Data Management in the APPA System

Abstract

International audienceCombining Grid and P2P technologies can be exploited to provide high-level data sharing in large-scale distributed environments. However, this combination must deal with two hard problems: the scale of the network and the dynamic behavior of the nodes. In this paper, we present our solution in APPA (Atlas Peer-to-Peer Architecture), a data management system with high-level services for building large-scale distributed applications. We focus on data availability and data discovery which are two main requirements for implementing large-scale Grids. We have validated APPA's services through a combination of experimentation over Grid5000, which is a very large Grid experimental platform, and simulation using SimJava. The results show very good performance in terms of communication cost and response time

Resource Discovery Considering Semantic Properties in Data Grid Environments

“…Many research works have used the P2P techniques for resource discovery in largescale environment [1], [2], [4], [8], [15], [23], [27] and [28]. But, only few studies have focused on the semantic aspect during this discovery.…”

“…However, all these solutions imposed their topologies. Other studies are based on a global ontology or schema shared by all peers [2], [4], [7] and [20]. APPA [4] is based on a common schema to distribute and share data between peers considering semantics.…”

“…Several other researches have emerged. We cite those based on a super peer topology [4], [7], [8], [27] and [30]. But, they require a certain mapping topology as in [3].…”

Analytical Model for Mobile P2P Data Management Systems

“…The potential advantages of P2P systems are the node autonomy, scalability, high availability through replication and performance through parallelism [4]. There are already some experimental systems that utilize some data management algorithm, but there are no commercial third generation data management solutions on the markets yet.…”

“…There are already some experimental systems that utilize some data management algorithm, but there are no commercial third generation data management solutions on the markets yet. Some example P2P data management systems include APPA [4], PeerDB [5], GridVine [6] and AmbientDB [7].…”