¡¡Chinese Journal of Computers   Full Text
  TitleQFMA: An Approach for Multi-Attribute Resource Addressing with Load Balancing
  AuthorsLIU De-Hui1) ZHOU Ning2) YIN Gang1) WANG Huai-Min1) ZOU Peng1)
  Address1)(School of Computer Science, National University of Defense Technology, Changsha 410073)
2)(China Institute of Electronic System Engineering, Beijing 100141)
  Year2008
  IssueNo.8(1376¡ª1382)
  Abstract &
  Background 		Abstract It is an effective way to make SOA uncentralized with P2P. In meta-data based P2P system, hot spot may result due to non-uniform of keyword¡¯s distribution and load, which will influence the usability of the system seriously. This paper proposes QFMA, an algorithm for load balancing in meta-data based P2P systems based on MAAN. The main idea is that, record the status of the high load node along the query path, the other query will change their destination based on the status recorded. The analysis and simulation result show that QFMA can reduce the load of hot spot and achieve load balance, the query path length is also O(logN).
Keywords SOC; SOA; P2P; QFMA; load balance; hot spot
Background This work belongs to P2P data management, and can be applied to Service Oriented Computing. Load balancing is one of the most important mechanisms in structured P2P system. There are two main kinds of methods to achieve load balancing: duplicate and caching, virtual servers. However, the approaches proposed do not work well in a special kind of P2P systems named meta-data based P2P systems, because the data to be dealt with are just meta-data of resources. To overcome the shortcomings, the authors make use of the status feed back by busy nodes and give a novel method based on MAAN. The experiment results show that it can achieve load balancing in meta-data based P2P systems, with little route path length sacrificed, and will not bring invalid meta-data to the network, and will not increase the difficulties of duplicate management.
This work is supported in part by the National Natural Science Foundation of China under grant No.90412011, and National Basic Research Program of China (973 Program) under grant No. 2005CB321804.