Science in stock with Meso@LR

At a time when science increasingly needs to work with data, the University of Montpellier, with financial support from the Region and the the Metropoleprovides the scientific community with a shared storage space. More efficient, more secure and more environmentally friendly, it is part of a regional initiative that Inserm is also involved in. Anne Laurent, vice-president in charge of open science and research data, explains.

Inserm and the University of Montpellier have just signed a cooperation agreement to set up a shared data management environment. Why is this important?
It's important because, in order to do science and produce scientific results, engineers, researchers, teacher-researchers and doctoral students increasingly rely on data. Data is a material that is stored, manipulated, exploited and enriched; and all this requires enormous storage and processing space, which must be shared. Meso@LR provides these spaces for our scientific communities.

What is Meso@LR?
It's a sort of gigantic computer divided into two parts: intensive computing and storage. On the computing side, there are huge processors. On the storage side, Meso can be compared to a super hard disk, or rather two super hard disks placed in two different rooms to double security. It is this super hard disk that is gradually being made available to the scientific community for work. This convention is a kind of launch signal for this collective construction, which everyone is invited to join.

And those big hard disks, how much storage capacity do they have?
The Meso@LR data infrastructure can store a whopping 15 petabytes! 15 petabytes corresponds to 15 million billion bytes. That's the equivalent of about a million USB sticks!

Will this platform be open to all Muse partners?
It's broader than that. It's one brick in a regional construction. Meso@LR is a structure managed by the University of Montpellier, but it is intended to cover all the IT needs of research in Eastern Occitania. It is co-financed by the Region and the Metropole, with the UM assuming responsibility for hosting, human resources, electricity, etc. from the outset.

So the agreement with INSERM is the first in a series?
Yes, INSERM is the first organization to sign, but we do work with other partners. And this cooperation is reinforced by Inserm's leading role in the construction of the Montpellier Data Science Institute (ISDM).

And what exactly is planned?
Inserm now has a 2 petabyte storage slice on Meso@LR, in which they will be able to build their offer and services to the scientific communities associated with Inserm. They have the keys to it, and are managing it thanks to the skills of their IT teams.

The UM has no access to their storage space. Isn't this data sharing?
No, Inserm defines its own data-sharing policy, and these data are and remain the property of the supporting institution. There are other infrastructures for sharing and opening data in the national context.

And what's in it for them?
By pooling these infrastructures in highly secure environments, we enable our IT teams to work more serenely, by reducing fears of problems, such as air conditioning, by enabling them to acquire top-notch equipment, and by sharing their experience and skills. In this way, we can develop even more ambitious projects.

And for users?
Value-added services, easier, more efficient data handling environments. It's a paradigm shift, like not having your own tools at home, but having better ones - there's everything to gain. What's more, Meso's scientific projects involve joint research units, and are designed to provide a harmonized response to the questions that all researchers are asking.

What kind of services will we be able to offer?
To give an example, the cloud offering we're currently building will enable us to offer "virtual machines". Instead of having to install equipment and software for accessing and using data on each of our computers, they will be pooled in this space. What's more, this will boost performance and security.

By the way, who technically manages Meso?
It's a team of top-level engineers, and I stress this because it's really impressive. Here too, it's a collective and mutualized effort, with each partner agreeing to make engineers available for 20 or 30% of their time, sometimes more. On these subjects, no one can do it alone, we really have to work collectively, otherwise we wouldn't be able to move forward.

And from an ecological point of view, is there any point in pooling?
Today, we know that placing these environments in dedicated, high-performance energy-efficient buildings reduces their impact. But this gain also depends on how we use them, both individually and collectively. I was the first to fill up my hard disk because I had the space - it's like having a cupboard! We need to learn to kill, to delete our data when it's not needed. In this respect, the work carried out to support our employees is essential, and a project involving several departments is being carried out at the UM on this issue.

By creating a new storage space when laboratories already have their own solutions, isn't there a risk of doubling storage?
The idea is that, as we go along, we should be able to disconnect local solutions that cost energy in terms of human time and electricity, except when they're needed, and sometimes they are. That's why we've included it in the agreement.

A cost of 300,000 euros. What does it correspond to?
It's the full cost: electricity, human resources, hosting and storage for this Inserm tranche over 5 years. Inserm also contributes human resources and skills, which are invaluable in providing a quality service.

Are there any links between Meso@LR and the Data Science Institute?
Yes, we're convinced that data science and data for science are going to be at the heart of science. And this poses major challenges for making this data easy to find, accessible, interoperable and reusable(FAIR). To manage and exploit this data, we need skills, we need support... This is one ofISDM's missions. Some of these issues are also research topics in which the laboratories on the Montpellier site are very active and recognized.

And how do you get access to these support services?
The UM's "interdirection pour la science ouverte" group, in association with ISDM, Méso@LRand all Muse partners, is the answer. This support is available through the ENT catalog of services. We are currently working together to enhance its visibility and services.

Is open data a goal?
It's sometimes said: " data: as open as possible, as closed as necessary. " Once again, we need to give researchers time to exploit and add value to their data, and Meso is precisely designed for science in the process of being "FAIR", so to speak. It's neither a data warehouse nor a data portal. However, data governance is an issue that needs to be addressed in order to collectively build environments that are designed to be shared and opened up. We want to make Montpellier a place where scientific data is not only produced, but also exploited and, where possible, shared as a common good.