Desktop to Cloud Migration of Scientific Computing Experiments

TL;DR

This paper introduces a tool suite that simplifies migrating scientific computing environments to the cloud, enabling easier setup, management, and optimization of cloud-based experiments for scientists without deep cloud expertise.

Contribution

The paper presents D2CM, a novel desktop-to-cloud migration tool that automates environment transfer and supports lifecycle management on cloud platforms like Amazon and Eucalyptus.

Findings

D2CM simplifies cloud migration for scientific experiments.

The tool enables optimization of computational costs.

Case study demonstrates effective application in supercapacitor research.

Abstract

Scientific computing applications usually need huge amounts of computational power. The cloud provides interesting high-performance computing solutions, with its promise of virtually infinite resources on demand. However, migrating scientific computing problems to clouds and the re-creation of software environment on the vendor-supplied OS and cloud instances is often a laborious task. It is also assumed that the scientist who is performing the experiments has significant knowledge of computer science, cloud computing and the migration procedure, which is often not true. Considering these obstacles, we have designed a tool suite that migrates the complete software environment directly to the cloud. The developed desktop-to-cloud-migration (D2CM) tool supports transformation and migration of virtual machine images, reusable deployment description and life-cycle management for applications to be hosted on Amazon Cloud or compatible infrastructure such as Eucalyptus. The paper also presents an electrochemical case study and computational experiments targeted at designing modern supercapacitors. These experiments have extensively used the tool in drawing domain specific results. Detailed analysis of the case showed that D2CM tool not only simplifies the migration procedure for the scientists, but also helps them in optimizing the calculations and compute clusters, by providing them a new dimension -- cost-to-value of computational experiments.