
The GÉANT Cloud Flow (GCF) platform, available in the form of a software package, is available to enable the first pilot implementations in the research community.
The first cross-border containerized community cloud offering provides hybrid cloud software and tools for the Research & Education community by allowing NRENs and Institutions to aggregate their cloud infrastructures to manage them as a secure, robust and flexible integrated cloud platform. The platform allows researchers to define and implement compute resources and workflows within a single, unified and easy to use dashboard integrated with eduGAIN AAI.
For NRENs
NRENs can install the GÉANT Cloud Flow software package to the current cloud assets and start delivering container cloud resources to their institutions.
For Institutions
Institutions can deploy the GÉANT Cloud Flow platform to their current campus infrastructure allowing their internal IT team to offer container cloud resources to their researchers.
For Researchers
The GÉANT Cloud Flow platform enables you to define and implement compute resources using a single easy-to-use dashboard. The workflows you define is executed on the available resources, whether that be local or shared with your NREN.
Current Pilots
The GÉANT Cloud Flow platform is being used in several pilots in academia and industry. For an overview of pilot projects, please visit
https://clouds.geant.org/community-cloud/cloud-flow-pilots/
Typical use cases for GÉANT Cloud Flow
- Natural language processing analysing qualitative surveys)
- Chemistry, physics (virtual labs for constructions, molecule reactions)
- Meteorology (weather & climate simulations)
- Earth Observation (earthquakes, simulations of satellite data, including archaeology)
Resources
The beta programme is delivered on resources hosted in GARR, CSC and GRNET
During the initial beta programme phase the offerings will be available free of charge for qualifying NRENs. For inquiries on qualifying to be part of the beta programme, please contact clouds@geant.org
Platform Offerings
The range of Community Cloud offerings available in the beta programme includes:
GARR Container Platform https://cloud.garr.it/containers/
-
- The GARR Container Platform uses the same accounts as the GARR Cloud Compute Platform.
- Features nearly 5,000 cores and 5.5 PB of storage space; Nvidia Tesla V100 GPU Node
CSC Rahti Container Cloud
-
- Rahti runs on OKD, the community distribution of Kubernetes that powers Red Hat OpenShift.
GRNET Okeanos https://okeanos.grnet.gr/home/
-
- Okeanos runs OpenStack - Max number of VMs 10,000
- Storage 300 TB
Architecture
The architecture above illustrates the coordinated effort of GÉANT community cloud team to deliver large scale workflow analysis across international borders. The GÉANT community cloud offering is a micro-service implementing the Global Alliance for Genomics and Health (GA4GH) Workflow Execution Service (WES) API specification for the execution of workflows written in the Common Workflow Language (CWL).
cwl-WES is a Flask/Gunicorn application that makes use of Connexion to implement the GA4GH WES OpenAPI specification. It enables clients/users to execute CWL workflows in the cloud via a GA4GH Task Execution Service (TES)-compatible execution backend. Workflows can be sent for execution, previous runs can be listed, and the status and run information of individual runs can be queried. The service leverages cwl-tes to interpret CWL workflows, break them down into individual tasks and emit GA4GH TES-compatible HTTP requests to a configured TES instance.
Hosting NRENs
- GRnet -Greece
- Consortium GARR - Italy
- CSC - Finland
Development Team
NRENs in the GN4-3 Project Community Clouds Team
Demo: Scheduling Scientific Containers on a Cluster of Heterogeneous Machines
Access to the Community Cloud Beta Programme for approved, on-boarded users: https://ui.gcf.cloud.grnet.gr
For more information please visit https://wiki.geant.org/pages/viewpage.action?pageId=159055907 (Federated access required)
The software is available in GitHub to use but will not be developed and maintained further by the GÉANT Cloud Team.