Browse Prior Art Database

A Buddy-based Load Balancing Algorithm for Multi-server File Systems Improving Server Buffer Hit Probability

IP.com Disclosure Number: IPCOM000105198D
Original Publication Date: 1993-Jun-01
Included in the Prior Art Database: 2005-Mar-19
Document File: 1 page(s) / 51K

Publishing Venue

IBM

Related People

Dan, A: AUTHOR [+3]

Abstract

Disclosed is an approach to load balancing in a multi-server environment which allows the optimization of desired performance criteria such as response time by balancing the load while maintaining good server buffer hit probabilities. This is achieved by controlling replication of filesets in the buffers of multiple servers. The method is client-based and takes into account the server load.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 57% of the total text.

A Buddy-based Load Balancing Algorithm for Multi-server File Systems Improving Server Buffer Hit Probability

      Disclosed is an approach to load balancing in a multi-server
environment which allows the optimization of desired performance
criteria such as response time by balancing the load while
maintaining good server buffer hit probabilities.  This is achieved
by controlling replication of filesets in the buffers of multiple
servers.  The method is client-based and takes into account the
server load.

      Dynamic load balancing across file servers is required in a
multi-server file system environment to avoid the problem of server
bottleneck.  However, in general, balancing the server load in such
an environment across all the servers also results in reduction of
server buffer hit probability as the pages of a file may be
replicated in the buffers of different servers [*].  Therefore, an
intelligent load balancing algorithm should maintain good server
buffer hit probability while distributing the load so as to optimize
desired performance criteria such as response time or throughput.

      A good server buffer hit ratio can be achieved by an algorithm
that controls the replication of the pages of a file in multiple
server buffers.  In most file systems, files are grouped into
filesets for the purpose of administrative control and maintenance.
In our algorithm, each fileset is assigned a primary server.  To
maintain good server buffer hit ratio the...