I was recently at the NGS (National Grid Service) Innovation Forum 2008 to find out what existing users of the NGS were doing and to see what the reaction was to future plans for NGS Phase 3. The first, very encouraging, point was that there were more users there this year than there have been for previous years. Secondly, these users were more diverse, with representation from researchers, e-science centres and support functions for researchers such as IS and research computing directors.
Day one started with presentations from researchers in biology and physics biomolecules with representatives from other research areas being amongst those at the event. It has been particularly encouraging for JISC as a funder to see this transition of the NGS from providing resource predominantly for those in the ‘hard’ sciences such as physics and chemistry to greater provision for those in social sciences and the arts and humanities. One message that remains, however, is that if the NGS is to get more users from a wider range of disciplines then they need to offer alternative methods of accessing the service to the command line and these need to be easy to use. The benefits are very tangible, with one presentation reporting that modelling time had been taken down from one month to six hours.
Michael Wilson then described how EGI (the European Grid Initiative) could involve the NGS amongst others and sparked off a very lively debate on who would take the NGI (National Grid Initiative) role for the UK that was required by EGI. Whilst the UK and other countries have expressed an interest in EGI there is still no firm commitment and Michael’s talk stressed that EGI was only a co-ordinating body for European provision of grid infrastructure, not a funding body for national facilities, as has previously been the case with bodies such as EGEE. This meant there needed to be national commitment to ensure that the UK was appropriately represented.
From the European perspective we moved to Daniel Katz’s presentation on TeraGrid, the American national grid. There were a number of points that were particularly notable in the presentation, out of which the most interesting one was the concept of Campus Champions. Campus Champions help promote TeraGrid and grid usage within their campus in exchange for attendance at TeraGrid meetings and a t-shirt! More to the point, they are people who would like to encourage grid usage and work with those who are new to the grid to help them carry out their research more quickly or simply do new research. It is something that we see happen on an ad hoc basis in the UK but gives food for thought on how we get phase 3 of the NGS to encourage new users. Also of interest for me, with my access management hat on was TeraGrid’s experimental use of InCommon to access grid resources.
After lunch, the programme moved onto grid technologies. There was a good section on Condor for managing campus grids. Whilst there is often not much attention paid to grids within an institution they form a vital part of the infrastructure available to researchers. Hugh Beedie also pointed out that they could be a very effective green alternative to high performance computing, especially given modern machines’ power efficiency. Next up was a session on Clearspeed from Steven Young. He described how there were four of these maths acceleration cards that now feature at the Oxford node of the NGS. At this stage, there isn’t much use of them but they look promising for jobs that are maths intensive.
The day finished with presentations on the training available on the NGS (from David Ferguson) and Andy Richards talking about NGS Phase 3. Both provoked lively debate from the audience and there was a great deal of interest in David’s offer to run training on a regional basis so if you couldn’t attend the event and you read this then get in touch directly with the training team and find out about courses at http://www.nesc.ac.uk/training/.
Day two was a chance to tie up with the campus grid SIG and to look at what the experience was for those who had joined the NGS. The overall conclusions seemed to be that whilst it wasn’t easy to set up the software, the NGS had a very active support community that made the whole process a easier and that there were tangible benefits from going through that process. This led into how to make the NGS sustainable, which follows the general trend with projects in JISC that are moving to be a service. It was a topic that received a good deal of audience feedback and I am hoping that this can be followed up after the event as it is not going to be an easy task keeping access as easy as possible whilst making sure that institutions are appropriately recompensed for what they contribute.
The day finished with presentations on new directions for the NGS. Keir Hawker went through what data services were on offer, with a range of options from Oracle through to MySQL. Mike Jones then went through how the SARoNGS project was working to allow users who were members of the UK Access Management Federation to get access to NGS resources.
So, what were the key points to take away from the meeting? I think they were:
- Research is global and the grid offers a good way of working collaboratively within a trusted infrastructure. It will be interesting to see how this ties into ongoing work on interfederation and virtual organisations in the identity and access management area;
- The NGS has a great deal to offer the researcher and they are very keen to engage with active researchers to help them carry out novel research or to make what they do more efficient;
- There are no doubt potential users of the NGS who could benefit enormously from using it so it is well worthwhile attending a training event or one of the e-Research Roadshows to find out more;
- Whilst there are resources to try the NGS that are free at the point of use, this model will not scale infinitely so there need to be equitable models for sustainability;
- There is a growing community of researchers from an increasingly wide range of disciplines but there still needs to be a focus on growing that further;
- The institution needs to get involved in helping its researchers access grid facilities as more and more research is collaborative in nature. This is not just providing access to the NGS but includes grid resources on campus so that researchers have a range of resources available to them;
All in all, it looks to be an exciting future for the NGS. The next major decision point is whether approval is granted by JSR for the Phase 3 proposal. My thanks to Andy Richards and the team at NGS for a great event and inviting me along and paying for my accommodation.