Earticle

초록

영어

This paper presents, in the form of a case study, the design, implementation and performance evaluation of Phoenix: a prototype of a fault-tolerant distributed storage web service based on URLs. Our system supports a set of redundancy strategies including simple file-replication, error-coding techniques and a hybrid adaptive technique. The web storage middleware (WSM), which is at the core of the prototype, reuses the URLs of the files to dynamically accommodate data and redundancy load.
Experimental assessments show that distributing the users to several web servers compensates for the overhead produced both by generating and distributing data redundancy. We also found that during a non-critical number of web server failures, users do not perceive but a lengthening on the system response times. As expected, peak user requests and network congestion are the most important factors affecting the storage system’s performance.

목차

Abstract
 I. INTRODUCTION
 II. RELATED WORK
 III. PHOENIX DESIGN PRINCIPLES
  A. WSM: A Web Storage Middleware
  B. Redundancy Strategies
  C. Redundancy based on Simple Replicas
  D. Redundancy based on IDA
 IV. DESIGN OF EXPERIMENTS
  A. Studied Configurations
  B. Studied Measurements
 V. RESULTS
  A. Upload Operations
  B. Download Operations
  C. Degraded Mode and Redundancy Costs
 VI. CONCLUSIONS
 VII. FUTURE WORK
 ACKNOWLEDGMENT
 REFERENCES

키워드

저자

• J. L. Gonzalez [ Dept. Systems and Computation Cd. Valles Institute of Technology Carretera al Ingenio Km2, Cd. Valles, S.L.P. Mexico ]
• Ricardo Marcelín-Jiménez [ Electrical Eng. Dept. UAM-Iztapalapa Atlixco 186, 09340 Iztapalapa, Mexico City ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  JoC [Journal of Convergence]

• 간기

  계간

• pISSN

  2093-7741

• eISSN

  2093-775X

• 수록기간

  2010~2015

• 십진분류

  KDC 004 DDC 004