Scalable, Extensible, Distributed Query Centric Data Server (DStorm)
|
||||
Motivation Features of DStorm
|
Features of DStorm Scalability: DStorm is scalable -- new PCs can be added in to the cluster dynamically, and workload and data will be distributed based on the configuration accordingly. The system is query centric as it balances the load and distributes the data based on query access patterns and current system loads. Extensibility: DStorm supports dynamic addition of new access methods, indexes, and buffer management strategies into the system. Indexes such as the R-tree, Extensible hashing and B+-tree are however implemented as part of the system.
PCs in offices are typically under utilized as they are mainly used
for simple word processing and emailing, and as dumb terminals.
DStorm could be used to set up a virtual PC cluster
out of these Office PCs and
to utilize "spare" resources for
storage and manipulation of shared and private documents/data. It could
be used as a distributed backend for Electronic Document Registry, Document Archival
System, and Distributed Database System.
DStorm is implemented in JAVA, and hence is platform
independent.
Should you wish to find out more about DStorm, please contact Project Members: Ooi Beng Chin, Stephen Bressen, Lee Mong Lee, Tan Kian Lee, Jenson Goh, Lan Bin, Anirban, Ng Wee Siong
|
|