Analysis and comparison of algorithms in advanced web clusters solutions

Abstract

Today's websites (applications) represent an essential part of nearly every business system, therefore it is unacceptable for them to be unavailable due to the ever-increasing competition on the global market. Consequently, such systems are becoming more and more complex to allow for their high availability. To achieve a higher system availability, a greater scalability is used in creating the so-called Web farms or Web clusters. This system requires much more computer power than the traditional solutions. Since such systems are very expensive and complex in nature, the question is how to obtain the best possible results with the least amount of investment. To achieve that, it is necessary to take a look at the component which contains the information on the request or traffic amount, and that is the HTTP/HTTPS load balancer. The system is based on several algorithms, however, there are no comprehensive analyses that indicate which algorithm to use, depending on the Web cluster and the expected amount of traffic. For this reason, this paper provides a detailed comparison of several frequently used algorithms in several different Web cluster scenarios, i.e. loads. Also, examples are given as to when to use a certain algorithm.