http://projects.oucs.ox.ac.uk/lowcarbonict/conferences/conf-1.htm.
There were presentations on:-
- Low carbon ICT in context of Oxford University
- Low carbon approach in large hospital using virtualised servers
- A wonderful overview of reducing carbon in data centres
- Greening the full lifecycle (from manufacture through to recycling)
- Facebook meets green business
- Green technologies in the pipeline
Just 0.5% of the energy of the fossil fuel is actually used to deliver power for computing.
When Liam's slides become available on the above web address, take a look and send your comments.
There is considerable scope for increasing the percentage if useful power used.
The Green Grid (http://www.thegreengrid.org/home) is a very interesting site. "The Green Grid is a global consortium dedicated to advancing energy efficiency in data centers and business computing ecosystems."
The conference agreed that the low carbon ICT area is going to grow and grow in importance.
8 comments:
This has indeed been a useful and throught-provoking conference. Some extremely serious issues have been highlighted and more than ever before I will be thinking carefully about energy use when making decisions or giving advice on IT provision and support.
The difference between energy actually used in manufacture of a system (embedded energy) and the energy it uses during its life is an important factor that has been well-highltghted today also.
A related Press Release which describes an Oxford project to reduce power use by desktop devices is at: http://www.physorg.com/news125068589.html
I was entirely enthused by the presentation given by Liam Newcombe (http://dcsg.bcs.org/). As you know we are in the early planning stages of both a new University data centre and the refurbishment of our existing computer room. Having attended this conference I think we should be more bullish in making energy efficiency the primary driver for both the refurbishment and new-build. Appropriate levels of resilience, the traditional driver, we have already concluded will now come through other means (not least, having more than one data centre).
I am sure that our University, in common with other institutions undergoing similar projects, would benefit from the fresh thinking ('crop rotation' for example) which Liam epitomised.
Our own energy costs associated with the Computer Room have tended to remain background costs. I think we should start to foreground and address them. Attempting to apply a more holistic approach. So, whilst we can champion the data centre projects as part of the University's sustainability strategy, we should also, for example, be including energy costs in the TCO of any procurement of enterprise systems and other significant hardware components (I suspect it's rare even to calculate and include any increased energy costs associated with new services in the recurrent budget).
The Computer Room refurbishment project can concentrate on energy efficiencies associated with the M&E whilst at the same time we can continue to investigate policy for virtualisation/grid platforms. The new build should be able to take a more holistic approach from the beginning, agreeing the standards and policy for hardware deployed in the new facility.
I very much enjoyed the conference and this is largely due to so many people making excellent contributions, from the speakers, companies that hosted stands, and the delegates who engaged enthusiastically with the presenters and each others questions. I've rarely seen such, dare I say it, 'passion' in an ICT-related conference, the goal of green or sustainable ICT really seems to gel a wide range of expertise into a common cause. The mixture of physical and computing engineers, service providers, energy experts lead to a prolific surfacing of ideas and thoughts.
While I have hardly had time to reflect, the following thoughts have occurred to me:
1. It should be possible to calculate how many CPU cycles the planet has enough energy to support through sustainable energy supplies. Of course this will be a moving target as we innovation continues in chip design, cooling, automation of virtualisation and of course energy supply
2. There seems to be little point in cap and trade of the 'IT industry' since IT is infrastructure across all industries. The pressure on IT providers will be from the business that they support so for OUCS it will be from each department, or the University itself. The crucial point though will be that IT providers will have to measure how much energy a 'service' consumes.
3. On this last point, it was mentioned a lot that IT service providers tend to put robustness above energy conservation so over-spec hardware/ software. As a solution some delegates proposed that IT departments should pay for the energy they use. I don't think this is the real issue; the real issue is whether IT departments can account for the energy consumed by a specific service.
4. It seems that one of the best ways to improve the efficiency of data centres/ machine rooms is to site the hardware close to a green energy supply. While Canada and Iceland are being reported as ideal for this purpose, I wonder where this will be in the UK. I’ve heard one of the biggest builds of nuclear power station is planned for oxford but I wonder whether its possible to create enough solar/ wind/ geothermal energy to power the most efficient type of data centres.
5. One of the most elegant things about an automated virtualised data centre is the ability to literally move processes between data centres. This could be employed to chase the 'greenest' energy supplies around the country/ world. I’ve heard that Google already do this to chase the cheapest energy...
6. If we can monitor and report on the cost of a service, then inevitably a value for this service will emerge through use. This will result in a service marketplace evolving. This seems anti-web which still has the 'free' aura surrounding it but just like there is a limit to the number of cars that can fit on the roads; I think we will reach a limit to the number of services that we can supply with sustainable energy. This is a calculation that needs doing however. this point fits in with point 1 above, I think it is helpful to produce a top end figure so that people realise there the planet does have a limit to how much computing it can support (in the same way that we're learning there's a limit to all natural resources e.g. oil, water).
7. One strong theme that emerged through the conference is the need for suppliers to calculate the embedded energy costs of their manufacturing processes. To make this happen it seems that a regulator will need to monitor the self reporting by suppliers.
Anyway, as I said, I want to spend some time going through the presentations and doing some more research. I've had some ideas for the next conference, which many delegates have started to ask about. I'm thinking the next event needs to have quite a different format, and most certainly use technology to do as much 'work'/ learning as possible before the actual f2face meeting. I'd welcome ideas on this subject.
One small point: where people here are using the term "embedded" energy they should probably substitute it with "embodied" energy. "Embedded" is usually used in the context of e.g, "embedded microgeneration in a building", etc.
See here or here
Post a Comment