Thursday, November 10, 2011

Cracow Grid Workshop 2011 - CGW2011

I have just returned from a very stimulating and rewarding three days at the 2011 Cracow Grid Workshop. Apart from discovering that Cracow is further South than Prague, just, I also learnt more about grid-related initiatives that relate to the Polish research community. I was pleased to see a good turn out to all the sessions, to see a wide age range with a strong new generation embracing distributed computing and, most importantly, an attentive and questioning audience.

The ongoing discussion about cloud versus grid computing paradigms surfaced in a number of areas with a clear dominance for the cloud model. Personally, I find this interesting but feel that it should not become the dominant debate. As we continue to embrace virtualisation as a mechanism for introducing greater flexibility to the way in which services are delivered we should be able to concentrate on delivering the right services to the right researchers rather than having philosophical discussions about infrastructure.

The other keynotes and talks covered a diverse spectrum of research areas. A number of talks presented WLCG and the Polish contribution to LHC work. Life Sciences was well represented too both at the level of the Life Sciences Grid Community (LSGC) which has been working closely with EGI over the last year and also we heard from the Virtual Physiological Human community. A number of other EU-projects with significant Polish involvement such as UrbanFlood and also the International Desktop Grid Federation were also featured.

Having been fortunate enough to be invited to participate in the jury to judge the posters I had the opportunity, not to say obligation, to study them in some detail. The standard was high so I felt that I came away feeling that I had acquired a good overview of activity in Poland. Posters are important and well worth the effort that goes into producing them; they are a wonderful source of information for others and an opportunity for researchers to rethink the goals and purpose of their research. We at EGI should make more use of these - perhaps more competitions and bigger prizes at our forums?

The strongest message that I took away from CGW11 was that whilst there is much good work being done, too many projects are not working in a sustainability-focused way. Within the context of EGI as a whole we have been promoting the move to a tiered approach modelled on the cloud paradigm with the intention of clearly separating the layers of infrastructure, core middleware services and then user services at the top. Many of the projects that I saw were re-creating elements from all of these layers. Whilst this works at a proof of concept level and enables research to be done in the short term; in the medium term the costs for sustaining all of the elements are high. EGI, that is the organisation in conjunction with the relevant NGIs, need to analyse these successful projects and identify the key elements and migrate these to the relevant tiers in the infrastructure. Typically, there will be elements that already exist and so we can then focus on the original elements with the greatest potential for impact in the research domain and add these to the production infrastructure.

I am pleased to report that the workshop included a wonderful conference dinner at a characterful micro-brewery restaurant in the old city in the centre of Cracow. Always a useful opportunity to push forward discussions from the earlier sessions as well as learn more about the rich history of this fascinating part of Europe from the local hosts.

No comments: