The SNIC Science Cloud is a Community Cloud. This means that use of the IaaS resources is open free of charge to Swedish PIs and their collaborators. To use the resource, it is therefore necessary to fill out a project proposal via SNIC’s user management platform, SUPR. This procedure follows standard practices used today for other SNIC resources, and require a PI-initiated resource request. User membership of a project is then delegated to the PI. Actual models for resource quotas are currently under development.
The SNIC Science Cloud is currently run as a project by SNIC. This means that at present, the resource is offered with limited support and no QoS guarantees. Despite this, it is a very useful platform to develop knowledge of and explore the usefulness of cloud computing for your research. You can read more about the different service layers typically associated with cloud computing (IaaS, PaaS and SaaS) and how the SNIC Science Cloud maps to those by following links in the left menu.
The SNIC Science Cloud is not meant to be a replacement of SNIC’s existing supercomputing resources (HPC clusters), but rather a complement, offering advanced functionality to those users that need more flexible access to resources (for example more control over the operating systems and software environments), want to develop software delivered as services (need to run e.g. webservers, open custom ports etc), want to explore use of modern technology such as Docker, or want to provision e.g “Big Data” stacks such as Apache Hadoop/Spark. In addition, cloud computing resources open up for models of computing not possible to support on SNIC’s batch cluster resources, such as Interactive Computing. In addition, Cloud Computing resources is a good match for typical Many-Task Computing (MTC) and High-Throughput Computing (HTC) problems, especially if they require more customization of software stacks, and/or exposure to users via e.g. WebUIs.
The major aim for the SNIC Science Cloud going into 2016 is a rapid convergence to a stable production quality IaaS resource and scaling of that resource to multiple regions in Sweden. This will be followed by selected PaaS offerings pending user input. If you have needs of e.g. Apache Hadoop/Spark or similar environments in your research, we would like to hear from you. Also, we can always use a helping hand. If you are skilled in cloud computing, service automation, Big Data, or a related field, please do not hesitate to contact us to get involved in the project.