Scheduling strategies for optimal service deployment across multiple clouds

The current cloud market, constituted by many different public cloud providers, is highly fragmented in terms of interfaces, pricing schemes, virtual machine offers and value-added features. In this context, a cloud broker can provide intermediation and aggregation capabilities to enable users to deploy their virtual infrastructures across multiple clouds. However, most current cloud brokers do not provide advanced service management capabilities to make automatic decisions, based on optimization algorithms, about how to select the optimal cloud to deploy a service, how to distribute optimally the different components of a service among different clouds, or even when to move a given service component from a cloud to another to satisfy some optimization criteria.In this paper we present a modular broker architecture that can work with different scheduling strategies for optimal deployment of virtual services across multiple clouds, based on different optimization criteria (e.g. cost optimization or performance optimization), different user constraints (e.g. budget, performance, instance types, placement, reallocation or load balancing constraints), and different environmental conditions (i.e., static vs. dynamic conditions, regarding instance prices, instance types, service workload, etc.).To probe the benefits of this broker, we analyse the deployment of different clustered services (an HPC cluster and a Web server cluster) on a multi-cloud environment under different conditions, constraints, and optimization criteria.

► We propose a novel cloud broker architecture adapted to multi-cloud environments.
► We focus on the scheduling module and we explain different scheduling policies.
► We evaluate the performance of an HPC and a Web server cluster over the cloud.
► We present a profit analysis of the cloud scheduler applied to these use cases.