A self-adaptation, high availability cluster file system for network applications
Qiang Cao, Chang Sheng Xie · 2002
In this paper, we present a self-adaptive and high availability cluster file system in which we employ lots of commodity machines and their local operation system to build an integrative file system that can provide high availability, performance, and self-adaptation by automatically reduplicating data and efficiently using them, which supports logging and recovery from node failures. At that time, the data in the file is automatically replicated or migrated on demand to a number of requests from customers who use the data through the network, in order to decrease the workload and single node fault. The file system is satisfactory for the requirements of network applications, such as VoD, WWW, etc.