Knapsack-Sharing Model for Hybrid Hosting of Enterprise Resource Planning Software on IaaS Clouds

Cloud Computing and Enterprise Resource Planning constitute hot research and application topics thanks to their direct impact on the success or failure of contemporary organizations. The competitive value of an organization relies on how much it adopts Cloud Computing and Enterprise Resource Planning solutions. This paper describes a knapsack-sharing model for the hybrid hosting of Enterprise Resource Planning software in a cloud environment. Kanpasack-sharing is a combinatorial optimization approach that uses Integer Linear Programming models. On the one hand, the proposed approach allows the Cloud Service Provider to estimate the cost of hosting such complex software by solving a knapsack-sharing problem instance. On the other hand, it helps the organization's management in the fair distribution of the budget among departments and business areas since budget constraints are often present. In addition, the proposed approach helps the organization's management decide what processes to move to the cloud platform and what other processes remain on in-house servers. Moreover, we describe a new improvement of an exact algorithm to solve the knapsack-sharing problem instances with a large number of classes. The experimental work reveals a substantial improvement in running time compared to the most recent algorithm in the literature.