A Quantitative Justification to Dynamic Partial Replication of Web Contents through an Agent Architecture

The most usual solution to improve the performance of a Web server is based on building a distributed architecture, where the Web server is offered from a set of nodes. The most widely distributed architecture is based on Web clusters including a Web switch. The Web switch is responsible for deciding which site's node must attend which request. When deciding where elements are stored the classical solution was to fully replicate all contents in every server node. However, partial replication may require a fraction of storage while offering the same level of reliability. In this paper we report a solution based on dynamic partial replication where the number of replicas for each file and its management is handled by an agent architecture. We compare our solution with full replication and with static partial replication both in terms of storage capacity consumption and service time. Our results show that our proposed solution provides equivalent performance with a better use of disk storage capacity.

[1]  Roy T. Fielding,et al.  Hypertext Transfer Protocol - HTTP/1.1 , 1997, RFC.

[2]  Bruce A. Mah,et al.  An empirical model of HTTP network traffic , 1997, Proceedings of INFOCOM '97.

[3]  Paul Barford,et al.  Generating representative Web workloads for network and server performance evaluation , 1998, SIGMETRICS '98/PERFORMANCE '98.

[4]  Hyoung-Kee Choi,et al.  A behavioral model of Web traffic , 1999, Proceedings. Seventh International Conference on Network Protocols.

[5]  Byrav Ramamurthy,et al.  Scalable Web server clustering technologies , 2000, IEEE Netw..

[6]  Philip S. Yu,et al.  Geographic load balancing for scalable distributed Web systems , 2000, Proceedings 8th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (Cat. No.PR00728).

[7]  Philip S. Yu,et al.  The state of the art in locally distributed Web-server systems , 2002, CSUR.

[8]  Philip S. Yu,et al.  Request Redirection Algorithms for Distributed Web Systems , 2003, IEEE Trans. Parallel Distributed Syst..

[9]  David E. Singh,et al.  A Highly Available Cluster of Web Servers with Increased Storage Capacity , 2006 .

[10]  Oscar Sanjuán Martínez,et al.  Management of Distributed and Redundant Storage in High Demand Web Servers for Heterogeneous Networks Access by Agents , 2008, DCAI.

[11]  B. Cristina Pelayo García-Bustelo,et al.  A Multi-Agent based Proposal for the Management of Distributed and Redundant Storage , 2008, IC-AI.

[12]  E. Torres,et al.  An Architecture For Management Of Distributed And Redundant Web Storage With Ingelligent Agent Systems And Emerging Techniques , 2008, IEEE Latin America Transactions.

[13]  José Daniel,et al.  Propuestas arquitectónicas para servidores Web distribuidos con réplicas parciales , 2011 .

[14]  Jesús Carretero,et al.  A web cluster architectural proposal for balancing storage capacity and reliability by using partial replication , 2013, Comput. Syst. Sci. Eng..

[15]  Celia Gutiérrez An Analysis Architecture for Communications in Multi-agent Systems , 2013, Int. J. Interact. Multim. Artif. Intell..

[16]  Michael Kirsche,et al.  A New IEEE 802.15.4 Simulation Model for OMNeT++ / INET , 2014, ArXiv.