The future for science in Europe is bright – and full of clouds!

ESA's ESRIN facility in Frascati, Italy, where yesterday's
event was held. Image courtesy ESA.

It may be named after a dying star, but the Helix Nebula project is very much alive. Now half way through its pilot phase, the project is moving ahead with new organizations coming on board.

Yesterday, the Helix Nebula consortium held an event at the European Space Agency’s ESRIN facility in Frascati, Italy, to review the success of the project’s proof-of-concept phase. Helix Nebula aims to pave the way for the development and exploitation of a European wide cloud computing infrastructure. While this is initially based on the needs of IT-intense scientific research organisations in Europe, Helix Nebula intends to also serve governmental organisations and industry and, as such, will also reflect the needs of these stakeholders. According to the project’s website: “This pan-European partnership across academia and industry is working to establish a sustainable European cloud computing infrastructure, supported by industrial partners, which will provide stable computing capacities and services that elastically meet demand.”

 

“Helix Nebula is a partnership that was born out of a vision,” says Maryline Lengert a senior advisor in the IT department of the European Space Agency (ESA), a founding partner of the initiative. “We want to operate as an ecosystem. Today, the market is fragmented, but we want to bring it together and by doing so we will benefit from the stability of diversity.” To support this, Lengert highlights the fact that the Helix Nebula consortium has already grown from 20 to 34 partners in just the last year alone. Hans Georg Mockel, who is director of human resources, facility management and informatics at the ESA, says that the project has so far worked very well for his organization. “The proof-of-concept phase has demonstrated the feasibility of the [Helix Nebula] strategy and now we’re starting the next phase.”

“Space missions and their expectations have changed with the evolution of the internet,” explains Mockel. “People want to have access to large amounts of mission data immediately. We need to do it [dissemination] in an efficient, economical way and we want to do it with other scientific organisations in similar situations.” He says that the solution is for organizations to go beyond providing their own infrastructure and to federate things on a European level through the science cloud.

Flagships flying high

ESA, is working in collaboration with the French and German national space agencies, as well as the National Research Council in Italy, to create an Earth observation platform focusing on earthquake and volcano research. However, the maps created through this project can take over 300 hours of sequential computation time to complete, explains ESA’s Sveinung Loekken. “We want to put the processing of the maps onto the cloud, rather than on somebody’s workstation, which obviously struggles to handle it,” says Loekken. “We want to give people access to large data processing capabilities. This is the raison d’être of the scheme.”

This project is one of three flagship projects undertaken during Helix Nebula’s two-year pilot phase. Ramon Medrano Llamas presented findings from CERN’s flagship project, which has seen the organization gain access to more computing power to process data from the international ATLAS experiment at its Large Hadron Collider accelerator. This has allowed CERN the possibility to dynamically acquire additional resources when needed. “The proof-of-concept deployment has been very successful,” concludes Llamas. “Processing in the cloud clearly works.” Over the longer term, it is also hoped that use of commercial cloud resources could become a useful addition to very large data centres owned and managed by the scientific community.

Helix Nebula’s third flagship project involves the European Molecular Biology Laboratory (EMBL), headquartered in Heidelberg, Germany. Rupert Lueck, head of IT at the organization, explained how they have been working to establish a large-scale next generation genome analysis facility on Helix Nebula. “Biology and life sciences have a big need high-performance computing and for e-infrastructure which can cope with a large amount of data, says Lueck. “In order to understand DNA, we need to analyze sequences of up to three billion base pairs. It’s a lot of information and you have to make sure all the information is in the right place.” However, not all biology laboratories have the infrastructure to process this amount of data, explains Lueck. “That’s why we have need for fast, shared file systems,” he says. Many of the computational tasks carried out by EMBL researchers can also take a very long time to complete ­– often over a week. This dictates the need for the cloud storage to be very stable, as researchers certainly don’t want to have to restart these jobs if something gets lost. However, Helix Nebula has so far proven up to the task, reports Lueck.

Interdisciplinary innovation

“It’s not enough to simply produce data, we need to make sure data is fully exploited and that there’s an economic return in what we find,” says Thierry van der Pyl, the European Commission’s director of ‘excellence in science’. “Europe is in a difficult economic situation. This means, more innovation. And for science it means it’s important to better translate research into products and values… this is the driving force.” Pyl believes that the Helix Nebula project already provides a good example of exactly how this should be done.

Despite this, Pyl is keen for Helix Nebula to branch out into new fields in the future.  “How can we use the science cloud as ‘a lead market’ to stimulate development of cloud in Europe beyond science?,” he asks. “I urge you to work beyond silos and to find commonalities between disciplines. We need to avoid re-inventing the wheel each time, as we all too often do,” warns Van der Pyl. He concludes: “Helix Nebula is paving the way for working across disciplines and is helping to make Europe more innovative.”

Finally, Robert Jenkins, CEO of CloudSigma, one of the participating organizations on the supply side of the project, also spoke at yesterday’s event. He neatly summed up the situation for the project during his talk: “Helix Nebula is really trying to push the envelope in terms of what‘s possible with cloud and we’re dealing with some very thorny problems. That’s why we’ve come together to try to solve them.”

Laurence Field from CERN tells us how data is being managed for CRISP

Laurence Field is the CRISP (Clusters of Research Infrastructures with Synergies in Physics) IT & Data Management Topic Leader, based at CERN. He spoke in the Tuesday morning at the EUDAT 1st meeting in Barcelona, but the Gridcast team had already caught up with him before the meeting.

GISELA Latin American Institutions Contributed to it! (Higgs boson)

The CERN Director General has announced this morning (see below) the first observation of a particle consistent with the Higgs boson. The news is transforming to a snow ball being rolling all around the world in the TV and radio flashes.

The GISELA Community is particularly excited by the information. Indeed several Latin American Institutions, members of GISELA, such as UFRJ, UNAM, ULA, etc. are using its e-Infrastructure, in the framework of the Computing model of the LHC experiments in which they participate. The figure below shows the use of the GISELA e-Infrastructure by these institutions, for their various computing tasks such as simulation of physics processes, detector data processing and physics analysis. One should emphasize that Grid e-Infrastructures is a must for High Energy Physics experiments which could not otherwise deal with the sophisticated, CPU-consuming treatment of the data provided by their particle detectors. Hence it is really wonderful to observe, on a sensational case, the actual discovery power of Grid e-Infrastructures.

We take out of the event that Latin America made the right choice when investing, since more than 6 years in Grid e-Infrastructures, through the EC co-funded projects EELA (http://www.eu-eela.org), EELA-2 (http://www.eu-eela.eu/) and GISELA (http://www.gisela-grid.eu/). Finally we take the opportunity to underline the urgency to have the Latin American e-Infrastructure operation and its User support consolidated in the long term, to sustain e-Science over the continent, beyond its current outstanding outcomes.

Read the CERN Press Realease!

Strong hints of the Higgs - live from CERN

Good morning from the restaurant at CERN where I've been watching the announcement of the results from the CMS and ATLAS experiments at the Large Hadron Collider on the screens. These two experiments have been competing in the hunt for the Higgs boson particle, the missing piece from the standard model jigsaw. The announcement happened simultaneously at CERN, a bit further away on the site from where I was sitting, and at ICHEP in Melbourne (a loooong way from where I was sitting).

As an ex-CERNoise (as they call themselves) I was particularly interested in the announcement. I joined CERN in 2008 to run communications for the Enabling Grids for E-sciencE project, the grid computing infrastructure project that is now the European Grid Infrastructure (www.egi.eu). EGI, in collaboration with the World LHC Computing Grid (WLCG) provides the distributing computing infrastructure that has processed the petabytes of data streaming from the four LHC main experiments. It was good to hear a mention of the contribution that computing has made to this momentous global achievement during the talk from Joe Incandela of CMS and Fabiola Gianotti of ATLAS - by reducing the time taken to process the individual events, the results can be announced quicker. No small achievement when the US Tevatron is snapping at their heels http://www.bbc.co.uk/news/science-environment-17269647

While I was at CERN, the LHC started up in September 2008 to astonishing media and global interest... and then two short weeks later experienced a dramatic quench in one of the magnets deep underground, causing serious damage to the machine and a shut down in running of over a year. The mood at CERN switched from what probably counts as wild excitement in physicist terms (well, we had a staff party, and there was dancing of a sort) to one of pragmatic problem-solving and a collective rolling up of sleeves. The LHC has been a long time in the building, with many many problems to solve along the way  - here was just another phase to plan through. Just over a year later in November 2009, the LHC came back more stable and reliable than before and has since run pretty much flawlessly, ramping up the energy to where it is now at a record breaking 8 TeV.

This morning I was back at CERN for a different meeting but happened to be in residence on site in time for the joint announcement from CMS and ATLAS. I watched the auditorium fill up with students, physicists and a good selection of the world's media. As Peter Higgs arrived to take his seat, a round of gentle applause broke out. After a bit of milling about, a clearly rather nervous Joe Incandela from the CMS experiment took the podium to run through his slides. Following the Twitter feed online, these caused a certain level of bemusement among the Twitterati - the general thrust was I don't really understand this, but it sounds good. Despite working at CERN in the past, I was kind of with them, but the punchline came out pretty clearly at the end of Incandela's talk: We have found a new boson with mass of 125.3  GeV +/- 0.6 at 4.9 standard deviations or "sigma". A loud and extended round of applause in the auditorium followed. To put it in context, one sigma means the results could be random fluctuations in the data, 3 sigma counts as an observation and a 5-sigma result is a discovery. It is about a one-in-3.5 million chance that the signal they see would appear if there were no Higgs particle - so extremely unlikely. However, the CERN press release (http://press.web.cern.ch/press/PressReleases/Releases2012/PR17.12E.html) reminded us that the results presented today are preliminary, as the data from 2012 is still under analysis. The complete analysis is expected to be published around the end of July.

Fabiola Gianotti of the ATLAS experiment followed next. "It's tough to talk second," said Fabiola, "because all the clever things have been said." She then went on to very much prove that wasn't the case with a highly technical presentation (despite the comic sans font, which did not go down well online). However, when she got to the part about zooming in on the data... more spontaneous applause starts up (with a slightly ironic echo here in the restaurant). The ATLAS experiment results echo the CMS results - 126.5 GeV for a boson at 5.0 sigma, in agreement with the standard model of physics. In a personal statement, Gianotti said "It is very nice for it to be at this energy because we can measure it here at CERN in abundance – thanks nature!"

Rolf Heuer, the Director General of CERN had a final say. "It's been a global effort, a global success. It has only been possible because of the extraordinary achievements of the experiments, infrastructure and the grid computing." He continued, "We have a discovery, we have a new particle consistent with the Higgs boson. It’s a historic milestone today but we are only at the beginning. We can all be proud and it has global implications for the future. I think we can be very optimistic." Peter Higgs added his congratulations and announced himself "extraordinarily impressed". I suspect he isn't alone.

So a highly significant day for CERN, for the hunt for the Higgs and for me, as its pretty satisfying to be at CERN both for the LHC start up, and the announcement of the strongest hints of the famous particle so far. It's 28 years since the idea of the LHC was first suggested at a physics meeting, more than half my life time ago - I feel very lucky to have been here at this time to see the results of the work of over 20 years, thousands of people worldwide, not to mention six Director Generals. Congratulations!

Tell the truth: tell it first, tell it all

Image of particle collision from DESY Hamburg

Revisiting my notes on the ERF workshop on the socioeconomic impact of research infrastructures, held at DESY in Hamburg, I’ve had a chance to reflect on some of the positive aspects of the second day of talks. Of course, I’ve had a couple of weeks to think about it! In my previous post from ERF, I drew attention to the difficulty in measuring ‘socioeconomic impact’ as though it were a single entity. The easily quantifiable economic aspect tends to take over, at least at first. So we have to dig that little bit further to get to the qualitative measure of the woolier social impact.

‘Socio-’ itself can be broken down into several subspheres: social, of course; also the environmental and educational. We heard an extremely interesting talk from CERN’s head of communications, James Gilles on Living in the Media Spotlight. CERN fit the ERF membership criteria perfectly, a main stipulation of eligibility to join being that a member research centre should do its science (and everything else) ‘out in the open’. Social media, explained Gilles, had shifted the spotlight away from communications offices, such as the one he leads, onto the scientists themselves. At CERN, and other ERF member institutes across Europe, this is a trend that continues to gather pace. As for doing everything ‘in the open’, that includes mistakes. CERN had observed that, when EU taxpayer money is at stake, there’s only one mantra that counts: ‘Tell it all, tell it fast, tell the truth…’  – even when that truth meant explaining why the LHC broke so soon after it was switched on.

One of the most exciting developments e-science in recent years is citizen cyberscience. Projects such as Fold It, Galaxy Zoo, and now Moon Zoo (and others) have shown that the general public can make a real difference when it comes to big science. Graham Higley, Head of Library and Information Services at the London Natural History Museum explained how natural history was the perfect pairing for citizen science projects. NHM have run several scientifically important citizen cyberscience projects in recent years: tracking the influx of invader bluebell species; assessing water quality through river fly numbers; counting bugs (surely one to excite children of any age), and crowdsourcing envronmental science engagement. Bioblitz was a case in point, finding new species in Alexandra Park, North London. Web technologies have allowed members of the public to make useful contributions to science, but its down to forward-thinking organisations like the NHM to take advantage of them, and be brave enough to make the point that participation in the business of science is for everyone. National research infrastructures that follow the lead of the likes of CERN and the NHM really are having a positive socioeconomic impact, and are prepared to acknowledge that the ‘socio-‘ component is every bit as important as the economic factors when assessing how they’re performing.

Volunteer Computing to Volunteer Scientific Development

Ben Segal presented today their talk "Building a Volunteer Cloud" in the CLCAR 2009. Prof. Segal discuss about the opportunities of Volunteer computing to build realistic possibilities of High Performance Computing, addressing Cloud Computing. Using BOINC and Virtualization in a Cloud context is possible to investigate the use of either emulation or virtualization for BOINC, driven by the porting considerations just described. We quickly showed that emulation was too slow but that virtualization entailed only a few percent of CPU overhead and looked very promising, given the scale of the potential volunteer base.

On the other hand, Segal's talk presents BOINC and virtualization technology as a great opportunity to regional and multidisciplinary projects in LatinAmerica showing good examples of developments in Asia, Africa and Europe. Prof. Segal was made an extended call for motivation and participation in BOINC projects.

Thoughts from the coordinator of the worldwide LHC Computing Grid

I recently interviewed Jamie Shiers, coordinator of the worldwide LHC Computing Grid at CERN, about the STEP'09 postmortem – a review of the large scale STEP test of the wLCG in preparation for the LHC startup. The bulk of what he says is reported in an iSGTW article.

I’ve also made an audio clip of all the other good stuff Jamie said which there wasn't space for in the article – his views on the importance of grid computing, people, collaboration, and BBQs for the LHC. Have a listen!