Research On Automatic Fault Handling Technology In Cloud Storage System

This paper demonstrated an extensible and mass cloud storage system named HugeCloud based on the current research of fault handling in cloud storage system and automatic management in storage system. Automatic fault handling technology of HugeCloud was researched in this article. Main works are described below:1) The author proposed an extensible and mass cloud storage system called HugeCloud, designed the architecture of HugeCloud. The metadata management mechanism, dynamic load balancing technology of management server and data faults tolerance method were researched based on the architecture of HugeCloud.2) A policy based fault management framework for cloud storage system named PCFMF was brought forward based on policy management technology. A policy define language for fault handling in HugeCloud was designed named HPDL, which prepares the ground for automatic fault handling in HugeCloud and simplifies the implementation. The implementation of policy mapping mechanism and the design of policy repository are put forward at last.3) Fault detection and fault recovery mechanism were researched and a highly reliable fault detection framework was presented in this paper. Based on the framework, automatic fault recovery algorithms for various nodes such as storage management nodes, storage nodes and storage agent nodes in HugeCloud were put forward to guarantee the serviceability of HugeCloud in the fault circumstances.