关键字 |
| MANET，文件共享，路由算法，文件可用性。 |
介绍 |
| MANET是一个广泛的网络。这里有不同的主机节点。它们像路由器一样移动，互相通信以传输数据。有两种manet，正常manet和断开的manet。前者在局部区域内具有相对密集的节点分布，而后者具有稀疏分布的节点，节点之间机会性地相遇。本地P2P模型提供了三个优点。首先，当没有基站可用时(如农村地区)，它可以实现文件共享。其次，通过P2P架构，可以避免当前基于客户端-服务器的文件共享系统中服务器过载的瓶颈。第三，它利用了移动节点之间原本被浪费的点对点通信机会。因此，在分布式MANET环境中，节点可以自由地、不显着地访问和共享文件，从而支持一些有趣的应用程序。 However the distinctive properties of MANETs, including node mobility, limited communication range and resource, have rendered many difficulties in realizing such a P2P file sharing system. File replication is an effective way to enhance file availability and reduce file querying delay. It creates replicas for a file to improve its probability of being encountered by requests. Unfortunately, it is impractical and inefficient to enable every node to hold the replicas of all files in the system considering limited node resources. Also, file querying delay is always a main concern in a file sharing system. Users often desire to receive their requested files quickly no matter whether the files are popular or unpopular. |
相关工作 |
| 为了最大限度地提高移动自组织网络的可用性，可以使用复制。如果使用的副本数量很少，文件共享就不能有效地进行。使用了不同的文件复制协议，但它们都存在一些问题，比如将有限的资源分配给不同的文件，其次是将存储作为副本的资源。为此提供的解决方案是全局最优的文件复制。康陈[2]使用了随机路径点模型和基于社区的模型两种模型。在RWP中，节点在选定的点上重复移动。所以遇到每个节点的概率是相似的。这里考虑的是随机获得的速度。在基于社区的流动性模型中，测试区域被划分为不同的子区域，称为洞穴。每个洞穴都有一个群落。 One node belongs to one or more communities. When node moves into its home community it has a probability Pin and when a node visits foreign community it has a probability 1- Pin .In case of optimal file replication, the meeting ability of a node as the average number of nodes it meets in a unit time and use it to investigate the optimal file replication. The probability of being encountered by other node is proportional to the meeting ability of the node. It indicates that files residing in nodes with higher meeting ability have higher availability than files in node with lower meeting ability. While creating the replica the memory is occupied. The probability of being met by others is decided by the nodes meeting ability so replica consumes both storage resource and meeting ability of the node. |
| 根据Yu-Chee-Tseng[3]， manet的特性可以是动态变化的拓扑结构，没有基站支持，和多跳通信能力。为了交流，他们使用跳跃的概念。当两个节点在无线电范围内时，它们之间使用单跳功能进行通信。这里讨论的问题是关于广播泛滥的问题。广播的问题是风暴问题。为了这个转播可以做，这是做及时的基础。广播的问题在于覆盖面较低、冗余、争用和冲突。本文考虑了这些问题，解决了广播问题，与泛洪相比提高了到达能力，降低了延迟。使用概率路由和文件发现协议[4]~[6]避免广播。它们将查询转发到满足目标的概率更高的节点。另一个需要考虑的问题是阈值。 Threshold is the constant defined which gives the fixed host density. In this paper dynamic solutions to those problems are given which includes adaptive counter-based, adaptive location based, and neighbor coverage schemes. In adaptive counter based scheme each individual has capability to change or adjust its threshold based on neighborhood status. In adaptive location based scheme a host choose its threshold based on its current value of neighbor for determining whether to broadcast or not. Neighbor coverage scheme uses the accurate neighborhood information. |
| 梁中银[7]在自组织网络中采用了协同缓存的概念。为了有效地访问数据，使用了不同的协作技术。MANET的问题在于基础设施。这样数据就像路由器一样从一个节点传输到另一个节点。当移动节点作为请求转发路由器时，可以节省带宽和功耗，减少延迟。在合作缓存中，缓存数据的共享和协调是在多个节点之间完成的。因此，通过使用合作缓存，web性能得到了提高。本文采用了CachePath、CacheData和HybridCache等方案。在CacheData中，受欢迎的项被本地缓存。中间节点缓存数据，然后为将来的请求提供这些数据。 For caching the data space is required. The problem with CacheData is that same data item can be cached at two or more nodes. Because of which there is wastage of large amount of cache. To avoid this problem, the rule used is that, a node does not cache data if all requests for the data are from same node. In CachePath intermediate node knows that which node has requested which data because the path of the requesting node and destination is saved in the cache. So when other node request for particular data item, the intermediate node calculate number of hops (distance) and then data item present on nearest node is served. Which means that it cache the data path. Because of which bandwidth and query delay can be reduced. For saving the path, there is no need to save all node information as the path from current router to the destination can be found by underlying routing algorithm. In Hybridcache cache path and cache data schemes are combined means that when a data tem needs to be cached it uses CacheData and path for that data item can also be cache. According to Huang et al. [9], WiFi-based wireless networks based on node mobility pattern, AP topology and file popularity, caching files in servers is done for realizing the optimal file availability to mobile users. However, the file servers considered are fixed nodes connecting to APs. |
| Pitkanen和Ott[10]提出了DTN存储模块，利用DTN存储-携带-转发范式，使DTN节点在转发所需的较长时间内保留消息的副本。在自组织网络中，移动主机可以自由移动。要考虑的事情是它们在网络分区的范围内和范围外。因此，来自一个网络的主机不能访问来自另一个网络的日期项。它降低了可访问性。在移动主机上复制不是原始数据所有者的数据项的解决方案。Hara[8]的考虑是为了提高数据的可访问性，每台主机的内存空间是有限的。然后通过考虑非周期性的数据更新来扩展这些数据，因为在一个真实的环境中假设是中尺度移动自组织网络。这里可以实现数据项的共享。在自组织网络中存在的主机数量访问其他主机持有的数据项作为原始数据，每个移动主机创建每个数据项的副本并在其内存空间中维护副本。 No central server is present to determine the allocated replicas but mobile hosts asynchronously determine the allocation in a distributed manner as we know mobile hosts moves freely so some characteristics approaches need to be considered one is replicas are relocated in specific period, during every replications period replicas allocate is determined based on the access frequency from each mobile host to each data item & consideration of network topology is optional. Three replica allocation methods considered here are Static access frequency, DAFN Method, DAG. SAF allocates the replica of data items based on its own access. Frequently used item is replicated at host. The replica creation is done only when a data access to the data item is successful or the mobile host connected to other mobile host which host replica or original data. The problem with SAF is that every host has a replica which creates the memory problem is solved using DAFN. In this method replica duplication is avoided or eliminated among the neighbors of mobile hosts. The change with this scheme is that when replica duplication created frequency to data item changes the replica to another replica. In DAG the replica sharing is done on the lager group of mobile hosts than DAFN. The need for this is that network or group should be stable. |
| Wei Gao[11]提出了NCL选择方案，基于概率选择度量创建，并协调多个缓存节点，以优化数据可访问性和缓存开销之间的权衡。 |
工作范围 |
| 工作范围可扩展为: |
| 1.本文综述了利用复制和面向分布式路由协议最大化移动自组织网络中P2P文件访问可用性的不同方法。 |
| 2.提出了新的框架和方法。 |
| 3.对所提出的算法进行实际仿真，并对其性能进行评价。 |
| 4.对现有算法和提出的算法进行比较分析，以证明算法的有效性 |
问题定义 |
| 虽然有不同的文件复制协议可用，但它们的主要问题是缺乏将有限的资源分配给不同的文件进行副本创建的规则，以实现最小的全局平均查询延迟，这是有限资源下的全局搜索效率优化。他们简单地将存储视为副本的资源，但忽略了一个节点与其他节点见面的频率也控制了其文件的可用性。会议能力越高的节点，文件的可用性越高。因此，问题在于如何将网络中有限的资源分配给不同的文件进行复制，从而使总体平均文件查询延迟最小化。 |
视角的解决方案 |
| 我们可以通过创建副本来改进文件共享，从而提高路由效率。对于复制，考虑节点存储和节点满足能力。资源分配的控制是在平均查询延迟和一个最优的文件复制规则上，该规则根据每个文件的流行度和大小分配资源。因此，这里的文件复制协议基于规则，该规则以完全分布式的方式近似于最小全局查询延迟。可以提出两种节点移动模型，即普通manet的随机路径点模型和无连接manet的基于社区的移动模型，以实现有效的资源分配。这样可以提高P2P系统的有效性。 |
讨论及未来工作 |
| 以上所讨论的方法有助于提高文件的可用性，如使用普通MANET的随机路径点复制技术和分布式MANET的基于社区的迁移模型复制技术。当我们使用这两种模型时，效率提高了，节能了。我们可以扩展工作，从理论上分析更复杂的环境，包括文件动态和动态节点查询模式。 |
确认 |
| 衷心感谢审稿人对本文的审阅，为大大提高论文质量提供了投入。 |
参考文献 |
- AshwiniS。Jagtap，“提高P2PFile共享的路由效率”，《国际计算机科学新兴趋势与技术杂志》，2014年第2卷，第11期。
- 陈康，“通过复制实现高效文件共享，最大化移动Ad hoc网络中的P2P文件访问可用性”，2014。
- Y. Tseng, S. Ni，和E. Shih，“在无线多跳移动自组织网络中缓解广播风暴的自适应方法”，发表于《OfICDCS程序》，2001年．
- B. Chiara, C. Marco等人，“Hibop:机会网络的基于历史的路由协议”，2007年。
- A. Lindgren, A. Doria和O. Schelen，“间歇性连接网络中的概率路由”，2003。
- F. Li和J. Wu，“MOPS:在容忍中断的网络中提供基于内容的服务”，ICDCS, 2009。
- 殷殷杰，曹国强，“支持自组织网络中的合作缓存”，计算机工程，第5卷，第1期。1, 2006。
- T. Hara和S. K. Madria，“在ad hoc网络中提高数据可访问性的数据复制”，TMC, 2006。
- 黄毅，高毅，等，“在容忍延迟内容分发社区中优化文件检索”，ICDCS, 2009。
- M. J. Pitkanen和J. Ott，“移动dtn中的冗余和分布式缓存”，载于MobiArch的Proc.， 2007。
- W. Gao, G. Cao, A. Iyengar和M. Srivatsa，“支持容错网络中的合作缓存。ICDCS, 2011。
|
菜单