| Abstract |
Peer-to-peer systems have recently become a popular means for transferring large amounts of data from one peer to another. Peer-to-peer systems are distributed systems in which nodes can share resources, information and services, while they continue to remain completely independent. Despite the fact that the peer-to-peer systems have a great variety of applications, there are still application areas which have not been developed. One of these has to do with peer-to-peer networks and databases technologies. The main idea is the support and sharing of relational data that can be found in the databases in a peer-to-peer environment. Due to the dynamic nature of each node, neither the consideration of a global schema, nor the control of data placement is feasible in every node of the network. Besides, the sending of an SQL query to all nodes of the network is a non-practical and scalable action as the nodes will not be able to interpret the query due to the heterogeneity of the information. Moreover, there will be an overhead in network traffic as a consequence of the number of messages sent. Several techniques have been proposed for the creation of mappings among the schemas, in order to serve and route a query properly. In the present study, we present the design and implementation of the RDPeer application for the management and sharing of relational data, without the existence of a global schema or the creation of communities from peers based on the semantic content of the data. RDPeer is a hybrid peer-to-peer system, which allows every peer to participate in more than one communities simultaneously, to communicate with a database, to pose SQL queries and receive answers from the peers that have data related to the queries. RDPeer tries to resolve the problem of query routing by the dynamic creation of super-peers through a system of elections. The super-peers undertake to collect the schemas of relational databases and to process the queries. Processing of queries is accomplished in two phases. The first phase includes the creation of mappings among the SQL queries and the schemas of peers during querying. The second one transforms the queries based on the mappings. The transformation procedure results in the creation of SQL queries for the peers who contain relevant data and enables each peer to extract the data in a peer-to-peer way.
|
Διαμοιρασμός και Διαχείριση Σχεσιακών Δεδομένων σε Ομότιμα Δίκτυα
