A resource provisioning framework for bioinformatics applications in multi-cloud environments

Abstract

The significant advancement in Next Generation Sequencing (NGS) have enabled the generation of several gigabytes of raw data in a single sequencing run. This amount of raw data introduces new scalability challenges in processing, storing and analyzing it, which cannot be solved using a single workstation, the only resource available for the majority of biological scientists, in a reasonable amount of time. These scalability challenges can be complemented by provisioning computational and storage resources using Cloud Computing in a cost-effective manner. There are multiple cloud providers offering cloud resources as a utility within various business models, service levels and functionalities. However, the lack of standards in cloud computing leads to interoperability issues among the providers rendering the selected one unalterable. Furthermore, even a single provider offers multiple configurations to choose from. Therefore, it is essential to develop a decision making system that facilitates the selection of the suitable cloud provider and configuration together with the capability to switch among multiple providers in an efficient and transparent manner. In this paper, we propose BIOCLOUD as a single point of entry to a multi-cloud environment for non-computer savvy bio-researchers. We discuss the architecture and components of BIOCLOUD and present the scheduling algorithm employed in BIOCLOUD. Experiments with different use-cases and scenarios reveal that BIOCLOUD can decrease the workflow execution time for a given budget while encapsulating the complexity of resource management in multiple cloud providers.