Data Centres Europe 2011 Nice

Data Centres Europe 2011Last week I attended Datacentres Europe 2011 that was held at 5-6 May 2011 in Acropolis, Nice, France and organized by Broadgroup. In its seventh edition, this European summit is for senior IT executives, analysts and vendors to network, discuss and learn about the latest developments in data centers. A great event, featuring 85 speakers with 550 delegates from about 30 countries.

In his opening address Aaron Davis of Schneider Electric stated that data centers are vital to tackling environmental issues. He said that “We must decouple our growth from our energy consumption.” Mr Davis also said that, by centralising data processing, data centres were increasing energy-efficiency. In coming years, technology had to decrease its carbon impact by a factor of 10 to offset the effects of population growth.  Aaron Davis ended with the statement that “Data centers will be a major part of the next stage of human evolution.”

In the presentations and the panel discussions there was much emphasis on the following topics:

  • Energy consumption, Sustainability and Green in the Data Center
  • Cloud computing and the impact on Data Centers
  • Data Center business models
  • How to build flexible, scalable and modular Data Centers

Like last year a lot of speakers made a note or remark in the site line that there are still gaps between IT and Site  infrastructure people and that it is mandatory that those groups are aligned.

Also some explicit remarks were made about the consequences of the enormous growth in mobile/handheld devices such as smart phones, tablets and the lot. Referring to the fact that Metcalfe’s law states that the value of a communication network is proportional to the square of the number of connected users of the system (n2). Therefore the value of these devices for people will continue to rise and therefore also the need and usage of these devices. The data center will act as the man in the middle, because all lines of the network comes together in the data center. This will give an enormous rise in the demand for data centers.

Some speakers made the observation of the mutual interest between data centers and the smart grid a topic that will certainly get more attention in nearby future.

Notable missing discussions were e-waste  and data center life cycle management a form of Energy Accounting, the approach of measuring and analyzing energy consumption of the total
life cycle from the data center conception, through design and manufacture, to service and disposal.

The first day of the conference also the European Commission Code Of Conduct Special Awards were announced by Paolo Bertoldi of the European Commission Directorate General Joint Research Centre, Institute for Environment and Sustainability. The European Commission’s Code of Conduct (COC) on Data Centers Energy Efficiency is a voluntary code that organisations can endorse, committing them to implement measures to reduce energy consumption. The code is now established as the definitive industry practice for reducing carbon emissions by Data Centers and has been widely adopted with more than 122 endorsers.

As follow up the Data Centres Europe 2011 Awards will be hosted the second of June at the Gibson Hall in the City of London.

Cutting energy bill of data center by millions

Also in the digital economy location matters, but by moving your workload you can save money. The location of your data center can make a difference in energy costs. “So can this be the starting of moving your IT workload, on the virtualized infrastructure, to the place where the energy price is the lowest?” was the end of the blog entry “Power usage and money savings” earlier this year  about the huge price differences in electricity.

Now some very interesting stuff came up.

A recent study from researchers at MIT, Carnegie Mellon University, and the networking company Akamai suggests that such Internet businesses could reduce their energy use by as much as 40 percent by rerouting data to locations where electricity prices are lowest on a particular day.

Asfandyar Qureshi, a PhD student at MIT, first outlined the idea of a smart routing algorithm that would track electricity prices to reduce costs. This year, Qureshi and colleagues approached researchers at Akamai to obtain the real-world routing data needed to test the idea. The team then devised a routing scheme designed to take advantage of daily and hourly fluctuations in electricity costs.

“We introduce Power-Demand Routing (PDR), a technique that redistributes traffic between replicas with the express purpose of spatially redistributing the system’s power consumption, in order to reduce operating costs. Cost can be described in monetary terms or in terms of pollution. Within existing Internet services, each client request requires a meaningful amount of marginal energy at the server. Thus, by rerouting requests from a server at one geographic location to another, we can spatially shift the s marginal power consumption at Internet speeds.” as stated by Qureshi. Qureshi shows “how PDR can be used to reduce electric bills. He describe how to couple request routing policy to real-time price signals from wholesale electricity markets. In response to price-differentials, PDR skews client load across a system’s clusters and pushes server power-demand into the least expensive regions”. His conclusion is that existing systems can use PDR to cut their annual electric bills by millions of dollars.”

Setting a price for carbon emission will have a huge effect on (IT) economics. Its time to get the impact of a “carbon tax” on the radar screen of organizations as explained in the blog entry IT Carbon Emission is a million business. In his thesis, ‘Power-Demand Routing in Massive Geo-Distributed Systems’, Qureshi also show how PDR can be used to reduce carbon footprints. “Not all joules are created equal and in power pools like the grid the environmental impact per joule varies geographically and in time. We show how to construct carbon cost functions that can be used with PDR to dynamically push a system’s power-demand toward clean energy system”.

But beware we are not talking about sustainability we are talking about money: these ideas are not about energy or carbon savings but about energy and carbon costs savings.

Bookmark and Share

Data center doom scenario looming up because of carbon emission?

“There will be a price for carbon,” data center energy expert Jonathan Koomey said last Monday in his keynote at the Uptime Institute Symposium 2010. “We have to start thinking about how that price affects the economics of data centers. Carbon taxes will have an impact on where you locate your data centers.” He also said that this issue of the impact of a “carbon tax”is “not on the radar screen” of corporations yet. This issue was also echoed by other speakers at the Uptime event. “It’s not a question of if (carbon regulation) will happen. It will definitely happen” according to Mike Manos, the Vice President of Service Operations at Nokia.

One of the market-based approaches to control and reduce carbon emission is cap and trade, see Wikipedia. This mechanism is used in Europe. It has now the largest emission trading for greenhouse gases in the world, the European Union Emission Trading Scheme with The European Climate Exchange (ECX) in London as the leading marketplace for trading carbon dioxide (CO2) emissions in Europe and internationally.

If we can foresee that there will be a price for carbon produced by data centers and the mechanism are already in place, what will be  the impact on Data Center Power Usage and  Carbon Usage and Money spending? Is it worthwhile to monitor energy usage and the carbon footprint of a data center?

In my blog entry ‘Data Center Power Usage and Money savings‘ I made a rough estimate on money savings based on the precept that every kilowatt of computing related load saved (or wasted), equals 8760 kWH per year or about 35 Megawatt Hours over 4 years. Inspired by the Uptime Institute Symposium I now extended this rough estimate by including the cost of carbon and including a PUE conversion.

Carbon footprint

All power generation systems have a ‘carbon footprint’. There is some debate about how large these footprints are, but for an estimate I used some figures from the Parliamentary Office of Science and Technology of Great Britain:

  • Conventional coal combustion systems result in emissions of the order of > 1000gCO2eq/kWh.
  • The average carbon footprint of oil-fired electricity generation plants in the UK is ~650gCO2eq/kWh.
  • Current gas-powered electricity generation has a carbon footprint around half that of coal that is ~500gCO2eq/kWh.

Carbon price

For the price of carbon I used the EU carbon price from the European Union Emission Trading System (EU ETS). According to Reuters “Prices have averaged 13.66 euros in the first 5-1/2 months of 2010, meaning they will have to hold at current levels or rise further if they are expected to reach an average of 15.40 euros.”

Electricity price

To make a rough estimate of electricity costs I took a table of electricity prices from Wikipedia and used the exchange rate of 1.00 USD = 0.733783 EUR.

PUE factor

The PUE of data centers is very different, the EPA Estimated PUE Values for 2011 are between 1.2 and 1.9.

Table

A “worst case scenario” is used, where 1kW of dirty electricity (1000 gCO2eq/kWh) is producing 8760 kg CO2 per year or about 35 000 kg CO2 over 4 years. For the data center a PUE of 1.9 (to deliver 1 kW for the IT equipment you have to use 1.9 kW) is used.  The carbon price that is chosen is based on the 13.66 euro average price of the European cap and trade market. In this way we can get a first impression and a sort of upper limit. But be aware that the carbon price is dynamic and the EU carbon price is relatively low in comparison with 2009. Further in reality how green the supplied electricity is, is much more complicated and depends on the location of the data center and the power supplier. The PUE is a dynamic efficiency measurement because of the dependency on the workload that is being processed.

In the table below you can see the yearly and four yearly cost of 1 kW of electricity and the associated carbon costs, the total yearly and four yearly cost (electricity + carbon), the index on yearly electricity cost and total cost and the last column shows the rise of cost because of the adding of the carbon footprint. As you can see, you can save a buck with thinking 8760.

Electricity and CO2 costs

Electricity and CO2 costs

Problems are looming up

This table is all about kW and CO2, so to be more practical let us go one step further. We have servers in all kind of flavors that have a different power consumption. Making use of  the report ”Estimating total power consumption by servers in the U.S. and the world” of  Jonathan G. Koomey of Stanford University the power usage of low, mid and high range server are estimated on 180, 420, and  4800 Watt. So thinking 8760 this means:

  • low range server          =         180/1000*8760    =   1576.8  kWh/year
  • medium range server  =        420/1000*8760     =   3679.2  kWh/year
  • high range server         =     4800/1000*8760     = 42048.0 kWh/year

For a data center in Canada this give the following cost increase:

CO2 server costs

CO2 cost increase

Doom scenario

To get an impression what this means, let’s make a use case. If we take, for example, a data center of 5000 servers, with 50% low range, 25% medium range and 25% high range servers and we take the lowest electricity price (Canada), we can see the following cost increase:

  • low range server          =  €      102 310,-
  • medium range server  =  €      119 362,-
  • high range server        =  €   1 364 142,-

Total                                                 €   1 585 815,-  per year


Houston we got a problem!


Bookmark and Share

Iceland started green, emission-free data center, Vikings are back again

Yes the Vikings are back again, green IT Vikings. Yesterday, Nordic IT companies made several agreements to get Iceland first data center up and running.

Yesterday, 21 may at 12 noon Icelandic time, the President of Iceland and Iceland’s Minister of Industry has opened the first Icelandic data center, at Steinhella in Hafnarfjörður, the third largest city in Iceland, located on the south-west coast of Iceland, about 10 km south of Reykjavik. An Icelandic company, Thor DC, is behind the construction of the center and its operations.

Thor data center

Thor DC collaborates with a Spanish technology firm, AST,  that has become a leader in the production of so-called data containers. The data center has been built in 18 months and it is using electricity from geothermal and hydroelectric power plants. Thor data center therefore claims to use 100% green, emission-free energy. The data center has secured 3.2 megawatts of electricity to be provided by HS Energy on the Reykjanes peninsula, which is extendable up to 19.2 megawatts.

And they already have their first customer: Opera Software ASA . The Norwegian browser-creator becomes the first international company to move its data processing to an Icelandic data center. “The reason Opera Software came to sign an agreement with Thor Data Center is that the company can offer us energy from 100 percent renewable sources. In coming years Iceland will be able to provide a more stable supply of energy than most other areas in the world,” Stephenson von Tetzchner, founder of Opera Software, said. Opera Software currently has over 110 million users for its browsers across the world.

To move a significant part of its electronic data traffic to the Thor Data Center via sub-sea cables, Opera Software  also signed an agreement with E-Farice (an Icelandic / Faeroese company) for the provision of bandwidth between Iceland, London, and Copenhagen, using the FARICE and DANICE submarine cables. Opera’s Icelandic operations will generate a vast increase in data transfer through the FARICE and DANICE sub-sea cables. With the expected growth of Opera products, any increased traffic will in the future be transferred through E-Farice and Thor Data Center.  Farice offers service between Iceland, the Faeroe Islands and the UK via the new FARICE-1 submarine cable system. E-Farice offers also services between Iceland and Denmark via the DANICE submarine system. Points of Presence (POPs) are in the following cities:

  • Reykjavik, Iceland
  • Torshavn, The Faroe Islands
  • London, United Kingdom
  • Copenhagen, Denmark

Bandwidth is provided within the SDH technical frameworks with a minimum bit-rate of 45 Mb/s (DS-3). GigE interfaces are also available.

This hunt for green, emission-free energy for data centers can give a new boost to the future of the North Atlantic economies.  Some interesting background information can be found in the reportIceland The Ultimate Location for Data Centers” made by PricewaterhouseCoopers for The Invest in Iceland Agency, that is run by the Trade Council of Iceland and the Ministry of Industry.

Now its waiting how fast the Shetland data center initiative will create an up and running green data center (see this blog entry)

Bookmark and Share

Greetings from Lerwick, Shetlands: Colocation & Data Center Real Estate

Yes you read it right, Lerwick Shetland’s capital is set to be the location for the UK’s most northerly data center. Shetland Islands Council and developer Alchemy Plus has signed an exclusive Memorandum of Understanding and unveil plans to build a state-of-the-art 10,000 square foot data center powered entirely by renewable  energy and using a powerful high-speed fibre optic telecommunication link. Two Lerwick sites – at Black Hill Industrial Estate and Port Business Park – are currently being considered for the data center location.

This proposed  £12 million project must lead to a 1.1 PUE data center where 100% of its primary power comes from a renewable energy power source and will utilize the existing grid for a secondary back up supply. The 100% renewable energy power source will be based on wind and hydrogen technologies and access to a highly effective and well established district waste heating scheme.

According to Alchemy “Shetland is one of the most promising locations in the world for the development of renewable energy. Currently up to 20% of the islands’ electricity demand is generated by Burradale wind farm and other turbines. Permission is being sought by Viking Energy and a subsidiary of Scottish and Southern Energy for a new wind farm generating up to 540 Megawatts of power. A project on Unst has carried out research on the use of wind power to create hydrogen and wave and tidal energy developments may follow in the next few years.”

After Iceland yet another place with a cool climate and abundant supply of wind energy (instead of geothermal) that makes it a prime spot for IT data centers. This makes me thinking will the Faroes or Jan Mayen be next? I must have an atlas …

Lerwick

Lerwick

Of course the climate and alternative energy supply is very tempting but there is also Mother Nature that should be taken in to account when placing a data center. A lot of the requirements are focussing on the “internal” things of a DC. But what about the quality of the DC geographical surrounding? Yes environmental risk is everywhere but as stated in a recent blog entry the economic value that data centers are representing is huge. This makes it very interesting how these (remote) data center sites are selected. And here we are not only talking about the physical threat by Mother Nature of the data center itself, but also of the collateral damage. For example in the case of the Iceland volcano it was not only the physical aspects of the volcanic ash, “Can the roof collapse?” , “Has the fine dust penetrated the outside air filtration?” or “Are cooling towers clogged with ash sludge?”. There were also the logistic problems it created, “How to get people and equipment to and from the DC?”. This kind of issues is of course the same for hurricanes, earthquakes etc.

From my point of view an GIS (geographical information system) app could be very helpful in making geographical site analysis. Being curious I raised this question (“Is any one familiar with studies or the practical use of an GIS (geographical information system) for selecting data center locations?”) on several discussion groups but the response was low. Although there was a telecom operator using  ArcGIS Desktop to identify potential region to locate new data centers facilities. And someone else suggested  the use of Google Earth and stated that “It would be nice if there were FEMA overlays in Google Earth”.

A lot of times the statement is “You have to have a contingency plan for your information systems”. But what about a contingency plan for critical IT infrastructure (read data center)? Very nice that the data center is still standing after a growl of Mother Nature but if in the nearby surroundings the energy supply, roads or airports are destroyed or can’t be used what then? The chance that this happens can be small but the impact is huge. Still intrigued if and how this geographical analysis is be done.

Bookmark and Share

Icelandic data centers: boom, boom, out goes the light?

“Iceland says its cool climate and abundant supply of geothermal and hydroelectric energy makes it a prime spot for IT data centers. But in the wake of a major volcanic eruption earlier this week, companies may be thinking twice about hosting core IT assets in a place where Mother Nature is known for being a bit unstable.” according to ChannelWeb. They sure have a point that although the Icelandic climate and energy supply is very tempting there is also Mother Nature that should be taken in to account when placing a data center.

Maybe this event is also a wake up call. The trend of cloud computing with the development and usage of mega data centers makes us also vulnerable. The economic value that data centers are representing is huge. So how are data center sites selected?

There are some nice white papers on the topic of Data Center Site Selection. For example the paper of John Rate of Rate Consulting or
Geographic Factors for Data Center Site Selection” of Fortrust. A pity they only focussing on the USA. Similar studies for other parts of the world could be interesting also.

So whats new? You could say everything is already thought-out. But do we really have the insight about the risks of current and future data center locations? How are this decisions being made? For instance who came up with the idea of data centers in Iceland? Was it just cherry picking (“green is nice”) or was it an thoughtful proposition? Read the blog the Data Center Knowledge blog entry Iceland gets major data center project.

Iceland Data Center Location

Iceland Data Center Location ((c) Data Center Knowledge blog)

A lot of stuff you can find about data center site selection is about ‘creating a long list of criteria, create weightings, multiple the numbers, add the scores, then select the highest score as criteria site selection’. A flawed concept as stated in Green (low carbon) Data Center Blog. You also can question how these selections are being made with all this different kinds of geographical data.

This made me thinking about an GIS (geographical information system) app for data center locations. An GIS app where you use several map layers to filter out the bad and ugly areas to find the good areas for placing data centers.

The bad areas:
Areas where Mother natures has here surprises, volcanoes, earthquakes, tornados, hurricanes, floods etc.

The ugly areas:
Areas where we created are own risks: Proximity to hazardous materials manufacturing and storage, airports, Critical Areas for Transmission Congestion, black & brownout areas, social deprived areas, etc.

The good areas:
Areas with cool climate and abundant supply of geothermal and hydroelectric energy and abundant network capacity and human resources.

A first search for Data Center Site Selection and GIS didn’t give a positive result …
Is this really true? How are this complex selections else being made? Are Geographical Information Systems not used in site selection of a multi million business of data center building? Is it just old-fashioned handwork?

Bookmark and Share