Blueprint for an Open Source On-Premise Cloud Infra-structure to Serve as a Research Data Infrastructure for Universities

Vogl R., Hölters J., Ketteler-Eising M., Rudolph D., Blank-Burian M., Angenent H., Schild C., Ost S.

Forschungsartikel in Sammelband (Konferenz)

Zusammenfassung

In response of the significant increase in the amount and variety of research data, the European Data Infrastructure (EDI) and the European Open Science Cloud (EOSC) foster persistent, highly available and compatible data infrastructures where data from various disciplines can be stored and accessed. These infrastructures should not only provide storage but also tools for processing and analysis. To prepare the implementation of such an extensive research data infrastructure for a group of five Universities, Münster University (lead of the consortium) has invested substantial manpower in developing a versatile, scalable and performance optimized hyperconverged deployment of OpenStack (cloud stack for virtual machines and Storage as IaaS) and Ceph (as underlying Software Defined Storage) using kubernetes as container orchestration engine on industry standard hardware. This is the first instant that advanced leading edge cloud technology like kubernetes has been put to use in any of the participating university IT centers and we see this as pivotal for out future approach to system architecture. The Open Source approach was adopted for cost reduction and sustainability. Remarkable is the approach to build on community versions of the Open Source software only, without vendor support. A scaled down pilot system has been operational for well over a year now, and demand for such an infrastructure is mounting from numerous research groups from a wide range of disciplines. Implementation of the full scale cloud system is planned for mid 2019. This is an update on the very preliminary report on the project given at EUNIS 2018.

Details zur Publikation

Veröffentlichungsjahr: 2019
Sprache, in der die Publikation verfasst istEnglisch