201.
[17] Castro M, Druschel P, Kermarrec A M, et al. SCRIBE: A large-scale and decentralizedapplication-level multicast infrastructure[J]. IEEE Journal on Selected Areas in Communications,2002,20(8):1489-1499.
[18] Ratnasamy S, Francis P, Handley M, et al. A scalable content-addressable network[M]. ACM,2001.
[19] Tsoumakos D, Roussopoulos N. A comparison of peer-to-peer search methods[C]. Proceedings of theSixth International Workshop on the Web and Databases.2003.
[20] Zhao B Y, Huang L, Stribling J, et al. Tapestry: A resilient global-scale overlay for servicedeployment[J]. IEEE Journal on Selected Areas in Communications,2004,22(1):41-53.
[21] Kubiatowicz J, Bindel D, Chen Y, et al. Oceanstore: An architecture for global-scale persistentstorage[J]. ACM Sigplan Notices,2000,35(11):190-201.
[22] Dhara K, Guo Y, Kolberg M, et al. Overview of Structured Peer-to-Peer Overlay Algorithms[J].Handbook of Peer-to-Peer Networking,2010:223-256.
[23] Good N S, Krekelberg A. Usability and privacy: a study of Kazaa P2P file-sharing[C]. Proceedings ofthe SIGCHI conference on Human factors in computing systems. ACM,2003:137-144.
[24] Baset S A, Schulzrinne H. An analysis of the skype peer-to-peer internet telephony protocol[J]. arXivpreprint cs/0412017,2004.
[25] Liang J, Kumar R, Ross K W. The FastTrack overlay: A measurement study[J]. Computer Networks,2006,50(6):842-858.
[26] Zhao B, Duan Y, Huang L, et al. Brocade: Landmark routing on overlay networks[J]. Peer-to-PeerSystems,2002:34-44.
[27]龚海刚，刘明，毛莺池等．P2P流媒体关键技术的研究进展[J]．计算机研究与发展，2006，42(12)：2033-2040．
[28]刘亚杰．P2P流媒体内容分发关键技术研究[D]．长沙：国防科学技术大学，2005．
[29]杨戈，廖建新，朱晓民等．流媒体分发系统关键技术综述[J]．电子学报，2009，37(1)：137-145．
[30]秦丰林，刘琚．P2P网络流媒体关键技术[J]．电子学报，2011，39(4)：919-927．
[31]龚海刚，刘明，毛莺池等．P2P流媒体关键技术的研究进展[J]．计算机研究与发展，2006，42(12)：2033-2040．
[32] Foster I, Zhao Y, Raicu I, et al. Cloud computing and grid computing360-degree compared[C]. GridComputing Environments Workshop,2008. GCE'08. IEEE,2008:1-10.
[33] Armbrust M, Fox A, Griffith R, et al. A view of cloud computing[J]. Communications of the ACM,2010,53(4):50-58.
[34] Wang L, Von Laszewski G, Younge A, et al. Cloud computing: a perspective study[J]. NewGeneration Computing,2010,28(2):137-146.
[35] Geelan J. Twenty one experts define cloud computing[J]. Cloud Computing Journal,2009,4:1-5.
[36] Shankar S. Amazon elastic compute cloud[J].2009.
[37] Zahariev A. Google app engine[C]. TKK T-110.5190Seminar on Internetworking.2009:1-5.
[38] Xia T, Li Z, Yu N. Research on cloud computing based on deep analysis to typical platforms[J]. CloudComputing,2009:601-608.
[39] Joung Y J, Fang C T, Yang L W. Keyword Search in DHT-Based Peer-to-Peer Networks[C].Proceedings of the25th IEEE International Conference on Distributed Computing Systems, Columbus,OH,10June2005. Pages:339-348.
[40] Fu X, Feng B. Structured P2P search method to support complex query[J]. Mini-Micro Systems,2006,27(3):401-406.
[41] Jiang S, Guo L, Zhang X, et al. LightFlood: Minimizing redundant messages and maximizing scope ofPeer-to-Peer search[J]. IEEE Transactions on Parallel and Distributed Systems,2008,19(5):601-614.
[42] Chen H, Jin H, Liu Y, et al. Difficulty-aware hybrid search in peer-to-peer networks[J]. IEEETransactions on Parallel and Distributed Systems,2009,20(1):71-82.
[43] Gkantsidis C, Mihail M, Saberi A. Random walks in peer-to-peer networks[C]. Proceedings of theTwenty-third AnnualJoint Conference of the IEEE Computer and Communications Societies. NewYork: IEEE press,2004：120-130.
[44] Chen G, Low C P, Yang Z. Enhancing search performance in unstructured P2P networks based onusers' common interest[J]. IEEE Transactions on Parallel and Distributed Systems,2008,19(6):821-836.
[45] Khan S, Tokarchuk L N. Interest-based self organization in group-structured P2P networks[C].Proceedings of the6th Consumer Communications and Networking Conference. IEEE,2009:1-5.
[46] Kumar A, Xu J, Zegura E W. Efficient and scalable query routing for unstructured peer-to-peernetworks[C]. Proceedings of the24th Annual Joint Conference of the IEEE Computer andCommunications Societies. IEEE,2005,2:1162-1173.
[47]张一鸣，卢锡城，郑倩冰等．一种面向大规模P2P系统的快速搜索算法[J]．软件学报，2008，19(6)：1473-1480．
[48] Harren M, Hellerstein J, Huebsch R, et al. Complex queries in DHT-based peer-to-peer networks[J].Peer-to-Peer Systems,2002:242-250.
[49]钱宁，吴国新，赵生慧．基于贝叶斯网络的无结构化P2P资源搜索方法[J]．计算机研究与发展，2009，46(6)：889-897．
[50]邱志欢，肖明忠，代亚非．一种P2P环境下基于用户行为的语义检索方案[J]．软件学报，2007，18(9)：2216-2225．
[51] Zhou F, Zhuang L, Zhao B, et al. Approximate object location and spam filtering on peer-to-peersystems[C]. Middleware2003. Springer Berlin/Heidelberg,2003:997-998.
[52] Cai M, Frank M, Chen J, et al. Maan: A multi-attribute addressable network for grid informationservices[J]. Journal of Grid Computing,2004,2(1):3-14.
[53] Bharambe A R, Agrawal M, Seshan S. Mercury: supporting scalable multi-attribute range queries[C].Proceedings of SIGCOMM’04,2004:353–366.
[54] Liu B, Lee W C, Lee D L. Supporting complex multi-dimensional queries in P2P systems[C].ICDCS’05,2005:155–164.
[55] Shu Y, Ooi B C, Tan K L, et al. Supporting multi-dimensional range queries in peer-to-peersystems[C]. P2P’05,2005:173–180.
[56] Banaei-Kashani F, Shahabi C. SWAM: a family of access methods for similarity-search in peer-to-peerdata networks[C]. Proceedings of CIKM’04,2004:304–313.
[57] Novak D, Zezula P. M-Chord: a scalable distributed similarity search structure[C]. Proceedings ofInfoScale’06,2006:19.104
[58] Bawa M, Condie T, Ganesan P. LSH forest: self-tuning indexes for similarity search[C]. Proceedingsof WWW’05,2005:651–660.
[59]聂晓文，卢显良，李梁等．DHT负载均衡的必要性[J]．计算机科学，2009，36(9)：92-95．
[60] King V, Saia J. Choosing a random peer[C]. Proceedings of the twenty-third annual ACM symposiumon Principles of distributed computing. ACM,2004:125-130.
[61] Raab M, Steger A.“Balls into Bins”—A Simple and Tight Analysis[J]. Randomization andApproximation Techniques in Computer Science,1998:159-170.
[62] Handurukande SB, Kermarrec AM, Le Fessant F, et a1. Peer sharing behavior in the eDonkey network,and implications for the design of server—less file sharing systems[J]. SIGOPS Oper. Syst. Rev.,2006,40(4):359-371.
[63] Breslau L, Cao P, Fan L, et a1. Web caching and Zipf-like distributions: Evidence and implications[C].Proceedings of the IEEE INFOCOM’99. Los Alamitos: IEEE Press,1999,126-134.
[64] Jung J, Sit E, Balakrishnan H, et a1. DNS performance and the effectiveness of caching[J].IEEE/ACM Transactions on Networking,2002,10(5):589-603.
[65] Karger D, Ruhl M. Simple efficient load balancing algorithms for peer-to-peer systems. In: VoelkermGM, Shenker S, eds. LNCS.Berlin: Springer-Verlag,2005.131-140.
[66] Karthik A R, Lakshminarayanan K, Surana S, Karp R, Stoica I. Load balancing in structured P2Psystems. In: Kaashoek MF, Stoica I, eds. LNCS. Berlin: Springer-Verlag,2003.68-79[7]
[67] Li ZN, Xie GG. A load balancing algorithm for DHT-Based P2P systems[J]. Journal of ComputerResearch and Development,2006,43(9):1579-1585.
[68] Xu Z, Bhuyan L. Effective load balancing in P2P systems[C]. Proceedings of the6th IEEE Int’l Symp.on Cluster Computing and the Grid. Washington: IEEE Computer Society,2006.81-88.
[69]宋宝燕，高楠等．DLRD：一种支持动态负载均衡的P2P网格资源发现机制[J]．通信学报，2008，29(8)：94-99．
[70]刘德辉等．QFMA：一种支持负载均衡的多属性资源定位方法[J]．计算机学报，2008，31(8)：1376-1382．
[71]熊伟，谢冬青等．一种结构化P2P协议中的自适应负载均衡方法[J]．软件学报，2009，20(3)：660-670．
[72]司加全，杨武．一种结构化P2P存储系统的负载均衡方法[J]．南京理工大学学报（自然科学版），2011，35(1)：38-41．
[73]廖小波．结构化对等网络中基于访问热点的负载均衡算法研究[D]．成都：电子科技大学，2009：
[74]弭伟，张春红等．基于结构化P2P系统的SDYA负载均衡算法[J]．北京邮电大学学报，2010，33(6)：116-121．
[75] Zhu Y, Hu Y. Efficient Proximity-aware load balancing for DHT-based P2P systems[J]. IEEE Trans onParallel and Distributed Systems,2005,16(4):349-36.
[76]付志鹏，王怀民等．对等网络的抖动特性研究综述[J]．计算机学报，2011，34(9)：1563-1579．
[77] Rhea S C, Geels D, Roscoe T, et al. Handling churn in a DHT[M]. Computer Science Division,University of California,2003.
[78] Godfrey P, Shenker S, Stoica I. Minimizing churn in distributed systems[M]. ACM,2006.
[79] Yao Z, Wang X, Leonard D, et al. Node isolation model and age-based neighbor selection inunstructured P2P networks[J]. IEEE/ACM Transactions on Networking,2009,17(1):144-157.
[80] Legtchenko S, Monnet S, Sens P, et al. Churn-resilient replication strategy for peer-to-peer distributedhash-tables[J]. Stabilization, Safety, and Security of Distributed Systems,2009:485-499.
[81] Tan G, Jarvis S A. Stochastic Analysis and Improvement of the Reliability of DHT-basedMulticast[C]. Proceedings of the26th IEEE International Conference on Computer Communications.IEEE,2007:2198-2206.
[82] Anwitaman D, Roman S, Karl A. Query-Load balancing in structured overlays[C]. Proceedings of the7th IEEE Int’l Symp. on Cluster Computing and the Grid.2007.http://lsirpeople.epfl.ch/rschmidt/papers/Datta2007QueryLoad.pdf.
[83] Zhu C, Liu Z, Zhang WM, Xiao WD, Yang DS, Xu ZN. Load balance based on rumor mongering instructured P2P networks[J]. Journal of China Institute of Communications,2004,25(4):31-40.
[84]肖卓程，荆金华．层次式Chord：物理拓扑感知的结构化对等网[J]．计算机科学，2006，33(7)：25-28．
[85] Ganesan P, Gummadi K, Garcia-Molina H. Canon in G major: designing DHTs with hierarchicalstructure[C]. Proceedings of the24th International Conference on Distributed Computing Systems.IEEE,2004:263-272.
[86] Joung Y J, Wang J C. Chord2: A two-layer Chord for reducing maintenance overhead viaheterogeneity[J]. Computer Networks,2007,51(3):712-73.
[87] Xu Z, Min R, Hu Y. HIERAS: a DHT based hierarchical P2P routing algorithm[C]. Proceedings of the2003International Conference on Parallel Processing. IEEE,2003:187-194.
[88] Martinez-Yelmo I, Cuevas R, Guerrero C, et al. Routing performance in a hierarchical dht-basedoverlay network[C]. Proceedings of the16th Euromicro Conference on Parallel, Distributed andNetwork-Based Processing. IEEE,2008:508-515.
[89] Garces-Erice L, Biersack E W, Ross K W, Felber P A,Urvoy-Keller G. Hierarchical peer-to-peersystems[J]. Parallel Processing Letters,2003,13(4):643-657.
[90] Artigas M S, Lopez P G, Ahullo J P, et al. Cyclone: A novel design schema for hierarchical dhts[C].Proceedings of the Fifth IEEE International Conference on Peer-to-Peer Computing. IEEE,2005:49-56.
[91] Zoels S, Despotovic Z, Kellerer W. Cost-based analysis of hierarchical dht design[C]. Proceedings ofthe Sixth IEEE International Conference on Peer-to-Peer Computing. IEEE,2006:233-239.
[92] Srebro J M. Content-based addressing in hierarchical distributed hash tables[D]. Rochester Institute ofTechnology,2009.
[93] Koskela T, Kassinen O, Ou Z, et al. Improving community management performance with two-levelhierarchical DHT overlays[J]. Journal of Internet Technology, Special Issue on Internet ResourceSharing and Discovery,2010,11(2):167-181.
[94] Fotiou N, Katsaros K V, Vasilakos X, et al. H-Pastry: An adaptive multi-level overlay internetwork[J].Athens University of Economics and Business, Tech. Rep,2011.
[95] Gupta R, Somani A K. Pricing strategy for incentivizing selfish nodes to share resources inpeer-to-peer (p2p) networks[C]. Proceedings of the12th IEEE International Conference on Networks.IEEE,2004,2:624-629.
[96] Friedman E J, Halpern J Y, Kash I. Efficiency and nash equilibria in a scrip system for p2pnetworks[C]. Proceedings of the7th ACM conference on Electronic commerce. ACM,2006:140-149.
[97]彭冬生，林闯，刘卫东．利用无私节点改善基于支付机制P2P应用的性能[J]．计算机学报，2008，31(6)：953-959．
[98] Ma R T B, Lee S, Lui J, et al. Incentive and service differentiation in P2P networks: a game theoreticapproach[J]. IEEE/ACM Transactions on Networking (TON),2006,14(5):978-991.
[99]黄保华．对等系统的安全与激励机制研究[D]．武汉：华中科技大学，2006：
[100] Parkes D C, Shneidman J. Distributed implementations of vickrey-clarke-groves mechanisms[C].Proceedings of the Third International Joint Conference on Autonomous Agents and MultiagentSystems-Volume1. IEEE Computer Society,2004:261-268.
[101] Feldman M, Chuang J. Overcoming free-riding behavior in peer-to-peer systems[J]. ACM SIGecomExchanges,2005,5(4):41-50.
[102] Montresor A, Jelasity M. PeerSim: A scalable P2P simulator[C]. Proceedings of the IEEE NinthInternational Conference on Peer-to-Peer Computing. IEEE,2009:99-100.
[103] Gian P J. PeerSim HOWTO: Build a New Protocol for the PeerSim1.0Simulator[J].2005-12-24).http://peersim. sourceforge. net/tutorial1/tutorial1. pdf.
[104]李勇，彭宇行，陈福接．大规模视频点播磁盘cache替换算法[J]．计算机研究与发展，2000，37(2)：207-212．
[105]牛新征，佘堃，秦科．移动P2P网络的协作缓存优化策略[J]．计算机研究与发展，2008，18(4)：656-665．
[106]沈磊，罗惠琼．基于用户行为特征的P2P代理缓存的研究[J]．计算机应用研究，2007，24(11)：106-108．
[107] Gupta A, Liskov B, Rodrigues R. Efficient routing for peer-to-peer overlays[C]. Proceedings of theFirst symp. on Networked Systems Design and Implementation (NSDI).2004:113-126.
[108] Chu J, Labonte K, Bissias G, et al. A trace-driven evaluation of chord[R]. Technical Report04-38,Dept. of Computer Science, Univ. of Massachusetts, Amherst, MA1003,2004.
[109]张浩．基于结构化对等网络的高性能计算平台关键技术研究[D]．武汉：华中科技大学，2009：
[110]杨冬，董平，张宏科．对等网络Churn问题评估模型与分析[J]．通信学报，2007，28(6)：39-47．
[111]李伟，徐正全，杨铸．利用虚拟节点的动态结构化P2P网络性能研究[J]．武汉大学学报（信息科学版），2011，36(12)：1504-1507．
[112]张宇祥，杨冬，张宏科．P2P网络中Churn问题研究[J]．软件学报，2009，20(5)：1362-1374．
[113] Li J, Stribling J, Morris R, et al. Bandwidth-Efficient management of DHT routing tables[C].Proceedings of the2nd Conference on Symp. on Networked Systems Design&Implementation106(NSDI2005). Berkeley: USENI Press,2005.99114.
[114] Herrera O, Znati T. Static resiliency vs. churn-resistance capability of DHT-protocols[C].Proceedings of the18th International Conference on Parallel and Distributed Computing Systems(PDCS2005). Phoenix: IASTED/ACTA Press,2005.141-147.
[115] Kersch P, Szabo R, Cheng L, Jean K, Galis A. Stochastic maintenance of overlays in structured P2Psystems[J]. Computer Communications,2008,31(3):603619.
[116] Wu D, Tian Y, Ng KW. Analytical study on improving DHT lookup performance under churn[C].Proceedings of the6th IEEE International Conf. on Peer-to-Peer Computing (P2P2006). Piscataway:IEEE Press,2006.249-258.
[117] Stutzbach D, Rejaie R. Improving lookup performance over a widely-deployed DHT[C].Proceedings of the IEEE INFOCOM.Washington: IEEE Press,2006.1-12.
[118] Saroiu S, Gummadi P K, Gribble S D. A measurement study of peer-to-peer File sharing systems[C].Proceedings of the Multimedia Computing and Networking. San Jose, USA,2002:156-17.
[119]黄敏，蔡志刚．缓存替换算法研究综述[J]．计算机科学，2006，33(12)：191-192．
[120]张宇祥，张宏科．一种层次结构化P2P网络中的负载均衡方法[J]．计算机学报，2010，33(9)：1580-1590．
[121]魏文红，向菲等．一种结构化P2P系统的负载平衡算法[J]．计算机科学，2010，37(4)：82-85．
[122]孙海燕，数据网格副本管理关键技术研究[D]．长沙：国防科学技术大学，2005：
[123] Zoels S, Despotovic Z, Kellerer W. On hierarchical DHT systems–An analytical approach foroptimal designs[J]. Computer Communications,2008,31(3):576-590.
[124] Zoels S, Despotovic Z, Kellerer W. Load balancing in a hierarchical DHT-based P2P system[C].Proceedings of the2007International Conference on Collaborative Computing: Networking,Applications and Worksharing. IEEE,2007:353-361.
[125]宾雪莲，杨玉海，金士尧．一种有限优先级的静态优先级分配算法[J]．软件学报，2004，15(6)：815-822．
[126] ButtazzoG C. Rate Monotonic vs．EDF：Judgment Day [J]．Real-Time Systems(S0922—6443)，2005，29(1)：5-26.
[127] Li W, Kavi K, Akl R. A non-preemptive scheduling algorithm for soft real-time systems[J].Computers&Electrical Engineering,2007,33(1):12-29.
[128]巴巍，张大波，李琦等．一种新的组优先级动态实时调度算法[J]．计算机科学，2009，36(7)：230-233．
[129] Khataniar G, Goswami D. Avoidance of churn rate through temporal centralization in Chord[J].Peer-to-Peer Networking and Applications,2011,4(3):251-258.
[130]王丹．P2P系统资源查询机制研究综述[J]．计算机科学，2004，31(9)：57-59．
[131]海沫．分布式系统中的资源发现机制综述[J]．计算机科学，2010，37(12)：12-17．
[132] Peng Z, Ting-lei H, Cai-xia L, et al. Research of P2P architecture based on cloud computing[C].Proceedings of the2010International Conference on Intelligent Computing and Integrated Systems(ICISS). IEEE,2010:652-655.
[133] Liang C C, Lin H F. Proactive Peer-to-Peer Architecture for Sharing Files on a Cloud Network[J].Advanced Science Letters,2011,4(11-12):11-13.