Distributed Thinking

Ian Foster notes a project at University of Notre Dame near Chicago which is distributing research computing facilities in order to provide heat to campus buildings. This is an interesting trade-off: on the one hand, the distribution of resources means that the waste heat is put to good use; on the other hand, it's possible that each distributed installation is less efficient that a good-quality centralised machine hall. This is another example of where we need good quality models and well-measured example deployments to help us decide which approaches give the best results.

This week, I attended Healthcare Computing 2008 to get an update on the current state of e-health in the UK and to explore how grid technology can contribute. Health Informatics is a broad subject and it isn't possible to engage with the whole field, but I see three main areas of potential engagement. The path is most followed by the academic community to date is that of linking together data used in clinical trials or in health research. This is a natural fit for the e-science community as it extends existing work on secure access to distributed research data. The medical world imposes more security constraints, which adds academic interest, but is otherwise familiar to the e-scientists. It is also of a scale that is manageable in research projects. Successful projects include Psygrid , which is now deployed across the NHS research centres in mental health and in bioinformatics. So this is the first area of engagement. A natural question is whether this experience with rese...

There has been some rather confused coverage in the press about the grid infrastructure that supports the Large Hadron Collider at CERN. The Times has an article that cast the grid as a "superfast internet", with the emphasis on the high-bandwidth links that have been laid to support the LHC data dispersal. It also talks about the numbers of servers connected to the LHC grid, but without clarifying the distinction between bandwidth and processing power. It also implies that the LHC grid is the only grid, which perhaps we can forgive the journalists for, as plenty of technical people still refer to "the grid" as if there were only one. A Yahoo article , taken from Sky News , goes rather further, claiming that "the internet, as we know it, could be obsolete within a decade". The phrase, "as we know it", lends a wonderful vagueness to the claim. The article goes on to say that the Grid was the brainchild of CERN, which of course is an exaggera...

Ian Foster pointed readers of his blog at a good article about CaBIG, the Cancer Biomedical Informatics Grid. This major project is led by the US National Cancer Institute and also involves Cancer Research Labs in the UK, with the aim to share data between cancer researchers. The UK's OGSA-DAI system is a major component of the deployed system. It's particularly interesting to note that the system includes support for clinical trials. Clinical trials are time-consuming and expensive, so many people want better systems for managing them and a number of grid projects are tackling this area. PsyGrid is one such project in the UK. They're not exactly blowing their own trumpet about this, but their system is being used as the studies and trials platform for the Mental Health Research Network, and has been selected by the National Institute of Health Research (NIHR) as the platform for providing electronic data collection services to support studies and trial across the ...

I am very pleased with the line-up for our webinar on May 7th. We will continue our investigation into best practice for running energy efficient data centres, following our successful webinar last October. This time we will have an emphasis on the impact of virtualisation and on how one can model and measure the effectiveness of proposed improvements. Nic Barnes will explain what Merrill Lynch observed when they applied virtualisation on the desktop and in the server room. They found that the benefits were in some cases partially offset by losses elswhere. Nic will demonstrate the importance of measuring real gains in practice. Liam Newcombe will describe the work of the BCS in developing a model for predicting data centre efficiency. This approach will allow managers to plan and evaluate designs in advance of their implementation. Liam will show that the choice of metrics requires careful analysis. As always, viewers will be encouraged to ask questions. Any that we can...

Virtualisation has obvious benefits for much commercial IT, where existing servers often have utilisation rates of 10% or less. It's less clear whether virtualisation is so useful for High-Performance Computing (HPC), where systems are often kept running at utilisation rates above 90%. Nevertheless, there are potential benefits from adopting some aspects of virtualisation in these environments. The question is, do the benefits outweigh the cost in performance? This was the topic of yesterday's workshop on System Level Virtualisation for HPC (which was part of EuroSys 2008 ). The workshop was rather small but I did learn quite a bit. A couple of talks investigated the performance hits of using virtualised OS's instead of addressing the hardware directly. This varied, depending on the type of application; if IO was minimal, the slowdown was minimal too. An "average" slowdown seemed to be on the order of 8%. Samuel Thibault of Xensource looked at ways of usin...

Licensing has been an ongoing issue for virtualising and grid-enabling applications. Many licensing models do not transfer easily to a world where applications and virtual servers can be run on many processors and moved from machine to machine. Vendors want to ensure that they are paid for the full use of their applications, while IT managers want to know how much they can expect to pay in a given accounting period. Everyone is aware of the problem and several people have been trying to work out ways forward, but the issue is still unresolved. Grid Computing Now! are bringing together some of the key players in the industry for a webinar on Thursday 10 April at 14:00 GMT: David Gittins of Capgemini, who has been working with the public sector on this challenge Neil Sanderson, Product Manager for Virtualisation at Microsoft Mark Cresswell of Scalable Solutions, a leader in the provision of tools for monitoring and reporting usage. Ian Osborne, Project Director of Grid Computin...

Yesterday I attended the Low Carbon ICT Conference organised by Oxford University, with funding from JISC. This was a good an useful day, with speakers covering a range of topics, and a small number of exhibits. Some of the speakers were familiar - Zafar Chaudry gave his excellent talk on the benefits of virtualisation in one of our webinars , while Liam Newcombe will be speaking in our next-but-one webinar on May 7th. Others were new to me. The speakers' talks covered office ITC equipment, virtualisation, data centres, PC manufacturing, sustainability entrepreneurship and enterprise planning. I was more at home with the techie end of things, but this is not a problem that can be tackled by technology alone so it was good to see the broad participation of this conference. I was aware of most of the technical issues, although I was surprised at how much energy is used by desktop PCs and peripherals (see below). I was also interested to hear Juergen Heidegger recommend the use ...

"This is the Century of Information" - G. Brown, November 2007. This week, I attended a think-a-thon about a strategy document that the e-Science community in the UK is developing. The goal is to put in place the right mechanisms for ensuring that UK research can make the most of new computing technologies and methods. We have many success stories from the e-science programme; the question is, how do we build on those successes and make the techniques available to everyone? One point that arose from the workshop is that we need different types of successes. Most of the examples put forward were of good research enabled in a range of domains (GeoSciences, BioScience, Chemistry, Physics, Social Science, etc.). We also found examples of advances in Computer Science itself, rather than just using CS to support other fields; this is essential if we are to engage CS academics. Beyond academia, we need examples of knowledge transfer to industry. This is where the Grid Computi...

With the twin growth of supercomputing demand on the one hand and energy cost on the other, it makes sense for companies to produce supercomputers that use less power. IBM started the trend with the Blue Gene series, of course, although that project was nore about building the fastest computer while keeping energy consumption somewhat reasonable. More recently, others have joined the fold. One such is SiCortex , who claim to have designed their systems from the chip level up to minimise power consumption. Their SC5832 machine offers 5832 1GFlops 64-bit processors for 20kW, while for the SC648 they claim half a teraflop powered from a standard wall socket. That's not the only approach, of course. Floating-point accelerator boards and Graphics Processing Units are being used to boost computing power for specific applications while keeping costs low. The Register has a good overview article from November's SuperComputing conference. Meanwhile, in a talk at the Mardi Gras C...

This blog posting is an oddity - Bill St. Arnaud suggests a micropayment scheme whereby consumers pay for broadband services by increasing the cost of their household electricity (or car mileage) for a certain duration. The utility companies then pay this amount to the broadband service provider. So the service provider gets income from a source other than advertising; the consumer gets reasonably priced product, and the consumer reduces their energy usage because the price has gone up for that period. I don't see it happening, myself, but it's an interesting piece of lateral thinking.

In order to expand our "community of practice" on energy-efficient data centres, we have visited several people or groups who have an interest in the topic. Just over a week ago, some NeSC colleagues and I visited BRE's Scottish office . BRE used to be known as the Buildings Research Establishment and the folk at their East Kilbride office are particularly interested in sustainable development. So far they've been mainly working with housing and small businesses, but they seem potentially interested in modelling larger establishments. They are also interested in multi-level modelling of heat flow within buildings; something our e-science connections could definitely help with. We also had a good conversation about DC power circuits. This week, I was invited to speak at a meeting of the Russell University Group IT directors ( RUGIT ) - i.e. the people responsible for ICT at many of our leading universities. I presented some of the outcomes from the HTC week in N...

This week, the e-Science Institute launched a new research theme which should be of great relevance to industry as well as scientists - in fact, it may even help ordinary consumers to protect our own privacy online. The theme is about "Trust and Security in Virtual Communities". Andrew Martin , the theme leader, explained its aim in a webcast talk . The problem that Andrew is exploring is how we can trust a grid infrastructure to protect our sensitive data. In addition, how can we trust the results that we get back from running a job on "the computing cloud"? To give one concrete example, Andrew was involved in the climateprediction.net project, which encouraged people to contributed their PC's spare cycles to run climate modelling simulations. This raised several security issues. From the users' point of view, could they trust that the climateprediction.net program would not hijack their PC? Conversely, could the scientists trust that the data sets r...

I'm very pleased to say that our Knowledge Transfer Network, Grid Computing Now! , will continue to operate for at least another year. We have just received final confirmation from the newly-reconstituted Technology Strategy Board . This is welcome news; it means that we can continue our plans to bring users, vendors and academics together to address real problems in several sectors. So it seems a good moment to reflect on the state of KTNs and how they might develop. Innovate07 was the showcase for all 22 Knowledge Transfer Networks. This was my first time at Innovate and I was impressed by the range of technologu areas and delegates. It was also a good opportunity for networking between KTNs, which has led to some joint initiatives. This range of KTNs is in part a branding exercise, as some KTNs had previous existences as Faraday Institutes or other institutes. So we at GCN are in the odd position of being one of the first KTNs to be set up and at the same time among the o...

I delighted to see that Internet Villages International are building a major data centre to be powered entirely by "green" energy in Scotland. This is exactly the kind of development that some of us have been arguing for. The centre will be built next to a source of renewable energy and the waste heat is intended to be used for a new village and for local horticulture. This makes sense in many ways. Data centres do not need to be near the businesses that use them (think of how often you use Google and where their data centres are). Placing them near sources of renewable energy saves on transmission losses; indeed Google, Amazon and other firms are already doing this in the USA. Also, data centres produce a lot of heat and it makes sense to use this rather than waste it. Let's hope that the operators minimise the energy they do use as well, for example by running virtualised servers, building modular UPS and cooling systems, and using external air intakes rather tha...

We;ve just finished a week (well, four days) of talks, tutorials and discussion about High Throughput Computing. The event was opened by Miron Livny, leader of the Condor team, who gave an excellent introduction - the key point is that HTC is about the number of tasks that can be completed in a given time, whereas "traditional" High Performance Computing is about how much computing power can be brought to bear at a given time. As Miron puts it, Floating Operations per Year is not necessarily 60*60*24*7*52 Floating Operations per Second (FLOPS). We've hosted events by the Condor team in the past, but for HTC week we extended our range. In particular, John Powers and Dan Ciruli of Digipede flew over from the Bay Area to tell us about their product. A day of hands-on tutorials allowed delegates to compare the strengths of Digipede and Condor, and the evening discussions included ways the systems could be used together. Scheduled discussions looked at requirements for ...

Last week the e-Science Institute organised a "think tank" to review the state of e-Science and suggest opportunities for research. A major emphasis of the debate was the recent trend to use Web 2.0 tools to support scientists. Dave de Roure gave several examples he saw at recent conferences, including wikis and blogs such as Open Wetware and Useful Chemistry , as well as various data mashups. Tony Hey gave a public lecture on e-Science and Digital Scholarship which presented a similar story, including the use of utility computing (which now seems to be called cloud computing - you've got to love the constantly changing buzzwords in IT). Among the discussions, people mentioned the combination of Web 2.0 tools with semantic web technology, and the combination of structured queries and semi-structured information as in DBpedia . This growth of e-Science 2.0 (to coin a buzzword of my own) has mainly seems to have occurred largely in the life sciences, perhaps becaus...

The media and online world are buzzing with the news that HMRC have lost discs containing financial details of 25m people. My particular interest is to what extent the centralisation of the database contributed to the problem. If we consider the NPfIT programme for storing medical records, would they be safer in distributed data stores? At first glance, one might think that a security breach in one store would at least be limited to the set of data held there. But those distributed stores would have to be networked and to allow remote queries; would this increase security (by checking for mass requests) or decrease it (because people wouldn't know which remote requests to distrust)? In the meantime, here are a few URLs to comments that I found interesting. Philip Virgo worries (as do I) about those organisations that cover up their breaches rather than report them. David Lacey argues that certification of security practices is required to make sure policies are followed. ...

"Green IT" is a hot topic at the moment. In the UK, data centres contribute nearly 2% of the country's CO2 emissions, a figure which is similar to that of the much-vilified airline sector. We recently broadcast a webinar on this subject which went very well - several people have said that it was our best webinar yet. We were fortunate to have two excellent speakers. The first was Zahl Limbuwala of the BCS Specialist Group on Data Centres. This SG has developed an open source model for measuring server room efficiency, which will be published in January; Zahl presented the case for why energy efficiency is relevant and talked a little about this model. The second speaker was Kate Craig-Wood, who runs a carbon-neutral hosting company. Kate gave a lot of low-level practical suggestions for improving the energy efficiency, which her company has used in their new data centre. There are obvious advantages in making such improvements, both economic and environmental. The ...

One of the best talks at AHM2007 was on medical imaging, by Prof. Anders Ynnerman of Linkoping University in Sweden. Imaging researchers always have an advantage when giving talks because they can show much of their work in pictures; Anders took full advantage. The immediacy of visualisation in communicating information meant that some viewers found his medical pictures a little too gruesome. The computer science aspect of his talk was about techniques for 3-D reconstruction of body images from the 2-D “slices” taken by medical CAT and MRI scanners. As these scanning devices increase in resolution, so the size of data to be processed increases. In order to process the data, ZZX is using intelligent compression techniques, based on whether a voxel represents bone, blood or tissue. Aside from the technical content, I was very interested in the uses XXX has found for this work in addition to the usual pre-operative medical briefing. His team are now able to help the police by perfo...