Efficient Migration-Aware Algorithms for Elastic BPMaaS

Abstract

Abstract. As for all kind of software, customers expect to find business process execution provided as a service (BPMaaS). They expect it to be provided at the best cost with guaranteed SLA. From the BPMaaS provider point of view it can be done thanks to the provision of an elastic cloud infrastructure. The provider still have to provide the service at the lowest possible cost while meeting customers expectation. We propose a customer-centric service model that link the BPM execution requirement to cloud resources, and that optimize the deployment of customer's (or tenants) processes in the cloud to adjust constantly the provision to the needs. However, migrations between cloud configurations can be costly in terms of quality of service and a provider should reduce the number of migration. We propose a model for BPMaaS cost optimization that take into account a maximum number of migrations for each tenants. We designed a heuristic algorithm and experimented using various customer load configurations based on customer data, and on an actual estimation of the capacity of cloud resources.

A Genetic Algorithm for Cost-Aware Business Processes Execution in the Cloud

“…We propose an integer linear programming (ILP) model and a genetic algorithm that aims at finding the best allocation strategy while limiting the number of migrations per tenant. We improve our previous results [2] with a method that provides a better cost elasticity.…”

“…We proposed in [2] a method based on an iterative heuristic and time series segmentation. It computes the list of necessary cloud resources and a mapping of tenants on each resource, time slot per time slot.…”

“…In this part, we recall briefly our previous method [2]. It has two parts: first, we choose a migration strategy i.e the time slots where each tenant can migrate, and, second, we apply a heuristic using resources prices and capacity, tenants needs, and their migration strategy in order to obtain the cloud resources and the placement of tenants.…”

“…To evaluate it on the different individuals, we compute the cloud resources allocation and placement of the tenants on it. In our case, we run our iterative time slot heuristic [2] or a solver on our model presented in Section 2.4, for each individual (migration strategy) we need to evaluate. We keep the cloud resources and tenants assignation distribution in memory for the next steps, and the fitness score.…”

“…We have conduct tests with the cloud resource prices and sizes, and the seeds of our previous work [2]. We consider 12 configurations, each composed of two Amazon Web Services compute resources: one database resource (RDS) for the database, and one compute instance (EC2) for the application server.…”