Research IT as a Service
I would like to share with you a couple of points taken from Ian Foster's presentation about the future of research computing and the current trend of moving the administrative load to the cloud. While, the big science projects are the ones that most often get the required funds and resources, the small to medium projects on the other hand are struggling to take off. Small-to-medium projects may end up using inappropriate tools and fail to materialize! Research projects' requirements are substantial and they depend on reliable results which need to be processed in a reasonable time. Let alone the competitive element which a business may be relying on! The current approaches taken by the small-to-medium scale projects are unsustainable mainly due to the high administrative costs related to them. Therefore, the effort put in by the R&D staff should focus on the project goals or business value rather than on the administrative part. For that reason, the trend is to outsource the IT functions. For instance, the Cloud service model IaaS (Infrastructure as a Service) would offer great benefits in data analytics and content distribution whereas IaaS combined with SaaS (Software as a Service) would reduce research IT costs and provide access to powerful tools.
Ian Foster refers to this model as research IT as a service and gives a list of time-consuming tasks and examples related to science projects. The market is providing researchers with some choices such as, commercial cloud providers, open source grids, XSEDE supercomputers and campus systems. The concept of Research IT as a service may be seen as a combination of other services such as, research data as a service, data analysis as a service, etc. The whole concept would lead us to a new cyber-infrastructure strategy aiming at more capability for more people at less cost.