EGI-Engage interim impact report

Tiziana Ferrari answers an EC survey with figures from March 2015-August 2016  

In September, the European Commission's DG-CNECT sent an evaluation questionnaire to all e-infrastructure projects. The questionnaire defines five dimensions of impact: relevance, effectiveness, efficiency, coherence and EU added value.

Here is a summary of EGI-Engage's contribution to this initiative.

Relevance

Q: Provide measures of the usage of the services provided by your e-infrastructure and improvement over time. For example, traffic growth on the networks, usage of supercomputing facilities etc.

A: Cloud and Storage service capacity federated in EGI have been growing exponentially during the last years. In the last 12 months the use of compute services has increased by +26%.

Compute and Storage capacity KPIs (2016 QR3)

> 770,000 HTC cores (+26% yearly increase)

> 6,646 Cloud cores (in production since May 2014)

> 285 PB of Online storage (+12% yearly increase)

> 281 PB of Archive storage (+42% yearly increase)

Users

> EGI serves an international user community with researchers from all regions of the globe. The active registered users are estimated to be 48,000 in 2016 QR3 (+26% yearly increase).

Effectiveness

Q: Provide benchmarks comparing your e-infrastructure against e-infrastructures in other scientific powerhouses such as USA, China and Japan.

A: EGI is by far the largest High-Throughput Compute (HTC) federation in the world, with +300 data centres across 56 countries. The infrastructure is operated locally primarily at universities and research laboratories in Europe but also in Africa and Arabia, the Asia Pacific region, North and South America.  

> USA The largest HTC infrastructure in the USA is the NSF-funded Open Science Grid with about 100 distributed data centres. XSEDE, the collaborative federated HPC infrastructure for researchers in the US, federates access to 214,000 cores.

Efficiency

Q: Were things built and stand idle or are they used?

A: The EGI usage figures speak for themselves:

> HT Compute: 1,640 Billion CPU hours (last 6 months, 2016 QR3), +26% yearly increase (the EGI volume is double the amount reported by Open Science Grid in the US)

> HT Compute: 1.7 million jobs/day on average (last 12 months)

> Cloud Compute: 256,000 instantiated VMs (last 12 months)

Q: Does your e-infrastructure assist or harm private business development in an area?

A: EGI supports business development through a programme launched in 2015 aimed at:

> Access to compute, storage and research data services: EGI provides computing capacity to test workflows, models, and applications with dedicated support and consultancy.

> Platform integrators: e.g. Terradue (developer of the ESA geohazards and hydrology ESA exploitation platform) and CloudSME (provider of SaaS services for engineering and manufacturing).

> Providers of Infrastructure as a Service: To date the EGI Federation includes one cloud SME located in UK (100%IT).

Q: Some e-infrastructures are more domain-specific than others – how are infrastructure needs of different sciences treated?  

A: EGI is multidisciplinary and addresses all science disciplines according to the national research priorities. With Compute, Storage and Data Management services it supports 14 fully integrated Research Infrastructures and communities, including for example the LHC - Large Hadron Collider, CTA - Cherenkov Telescope Array or LifeWatch - European Infrastructure for Biodiversity and Ecosystem research.

Coherence

Q: To what extent your e-infrastructure project supports other activities (e.g. internationalization by making infrastructures global, open access policies by proving services for open data and open science, etc.). Do you consider your e-infrastructure neutral with respect to other research priorities?

> EGI services are developed, deployed and supported by open source software, to allow sharing of scientific applications in the form of cloud virtual appliances.

> EGI has a service-oriented approach in which research communities are not only consumers but also providers of innovation and services for community-specific capabilities in one integrated catalogue.

> Co-design is supported through Competence Centres involving Research Infrastructures and research communities that contribute to the development of human networks of experts, services providers and technology providers.

EU Added Value

Q: Could you provide examples about how your e-infrastructure adds value at EU level?

> EGI federates national e-Infrastructures and resources from international research organisations and provides open access to advanced research computing tools, in support of the European research Area (ERA).

> EGI supports the Open Science Commons, an overarching policy designed to overcome the barriers preventing the implementation of the ERA. This concept builds on the principles of Open Science, or Science 2.0, supported by the European Commission and others and also the commons principle for management of shared resources. EGI is seeking other e-Infrastructures, research infrastructures and stakeholders from the digital research community to collaborate in constructing the Open Science Commons.

> EGI contributes to digital innovation by providing access to the digital capabilities or resources available to the research sector for the exploitation of commercially viable research data, and to support testing and experimentation. The most advanced collaborations are those involving industry and SMEs doing business through the exploitation of Earth Observation data including Copernicus data.