Efficient way of Data Managing for Range Queries in Unstructured Peer to Peer Networks. P Lalitha Kumari Dept. of CSE ASR College Of Engg, Tanuku,Andhrapradesh,India. P Rama Rao Dept. of MCA Sri Sivani College of Engg, Andhrapradesh,India. Chinnam YuvaRaju Dept. of CSE ASR College Of Engg Tanuku,Andhrapradesh,India Abstract— Peer-To-Peer (P2P) networks have become very popular in the last few years. Nowadays, they are the most widespread approach for exchanging data among large communities of users in the file sharing context. Efficient way of managing storage and retrieval of multidimensional data is achieved by proposed framework which ensures robust query evolution. This framework is based on peer to peer network, where large collection of Data to be stored. This data is divided into subparts and built up an index on set of each compressed data and this data is to be distributed across p2p network .This compressed data supports efficient data extraction of information .A replication mechanism provides appropriate coverage of index and metadata by considering network conditions and query workload . Keywords- multidimensional data, indexing, compression and p2p network I. INTRODUCTION EER-TO-PEER (P2P) networks have become very popular in the last few years. Nowadays, they are the most widespread approach for exchanging data among large communities of users in the file sharing context. In order to make participants really autonomous, they should be imposed no constraint on storage and [1] computational resources to be shared, as well as on the reliability of their network connection. These requirements make traditional distributed frameworks unsuitable and suggest the adoption of a solution based on an unstructured P2P network, where peers are neither responsible of coordination tasks (such as super peers, which are called for a certain amount of resources and reliability), nor imposed to host specific pieces of data (as in DHT-based networks). Our aim is devising a P2P-based framework supporting the analysis of multidimensional historical data. Specifically, our efforts will be devoted to combining the amenities of P2P networks and data compression to provide a support for the evaluation of range queries, possibly trading off efficiency with accuracy of answers. [2] The framework should enable members of an organization to cooperate by sharing their resources (both storage and computational) to host (compressed) data and perform aggregate queries on them, while preserving their autonomy. The management of compressed data on unstructured P2P Networks is an intriguing issue, but poses several research Challenges, which we are discuss in the following. A. Compression A compression technique must be devised which is able to create “prone-to-be-distributed” data synopses supporting the efficient evaluation of aggregates, possibly affected by tolerable error rates [3]. However, in this case, although the cost of disk storage is continuously and rapidly decreasing, it may still be difficult to find peers for which hosting replicas of synopses has a negligible cost, while autonomy is a requirement in our setting? Using traditional compression techniques, synopses providing reasonable error rates may have a non-negligible size (usually not under 1 percent of the size of the original data set, e.g., 1’’ GB from a 1’ TB data set). Although compressing the data certainly makes replication less resource consuming, [4]replicating the entire synopsis each time would require storage and network resources that could be saved if only some specific portion of the synopsis could be replicated [2]. We recall that replication is mandatory in the P2P setting, both to contrast the volatility of peers (which threatens data availability) and to prevent peers from being overloaded (in the presence of many users interested in a data set, if the peers hosting these data were too few, they would be required to process a large amount of queries). These drawbacks would be overcome if the compressed synopsis were subdivided into tiny sub synopses which are Independently replicated and disseminated on the network when needed. Peers would, therefore, be asked to host replicas of small chunks of data. This way, the autonomy requirement would not result in a limit on the overall size of the synopsis B. Indexing A better way to address this issue is to design an indexing mechanism that supports the efficient location of the sub synopses involved in the query evaluation. In the literature, there are several works proposing distributed indexing techniques, where indexes are variants of R-Trees which are partitioned and distributed among the nodes of the network. [1] According to these approaches, nodes of the networks are assigned groups of nodes of the R-tree, and maintain references to hosts which are assigned other nodes of the R-tree. The association between hosts and R- tree nodes is fixed and the maintenance of the index is centralized. These solutions, as they are, were devised for relatively static scenarios, and they are not suitable for the dynamic scenario addressed by our proposal, where in order to guarantee peer autonomy, peers cannot be P P. Lalitha Kumari et al, / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 2 (5) , 2011, 2078-2083 2078