Asian Research Thesis Index

Abstract

Due to demand for the modern real time data warehousing, stream based join
algorithms have become need and are getting significant importance. The basic
problem in real time data warehousing is the joining of the fast stream input of source
level updates with the disk based relation. A number of stream based algorithms have
been proposed in literature to compensate the disk access cost over the fast incoming
stream of data with huge volume. Reduced Mesh Join (RMESHJOIN) is one of the
prominent approaches in this regard. Majority of work has been done on this
algorithm to enhance its performance using memory management, but still work
needs to be done to reduce I/O cost. In this research, an Optimized Reduced Mesh
Join (OR-MESHJOIN), an optimization of R-MESHJOIN, will be proposed. OR-
MESHJOIN algorithm will overcome the issues related to I/O cost thus resulting in
significant increase in overall performance. The new version will enhance the
efficiency of the R-MESHJOIN by parallel retrieval of data from disk based relation
R into main memory. This will be a significant feature of the proposed algorithm that
will thus enhance the performance by many folds. Finally, experimental study will be
carried out to justify proposed algorithm in terms of efficiency and accuracy.

Add/Update Thesis