DuraCloud Pilot Program: utilizing cloud infrastructure as an extension of your repository

Abstract

Cloud compute and cloud storage are now available from several commercial providers around the globe as a commodity service. Due to the extremely low cost, ease of use, and instant scalability the Academic community is taking a hard look at how to best utilize this resource as an extension of its own IT environment. Even though there has been much interest in using cloud infrastructure few organizations have seriously integrated the cloud as part of their environment due to several challenges foreseen, whether real of perceived. In a large survey undertaken by the DuraSpace organization in Winter 2010, some of the biggest challenges identified by our community were security, performance, reliability and trust. Little data has been published to either validate or discredit the key challenges noted within the academic community, although much anecdotal information exists regarding site outages, poor performance, loss of data and the like. The purpose of this presentation is to present the findings of a large scale pilot program utilizing cloud infrastructure from multiple commercial cloud providers, as a utility. The presentation will discuss the key challenges and benefits identified when using cloud storage and compute as a utility during the pilot program. The presentation will provide detailed analysis, where possible, across multiple cloud providers. The analysis presented will include, when applicable, what solutions were deployed to overcome security, reliability, performance and other identified technical issues.