At the Digital Infrastructures for Research Conference in Krakow, Poland, we had the opportunity to talk with Yannick Legré, Director of EGI, the European-rooted federation for advanced computing and data research. EGI's aim is to serve the research communities in Europe and provide them with the necessary resources to overcome the scientific challenges. EGI is federating slightly less than 900,000 computing cores that it is providing through its 325 computing and data centres in Europe and in more than 50 countries as well. The resources that EGI provides to the different research communities are very important. The more challenging part is to federate all the actors throughout Europe. This is what the EGI Federation has been successful at with all the different actors contributing and collaborating together to serve the scientists.
The goal is not to provide bare metal like some very well known commercial providers. The goal is to provide services to the research communities. With respect to that EGI has recently been working on a service catalogue to clarify the offer and what EGI can provide to the scientists. Over the last 18 months EGI has cleared its service portfolio. Now, EGI is just finalizing the service catalogue that is available for the entire scientific community wherever they are organized, such as the pan-European research infrastructure or the individual user or small groups of researchers. The service catalogue is also available online. At the EGI website you can get all the necessary information to interact or connect with the EGI Federation. However, it is more likely that individual researchers will first connect with the national contact point in their country because they are much closer to each other. The support is federated at a European level. Depending on the application area or the collaboration EGI established with the different researchers the support could be provided by a few countries or at a general level through the EGI Foundation.
Collaborating with the other infrastructures is always very important but is also a big challenge because the infrastructures are all coming from a different background. None of them can achieve everything on their own, so the collaboration is vital. Yannick Legré stated that they really need to consolidate what they are offering to the end users, whether they are a big research infrastructure or a small group of researchers. EGI is working more towards respect of standardization to also improve the quality of the level of services EGI provides to the user. The EGI Foundation is finalizing its ISO certification. By the end of the year, EGI will be certified on ISO 9001 and 20,000. It is also a very big step towards more maturity and more capacity to have a repeated outcome of what you are doing and being able to serve with the same level of quality on a permanent basis. It is really EGI's goal to provide a consistent high quality of service to the user.
EGI is collaborating with all e-Infrastructures, starting from GEANT, the underlying infrastructure on which the EGI Foundation has been built. All its computing and data centres are connected to GEANT. EGI is also collaborating with EUDAT for data curation and long term access to the data, and to build up some use cases with specific user communities. Of course, EGI is collaborating with OpenAire for all the publications concerned. Yannick Legré said that the scientists who are using the EGI infrastructure were asked to publish their data with a specific tag that can automatically be collected by OpenAire. In some way, EGI also collaborates with PRACE for some specific applications or some specific user communities that require to access both Cloud-type and high performance computing facilities. EGI is providing high-throughput computing, which was known in the past decade as Grid computing. EGI is also providing Cloud computing. PRACE represents the top level of high performance computing and is aiming at exascale computing but it is not yet there.
An important development for the European scale are the Key Performance Indicators (KPIs), according to Yannick Legré. Most, if not all, the e-Infrastructures are already monitoring their infrastructure because they are bound to do this by law. EGI is collecting a lot of metrics. The interest of collecting a KPI is to answer a very specific question. You really have to know what type of question or what kind of answer you want to get to be able to accurately collect a KPI. To be able to do that EGI has produced the service catalogue and the ISO certification. This helps EGI to clarify all the processes, all the procedures and to define the appropriate metrics and KPIs that EGI has to collect in order to collaborate with the other e-Infrastructures.
To compare one e-Infrastructure with another one would be a slightly different problem because each e-Infrastructure can only collect the metrics and the KPIs which refer to their own infrastructure. The analysis of these different KPIs and the harmonization between the different infrastructures will be something different that has to happen afterwards. There is a new project which has recently been approved by the European Commission and is called eInfraCentral that is exactly trying to achieve this. EGI is involved in this project, just like the other e-Infrastructures. In this respect, e-IRG also has the goal of collecting policy-related KPIs.
The Digital Infrastructures for Research Conference has been a great success. This is the first time that all the e-Infrastructures have come together to organize a user forum. About 400 people attended the conference. The organizers would love to welcome more user communities and scientists using the services. As this was the first edition, Yannick Legré was very pleased with the outcome of the conference and he hopes to do even better at the next conference. It has been very challenging and interesting for the organizers, as well as for the research community. It is now possible to see and to combine the services offered by the different e-Infrastructures.