원문정보
초록
영어
Cloud computing is becoming more and more important in research organization and IT industry. It has various advantages like scalability and availability of computing resources. Virtual cluster systems are also used in scientific computing environment. The system like vcluster, integrates computing resources from heterogeneous clouds. It maintains the information of all clouds which are involved in a federated environment. The virtual machine allocation is based on the available resources and information of utilized resources. In order to do that vcluster has to request the clouds to get the real time information of virtual machines which include huge synchronization overhead. In this paper, we propose a methodology to reduce overhead in synchronization mechanism. An analysis shows that the proposed mechanism minimizes the overhead of synchronization. It also has ability to discover, correct and report the errors.
목차
1. Introduction
1.1. Vcluster
1.2. HTCondor
2. Motivation
3. Proposed Method
3.1. One-to-One Correspondence
3.2. One-to-Many Correspondence
3.3. The Algorithm
4. Experiments
5. Conclusion and Future Work
References