Implementing a Linux Virtual Server [30 mn de lecture - paru le 9/19/2005 11:38:59 AM - Public : Expert]

Auteur

Vincent BEDNAREK Elève-Ingénieur Supinfo Paris Promotion SUPINFO 2005    Lui écrire    Tous les projets de cet auteur    Le mini-CV de cet auteur

Introduction

Some Internet sites have today significant requirements in term of network resources to answer efficiently to client requests.

To answer this criterion of effectiveness, various solutions was used: setting in cluster, increasing band-width... Among all these possibilities, one appeared like the most effective compromise, the load balancing solution.

It consists in group with several servers (we will name them "real servers") placed behind another principal servers (the load to balancer). This last receives the requests from the clients and transmits them to real servers according to a defined order. The load balancer thus constitutes the intermediary between the Net surfer and real servers. This is completely invisible for the Net surfer who thinks to be directly connected to a single server.

In some cases, and to still improve the availability, we will see that it is possible to couple 2 load balancers, in cluster, so that they can still ensure service if one of them breakdown.

Although those methods are not always very known, they however are very much used among Internet operators, in particular for the Web servers, Web site content is thus diffused more quickly.

Moreover, this solution offers a considerable fault tolerance, in addition to performance advantages.

LVS is not the only one to propose this type of functionalities. Apart from the free systems, there are also material equipments specifically dedicated and optimised for this type of task. Nevertheless, we will not approach the owner systems in this document.

This article will present you:

• Different types of load balancing that proposes LVS.
• Different algorithms that proposes LVS.
• Installation and configuration of a LVS system.