Scalable, Extensible, Distributed Query Centric Data Server (DStorm)
|
||||
Motivation
|
The last two decades have witnessed a growing need for information sharing. The Internet, Web, and data warehouses increased dramatically the use of databases because of many challenging applications. The large data volume and fast response needed for complex queries make it critical to apply parallel technologies. Given the technological advances, collections of computers including Clusters of Servers and Network of Workstations are fast becoming the primary computing infrastructure for high performance. Such systems are attractive because they are built from affordable commodity hardware such as Pentium based Personal Computers (PCs) and fast LAN such as Myrinet or ATM. Furthermore, they use standard software components such as UNIX, PVM (Parallel Virtual Machines) and the MPI parallel programming (Portable Parallel Programming with Message-Passing Interface) environments. DStorm is a scalable, extensible distributed query centric data server which aims to provide a low cost high performance data server that runs on any low cost PCs or workstations. The other main objective of DStorm is to exploit spare capacity of hardware resources of Office PCs for storage and manipulation of data without affecting ad hoc tasks running on a typical PC.
|
|