PuLSaR: preference-based cloud service selection for cloud service brokers

Ioannis Patiniotakis,Yiannis Verginadis,Gregoris Mentzas

doi:10.1186/s13174-015-0042-4

Abstract

Over the last few years, the vast increase of cloud service offerings that are available from heterogeneous cloud vendors, has made the evaluation and selection of desired cloud services, a cumbersome task for service consumers. In that respect, there is an increasing need for user guidance and intermediation during the service selection process but also during the cloud service consumption that should always refer to the best possible choice based on user preferences. In this paper, we discuss the Preference-based cLoud Service Recommender (PuLSaR) that uses a holistic multi-criteria decision making (MCDM) approach for offering optimisation as a brokerage service. The specification and implementation details of this proposed software mechanism are thoroughly discussed while the background method used is summarised. Both method and brokerage service allow for the multi-objective assessment of cloud services in a unified way, taking into account precise and imprecise metrics and dealing with their fuzziness. We cope with the fuzziness of imprecise metrics in the sense that this approach deals with linguistically expressed preferences and cloud service characteristics that lack a fixed or precise value and entail a level of vagueness which can only be captured using the Zadeh’s Fuzzy Set Theory. Furthermore, this paper reports on a number of experiments that were conducted in order to measure PuLSaR’s performance and scalability.

Highlights

Nowadays, enterprises are increasingly moving their IT environments into the cloud, reducing operating costs by converting from a business model reliant on hardware and software ownership, to one based on utility service consumption
In order to cope with all the meaningful metrics that should be used for an optimised use of cloud service offerings, we developed the Preference-based cLoud Service Recommender (PuLSaR)
7 Evaluation of PuLSaR In order to evaluate the performance of PuLSaR against an increasing set of requirements that could be met in a cloud service broker, we conducted a number of experiments and measured certain key performance indicators (KPI’s) that assess the behaviour of PuLSaR

Summary

Introduction

Enterprises are increasingly moving their IT environments into the cloud, reducing operating costs by converting from a business model reliant on hardware and software ownership, to one based on utility service consumption This has resulted in an unprecedented rise of cloud providers that serve their offerings as a service but at the same time has created additional challenges (e.g., regarding the quality-of-service [1], security etc.). We define an imprecision model as a set of qualitative cloud service metrics that cannot be objectively quantified or measured These metrics can be used for both describing a cloud service and expressing requirements during the cloud service selection phase. The precise metrics refer to those that include only crisp values (i.e., quantitative/measurable without any uncertainty) while the imprecise metrics refer to those that cannot be objectively quantified or measured and usually include fuzzy or linguistic values, for both describing and expressing a requirement for a cloud service offering. They extend the notion of membership of an element in a set, from binary (belongs or not belongs) to a grade of membership, expressed as a real number, usually in [0,1] interval

Objectives

Results

Conclusion