A Novel Approach to the Selection of Representative Skyline Web Services

Nowadays, there are more and more services with the same functionality and different Quality of Web Services (QoWS) published on the web. For QoWS-aware service composition, the large number of candidate services will result in the increase of the process time for composing services under user const...

Full description

Saved in:
Bibliographic Details
Main Authors: Gang, Dai, Chunli, Cai
Format: Conference Proceeding
Language:English
Subjects:
Communication networks
Composing
Computational intelligence
Conferences
Decision making
Decomposition
Mathematical models
Measurement
Optimization
QoWS
Quality of service
Representative skyline
Tradeoffs
Web services
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Nowadays, there are more and more services with the same functionality and different Quality of Web Services (QoWS) published on the web. For QoWS-aware service composition, the large number of candidate services will result in the increase of the process time for composing services under user constraints. In order to prevent to traverse all candidate services, a better approach is to offer a concise summary of candidate services that can represent different trade-off of QoWS, called representative web services. The representative web services can significantly reduce the number of candidate web services and offer potentially optimal set of services. Furthermore, it also reduces the processing time. In this paper we propose a solution that combines layer decomposition and least dominating representative's concept (LDR) to find the representative web services. The proposed solution consists of two steps: first, we use the layer decomposition to capture the different trade-off of QoWS and categorize all candidates into different sets. Second, we use the LDR to find the best web services which dominate all other services with the least increment. The result of experimental evaluation indicates that our approach can capture the trade-off of different QoWS, and can be executed much faster.
ISSN:2472-7555
DOI:10.1109/CICN.2015.201