Calcul Infrastructure

All of the IFB’s computing and storage offerings are supported by the NNCR (National Network of Computing Resources).

Community services are provided by the NNCR, which includes all the hardware (IFB-core + regional platforms) that contributes to ensuring the service offering. The NNCR relies on a distributed infrastructure consisting of 8 clouds and 6 clusters spread across 13 regional centers, and 2 centralized resources: the IFB-core servers (cloud in Lyon, cluster in Orsay), and nine high-performance computing platforms located in different regions. This network is open to French and international research communities, both public and private, in the field of life sciences.

The IFB also plays a key role in the national strategy for pooling and modernizing scientific digital infrastructure. Finally, at the ELIXIR level, this service contributes to the maintenance of Biocontainers and the coordination of the Galaxy community within ELIXIR.

Compared to data processing in other scientific fields, the specific characteristics of biological data make them dependent on a very specific hardware architecture, software environment, and data collections that are currently not available in national computing centers. To address these challenges, the IFB has organized its services by decentralizing physical, logistical, and human resources.

• Platforms participating in the Cloud federation: AuBi , Bilille , BiRD , GenOuest , IFB-BigEst-Cloud , IFB-core
• Platforms participating in the clusters : ABiMS , BiRD , GenoToul , GenOuest , IFB-BiGEst-Cluster , IFB Core , MicroScope, Migale , South Green

Such decentralization of resources offers at least four main advantages:

1. Service security in the event of a local outage
2. The distribution of data transfer flows
3. Geographical proximity
4. The joint development of best practices and advanced technical expertise for computing, processing and storing data.

The cloud federation offers a set of preconfigured environments, allowing scientists and bioinformaticians to choose the configuration best suited to their analyses. These virtual bioinformatics environments developed by some members of the IFB community are registered in the Biosphère catalog . Scientists and engineers can launch their own virtual environments with reserved resources that can be adapted to their own needs without interfering with other users. Activity is constantly growing, and the French IFB cloud has already been used for numerous scientific analyses and training sessions, workshops, hackathons or several recurring sessions of scientific schools or universities.

An account offers variable space (from 25 GB to 3 TB per virtual machine for the duration of job execution but deleted when the virtual machine is shut down) to which will soon be associated a persistent storage space shared between the virtual machines of a user or group.

You can also use the IFB cloud as a support for your training. To find out more

To request an account on the IFB Cloud, consult our computing offer.

All clusters on our regional platforms are dedicated to regional and thematic users.

What is the IFB-core cluster?

The Core Cluster is IFB’s national computing infrastructure. It aims to meet the computing needs of all communities in the health and biology fields, with a focus on users who do not have local computing resources.

The infrastructure is open to all users with an academic email address in France or in one of the ELIXIR member countries .

The Core Cluster is accessible via three modalities:

• SSH for direct cluster usage via SLURM,
• the Open Ondemand portal to use interactive tools such as JupyterLab or RStudio,
• the UseGalaxy.fr portal, the national Galaxy authority.

The Core Cluster is administered collaboratively. More than six engineers from five IFB platforms build and contribute to the project daily. To manage the multiple contributions, they are overseen by CI mechanisms (Ansible + Gitlab runner) connected to a common code repository (Gitlab).

Complete Cluster documentation: https://ifb-elixirfr.gitlab.io/cluster/doc/
Terms of Use: https://ifb-elixirfr.gitlab.io/cluster/doc/terms-of-usage/

Services

Accounts management
All academic users could request an account to access Core Cluster via our account registration and management portal : https://my.cluster.france-bioinformatique.fr. Beware, this account is usable only for the Core Cluster and won’t be recognize by other IFB services.

Submission of computer tasks
The primary use of the cluster is with SLURM from an SSH console.

If you are new to SLURM, please read the Core Cluster documentation and tutorial to learn how to submit your first jobs:
https://ifb-elixirfr.gitlab.io/cluster/doc/slurm/slurm_user_guide/

Interactive web portal
The Open Ondemand web portal allows you to run interactive tools like RStudio or Jupyterlab on cluster resources via a simple web interface. Find the overview page , documentation and dedicated video .

Galaxy
The IFB Core Cluster provides the computing resources for the French instance of Galaxy: usegalaxy.fr
usegalaxy.fr offer a large panel of bioinformatics software usabled in line. Some software are also usabled from thematics sub-domaines such as metabolomic, singleCell, covid19, etc.

Technical support
The support.cluster.france-bioinformatique.fr portal allows cluster users to contact our support team for any technical requests, including:

• Using SLURM or Open OnDemand,
• Installing tools on the cluster (excluding UseGalaxy.fr *),
• Management of user accounts or project spaces.

* Support for usegalaxy.fr is available on the IFB community forum

Support for the bioinformatics community
A community forum allows you to exchange ideas with biologists and bioinformaticians on bioinformatics: uses and options for a tool, setting up a workflow around a specific theme, etc.

Training hosting
The IFB Core Cluster can provide computing resources for your training session. Our support team offers a variety of options, including creating temporary generic accounts and reserving cluster resources.
To request cluster resources for your training, complete the request form .

Contribute to the IFB-core Cluster TaskForce

If you have skills in bioinformatics tools or system administration, join the IFB Core Cluster TaskForce to deploy your tools or contribute to the management of the IFB infrastructure and become part of the legend. Please do not hesitate to contact us by sending an email to:

contact-nncr-cluster@groupes.france-bioinformatique.fr Otherwise, if you wish to be trained in the technologies used on the infrastructure, training/tutorial sessions are offered regularly.

Continuous Integration (CI) & Collaborative Work
Resource administration is carried out collaboratively. In order to manage multiple contributions, these are managed by a continuous integration mechanism connected to a common code repository.

Traceability & Secure Contributions
All installation, configuration and maintenance actions must be traceable as much as possible. This is in order to:

• Debug
• Be able to go back if something goes wrong
• Inform other administrators of changes made

We chose to use a Git repository hosted on GitLab. Git meets all our traceability needs. The GitLab interface provides us with a space to exchange information and offers the ability to work with Merge Requests (Pull Requests) and host our own CI job runners. These MRs require the inclusion of a code review phase before going live. Each change or addition is validated by a peer review to avoid errors and ensure that at least two people are aware of it.

Task Force Members:

• Gildas LE CORGUILLÉ: CNRS/Sorbonne University, ABiMS, FR2424, Roscoff (co-responsible)
• Julien SEILER: CNRS IFB-Core/BiGEst, UAR 3601, Strasbourg (co-manager)
• David BENABEN: INRAe, Fruit Biology and Pathology UMR 1332, Bordeaux
• Nicole CHARRIERE: CNRS, IFB-Core/Genouest, UAR 3601, Rennes
• Manon CONNAULT: INRAe, IFB-Core/Migale, UAR 3601, Jouy-en-Josas
• Thomas CHAUSSEPIED: CNRS, IFB-Core/Genouest, UAR 3601, Rennes
• Jean-Christophe HAESSIG: CNRS, IGBMC/BiGEst UMR 7104, Illkirch
• Didier LABORIE: INRAe, GenoToul Bioinfo, Toulouse
• Guillaume SEITH: INSERM, IGBMC/BiGEst U 1258, Illkirch