Paper Info
Title
Who is more reliable?: an evaluation method for distributed-memory aggregation in the internet
Abstract
Traditional disk-oriented storage systems can no longer meet the high requirements of applications in latency and bandwidth, for which RAMCloud, was proposed to replace the disk with RAM. We consider using the memory of end-users in existing Internet to achieve a similar memory storage system. However the end-users' behavior are not as controllable as the servers in data center, which would lead to a high data loss rate. So how to ensure the data's reliability is the biggest challenge. To improve the reliability, as well as the resource utilization, we present a thought of Contributor scoring, establishing a scoring model according to Contributors' behavior. This model will classify Contributors into certain levels. Moreover, we train the Model with genetic algorithms. Experimental results show that the optimized model can effectively divide the Contributor, ensure certain reliability.It has great significance to this user memory storage system for improving the reliability and resource utilization.
Year
DOI
Venue
2012
10.1145/2413247.2413273
StudentWorkshop@CoNEXT
Keywords
Field
DocType
user memory storage system,traditional disk-oriented storage system,evaluation method,distributed-memory aggregation,scoring model,data center,certain reliability,resource utilization,contributor scoring,similar memory storage system,high data loss rate,optimized model,genetic algorithm
Data loss,Computer data storage,Server,Distributed memory,Bandwidth (signal processing),Engineering,Data center,Genetic algorithm,Reliability engineering,The Internet,Distributed computing
Conference
Citations 
PageRank 
References 
1
0.37
2
Authors
5
Name
Order
Citations
PageRank
Hongyi Chen131.09
Zhi-gang Sun24114.15
Bin Huang310.37
Tao Li4387.33
Xilong Mao562.24