A Petri net-based decision-making framework for assessing cloud services adoption: The use of spot instances for cost reduction

Abstract

Cloud services are widely used nowadays, especially in Infrastructure as a service (IaaS), with vendors offering several purchasing options and expanding the range of services offered on almost a daily basis. Cost reduction is a major factor promoting the adoption of cloud services among enterprises. However, qualitative factors need to be evaluated as well, thus rendering the decision regarding the adoption of cloud services among enterprises a non-trivial task for Information Technology (IT) managers. In this paper, we propose a place/transition or Petri net-based multi-criteria decision-making (MCDM) framework to assess a cloud service in comparison with a similar on-premises service. The framework helps IT managers choose between two such options, and can be used for any type of cloud service: Infrastructure as a Service (IaaS), Platform as a service (PaaS), Software as a service (SaaS), etc. Because its low cost is among the most important reasons for adopting cloud services, we also propose a Petri net to model cost savings using the spot instances purchasing option in public clouds. Through simulation of several scenarios, we conclude that spot instances present a very interesting cost-saving option in the auto-scaling process, even for simple business applications using few servers.