Infrastructure as a Service (IaaS) provides virtualized computing resources over the cloud.
NRENs can act as brokers for third-party providers that host hardware, software, servers, storage and other infrastructure components on behalf of users. Or NRENs may decide to build infrastructure and make it available over the cloud to users.
IaaS platforms offer highly scalable resources that can be adjusted on-demand and ideal for academic research characterised by workloads that are temporary, experimental or subject to unexpected change.
IaaS Delivery and Adoption Framework Agreements
GÉANT has established framework agreements with suppliers of Infrastructure as a Service (IaaS) solutions on behalf of NRENs and member institutions in order to enable R&E institutes to adopt and use a portfolio of IaaS cloud solutions in an easy, safe, predictable and controlled way. This portfolio of services provides standardized contract terms for use by the NRENs to bring to their client institutions, without the need to open individual tenders.
Services are connected to the GÉANT and NREN networks and are compatible with the community’s Identity Management capabilities.
To get an overview of the services and their capabilities, see the IaaS service matrix.
~okeanos is an Infrastructure as a Service platform is developed and operated by GRNET (Greek NREN). It is powered by an open source software stack called Synnefo (http://www.synnefo.org), which implements an OpenStack-compatible API on top of the Google Ganeti cluster management system (http://code.google.com/p/ganeti). In operation for several years, ~okeanos has spawned over 500,000 VMs with more than 5,000 active on any given day. Its main users are students, researchers and academics in Greece, while users from across the globe have downloaded and installed Synnefo on their premises.
~okeanos is under constant development. In 2016 the team developed a PaaS solution called Orka that offers Hadoop and Spark clusters on demand on top of ~okeanos for users who wish to handle big data without having to deal with details of Hadoop installations.
~okeanos is available via federated login to the global R&E community.