Resource Availability Optimization for Priority Classes in a Website

Abstract

High resource availability in computer networks has become a critical issue as it can be affected by the increasing number of network users. A methodology based on classifying website visitors into priority groups is proposed, assuring high availability for priority classes, based on reserved website resources that can be accessed only by these groups and simultaneously providing as many resources as possible to lower priority visitors. A birth-death process is proposed to model website visitors' arrival and service. An optimization problem is solved to determine the optimal trade off between resource availability for high priority visitors and free resource access to lower priority visitors. The major contribution of this paper consists in deriving formulas for the probability that a website visitor has no further access to resources and in determining the optimal reserved resources assuring the above trade off.

Optimal server resource reservation policies for priority classes of users under cyclic non-homogeneous markov modeling

“…Consequently, a way of balancing the needs of each class in resources and the priorities among users' classes has to be distinguished. The idea of reserving an amount of resources for each class was studied in [12,15] for two different cases and can be described by guarding a certain level of resource for the priority classes. Assume that the level of server resources that are reserved for class C is denoted by g 1 , for class B by g 2 and finally for class A is denoted by g 3 [12].…”

“…Another kind of client-server system that has been under extensive study is a computer system or network in which there are priorities among the users or client [3,5,15,19,29]. In such a system not only the needs of availability are increased in order to provide a satisfactory level of service but furthermore the high priority users have to be served in a more efficient way than others.…”

“…Based on the presented works for resource reservation on a website [15] and guarded cells in cellular networks [12], we propose to reserve resources that can be accessed only by higher priority classes. Nevertheless, reserving resources for high priority classes, in order to provide them with higher levels of resource availability, will eventually cause a denial of service for lower priority classes.…”

Software rejuvenation and resource reservation policies for optimizing server resource availability using cyclic nonhomogeneous Markov chains

“…In general, a client–server system should provide high levels of resource availability. A special case of client–server system is when priorities are assigned among users/clients . Priorities are assigned either because of a membership fee paid or not by the clients , or because of the status of the users, that is, in a university server .…”

“…In , a client–server system is studied, in which users are categorized into different classes with respect to priorities that have been assigned. To provide higher levels of resource availability and offer access to users/clients, the technique of guarding resources has been adopted according to the work presented in and . The work in focuses on modeling the resource reservation procedure for four different priority classes to provide higher levels of server resource availability.…”