Object clustering techniques for client/server and distributed software development

While object-oriented paradigm is known to be effective for developing web-based software, most of the current object-oriented methodologies including OMT and UML do not provide effective methods and instructions for modeling client/server or distributed applications. We present new effective object...

Full description

Saved in:
Bibliographic Details
Main Authors: Kim, Soo Dong, Rhew, Sung Yul, Kim, Chul Jin, Kim, Dong Kwan
Format: Conference Proceeding
Language:English
Subjects:
Application software
Buildings
Computer architecture
Internet
IP networks
Network servers
Object oriented modeling
Programming
Unified modeling language
Web server
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:While object-oriented paradigm is known to be effective for developing web-based software, most of the current object-oriented methodologies including OMT and UML do not provide effective methods and instructions for modeling client/server or distributed applications. We present new effective object clustering formalism and algorithms for client/server and distributed application development. The coupling degree among objects is often the key factor in grouping related objects into a cluster. If objects with large coupling degree are identified and clustered together, the cohesion in the cluster is increased and so network traffics among the clusters are largely reduced, yielding a high overall system performance. The paper presents the metrics for measuring object coupling and shows systematic algorithms for object clustering based on the proposed metrics. In addition, a case study is given to assess the proposed formalism and methods.
ISSN:1062-922X
2577-1655
DOI:10.1109/ICSMC.1998.725064