Metamodeling as an approach for better computer resources allocation in web clusters

摘要

Constant changes are inherent to information technology, the proof of which can be found in many challenges, such as the business sector's cloud computing. Because of the recent economic and financial crisis companies are forced to provide as best results as possible with the least amount of costs. The main issue of cloud computing are the computer resources (i.e. compute power). Due to their less than optimal usage, computer resources generate high-level costs. The term cloud computing encompasses a wide range of technologies, therefore this paper will focus only on Web clusters. It will describe the issue of improper use of resources in these systems and its main cause, as well as provide suggestions for further optimization. Additionally, the paper will present a new concept of HTTP / HTTPS traffic analysis which should enable a more efficient usage of computing resources. The concept has emerged from metamodeling of two methods. The first one is method for distribution of HTTP / HTTPS requests, and the other one is the Analytic Hierarchy Process (AHP) method which is used for classification and prioritization of requirements so that the entire Web cluster could operate as best as possible with the least amount of resources.