Abstract | ||
---|---|---|
This paper presents a server performance model based on the approach of probabilistic analysis. This model is built on the experimental results using the X-Code in a network environment with stable overhead. The model effectively optimizes the mean service time for client data request by determining the optimal number of servers which store the original data when the total number of the servers are given and choosing the optimal read scheme, therefore, it improves the overall throughput and performance of distributed storage system. We note that this model is built without any a priori knowledge of the traffic information. The results indicate that in a distributed storage system, the method can play an active role, as proper data redundancy can help to improve not only the reliability of a storage system, but also the efficiency of it. |
Year | Venue | Keywords |
---|---|---|
2006 | PDPTA | distributed storage system,mds code,performance,redundancy.,storage system,fault tolerant system,a priori knowledge,probabilistic analysis,redundancy |
Field | DocType | Citations |
Computer science,Fault tolerance,Distributed computing | Conference | 0 |
PageRank | References | Authors |
0.34 | 8 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jianbo Fan | 1 | 8 | 4.06 |
Lidan Shou | 2 | 370 | 48.66 |
qingfeng li | 3 | 0 | 0.34 |
Jinxiang Dong | 4 | 311 | 65.36 |
ningbo zhejiang | 5 | 0 | 0.34 |