The Cloud Computing paradigm Clause Samples

The Cloud Computing paradigm. In the original CompBioMed proposal we mentioned the fact that academically run HPC systems invariably support diverse groups of many users and so are subject to resource contention issues. Although within research environments this lack of quality of service guarantee is an inconvenience, scientists are resigned to it and simply work around the problem in order to benefit from the ultimate goal of producing high-quality research papers which make use of and often depend on access to large-scale computing resources. However, in environments subject to hard deadlines (for example, when decision support is required for urgent clinical intervention) it becomes important to be able to immediately and reliably allocate appropriate resources sufficient to meet the goal. When the original CompBioMed proposal was written this certainly was an important aspect to explore, two years later cloud providers have become key players in both academic and non-academic environments for the provision of on-demand computing. There are many reasons for this, the following two being the most important for us: (a) containers are progressively more powerful, flexible and efficient, even when deployed on traditional HPC resources, and (b) at the same time that governmental supercomputing centres are selling core hours on their resources, commercial cloud providers offer "bare metal" instances improving their HPC capacities. It is worth remarking the importance of containers: they are reduced versions of an operating system with all associated software that a given application needs to run. They are therefore an extremely flexible way of deploying applications reliably, reproducibly, and rapidly, in heterogeneous cloud environments. For these reasons, we strongly believe that Cloud Computing is the real novel HPC architecture today. Cloud computing advances for biological systems and its importance as an HPC resource are described in [3], a work partially funded by CompBioMed. This makes the observation that in the last ten years, virtualisation technologies underwent significant enhancements to their accessibility and ease of use. Considering also the current great abundance of hardware resources, not only full-stack but also lightweight virtualisation (containers) are becoming ideal platforms on top of which users can build their own "cloud-based platforms". Operating-system-level virtualisation, also known as “containerisation”, is an increasingly popular strategy today. There a...