Green Carrot energy usage app

Last August Tendril, a US-based energy platform company, announced that they were opening their API’s and launching an energy application developer program. The idea is to allow developers to build on Tendril’s cloud platform and to deploy the developed applications on Tendril’s Tendril Connect cloud platform.

For developers this is an opportunity to develop applications addressing the energy challenge and have them deployed in a ready-made marketplace of up-to 70 million addressable households. Similar to the Apple App store, Tendril offers co-marketing opportunities for developed applications.

Tendril provides developers with, not just the API’s, but also comprehensive documentation with a “Try it Now” capability as well as a discussion forum (so far lightly used) to have questions answered.

Tendril has also been promoting this initiative to developers by participating in Hackathons in San Francisco and more recently in New York. In conjunction with the New York Hackathon, Tendril ran a contest to see who could come up with the best apps using their API’s. The winner, eMotivator, won $3,000, while 2nd placed Green Carrot (screenshot above) won $2,000 from Tendril and another $1,000 from the Hackathon organisers for “best user experience”.

And I note that Tendril are listed as one of the Participating Organisations in the London Green Hackathon being organised by AMEE this coming weekend.

Of course, if Tendril really want to talk to developers, they should also be attending our RedMonk Monki Gras conference in London next week (Feb 1-2)! I’m not sure what the collective noun for developers is (I asked on Twitter and received the following suggestions – batch? class? scrum? repository?), but whatever it is, there’ll be a shedload of them there!

One of the interesting things about the Tendril open API initiative is that it should stimulate lots of creativity in the Smart Grid space. So far, as Tendril CTO Kent Dickson noted in a call with me the other day, no-one knows what the Smart Grid killer app will be, but crowdsourcing the ideas is far more likely to lead to compelling results.

I gave the keynote address at the Digital Trends 2011 event organised by HePIS and CEPIS in Athens recently. My talk was on Cloud Computing’s Green Potential and in my presentation, I claimed that Cloud Computing is NOT Green.

I started the talk by explaining what Cloud Computing is and the many advantages it can bring to companies. However, because none of the Cloud providers are publishing energy figures around Cloud computing, we can’t say whether or not Cloud computing is energy efficient.

I went on to point out that even if Cloud is energy efficient (and we have no proof that it is), that is not the same thing as being Green.

My slides are available on my SlideShare account and a transcript of my talk is here:

Okay, so my talk this morning is on Cloud Computing and its Green Potential. So a quick couple of words about myself.

So my name is Tom Raftery, I work for an industry analyst firm called RedMonk. My area of interest within RedMonk or the area I specialize in is energy and sustainability. We have termed the practice within RedMonk that concentrates on energy and sustainability GreenMonk. So the place that I blog at is at GreenMonk.net.

And a little bit about my past. I worked in an organization called Zenith Solutions back in the 90s and early 2000s, and Zenith Solutions was a software company creating what has now become termed cloud applications. At that time we called them web applications, they were web based software with the database backend online.

Then I worked for a company called Chip Electronics in the early 2000s and Chip Electronics was again a company which created Enterprise Resource Planning, ERP applications which were cloud delivered, at the time we called it Software as a Service. No at the time we called it active service provisions, since become Software as a Service. And I am also a co-founder and Director of CIX, which is a hyper energy efficient data center based in Cork in Ireland. So I know both from the hardware side and the software side.

I mentioned my blog on GreenMonk.net, I am on Twitter and twitter.com/tomraftery. My email address is there, my mobile phone number is there, please don’t ring it now. And this site here, slideshare.net the last line there and I am sorry for the bullet points, I don’t normally use them, but I did just here and in one other slide. slideshare.net is a site where you can upload a presentations.

So, this presentation I am giving this morning, I uploaded it to SlideShare earlier this morning, so it’s already online there at that site and if you go there now you’ll see it has already been viewed over 277 times so far. So, it’s a great site for getting your talks out all available, it’s also downloadable there.

One thing you’ll notice as well about the structure of my talks is a lot of them have images like this, but they also have this bit of text at the bottom which you can’t read, don’t try right now, but what they are is those are links to the source material. So, if at any point you do download the presentation you can go and click on the links, they are clickable links you can click on them and see where I’ve got the information from.

So that’s me, who are you guys?

A couple of questions, so how many people here have deployed applications to the cloud? Not very many. How many plan to? A few more, okay. How many people here think that cloud computing is green? Okay, good few people. Right. I hope to burst that bubble, unlike Nancy who spoke just a minute ago, I am not a, I am not a believer that cloud computing is green and I hope to explain why. I am a huge fan of cloud computing, I have to say, I use it extensively, going back to the slide for a second.

The Chip application, the Zenith stuff, the GreenMonk, Twitter, SlideShare even my email are all cloud delivered. Our organization RedMonk we use Google applications for domains for our email, so my email is cloud delivered as well. So I am a big user of and believer in cloud for lots of things. But I just don’t happen to believe it’s green.

So what is cloud computing? Well at kind of first blush it’s software that’s delivered in a browser, so that’s an very easy definition of it, something we can all kind of sign up to it. It’s a lot more complex than that at various other levels and I’ll go through a couple of those other levels as well, just very briefly to kind of give you that the kind of complexity that’s involved in it, but I am not going to go into any great depth. So it’s also nothing that’s very new, this is the original sign up screen for Hotmail.

Hotmail was an email application developed and sold to Microsoft back in ’97 for $450 million if memory serves. But this was before it was sold to Microsoft, this was the original sign up screen when they launched in July ’96 and it was one of the first widely used Software as a Service or cloud application.

So cloud is nothing new, it keeps getting rebranded, so the cloud name is newish alright, but the delivery mechanism is not that new. It actually harps back to mainframe computing back in the 60s.

So there are several types of cloud computing and the first type, the first level of cloud computing is kind of Software as a Service. That’s where you kind of take your packaged software and convert it into something as I mentioned already delivered in a browser. And I mean you probably are aware of these I mentioned Hotmail and its analogs the Google applications, there is also Zoho, there is social networking the Twitter that I mentioned, SlideShare all these kind of things, they are all Software as a Service.

So they are just basic applications that you access through a browser. But you can go back one level of abstraction from that to where you get to what’s called platform as a service. And don’t worry about these acronyms basically a lot of the times you don’t need to know this stuff, the platform as a service stuff is where you, as I could say, you go back one level of abstraction and you give people a platform on which to deploy cloud applications.

And the kind of platforms that you can get are ones like the Google app engine and Amazon and Microsoft’s Azure, these are the kind of platforms that are available if you want go down that route. Most people don’t need to go there, but if you do that kind of stuff is available as well. And then you can go back one further level of abstraction where you are actually delivering Hardware as a Service and this is called Hardware as a Service or Infrastructure as a Service and both names are valid, HaaS for hardware or IaaS for infrastructure as a service and that’s where you’re delivering stuff like networking, storage, compute, CPU cycles that kind of thing as a service.

And VMware, Rackspace, OpenStack again Amazon with their EC2 and their S3 services are those kinds of types of cloud computing. If that’s a little confusing and I know it can be, this is a slide which is also confusing, but if you actually stop and study it in your own time, you could download this application and if you are interested about it, this is a good way of seeing how the different types of cloud computing stack up as it were.

So over here on the very left, you have your traditional packaged software with the entire stack from networking up through applications where you manage the entire stack on your machine. So that’s the traditional Microsoft Office whatever applications, you do the whole thing.

Over on the other side you got your Software as a Service, something like Google apps or domains or one of these things where the provider the Google or whoever are responsible for the entire stack, all you have is a browser. And then in the middle you have the two other ones, the platform as a service, where the vendor managers up to here and you manage the applications and data or infrastructure where the vendor managers is just this part and you manage the rest.

So that’s the kind of way it stacks up. As I say on the deck itself there is a link down there to where you can find that image if you are interested in checking into it. It’s quite a nice way of seeing the differences between the different types of cloud computing.

And then just to complicate things a little further, there are different deployment mechanisms. You can have private cloud, private cloud is hosted by yourself on your own infrastructure behind your own firewall. You can have public cloud which is what most people are familiar with or you can have a hybrid where you have some stuff private, some stuff public and that’s one that a lot of people are looking at, because it means you can have your data behind your firewall, but the functionality you are accessing it from public. So your stuff remains on premise.

And that’s quite important, because as Nancy alluded to, there can be a lot of issues with the data in cloud computing, because for example if you are a European company do you really want your data hosted on servers in US territories where for example the data privacy laws are a lot more lax. So I have spoken to several European companies who have said categorically they will not use cloud computing if their data is going to be hosted in US territories. It’s only if it’s in the EU and only if they know where in the EU. So you are noticing cloud providers taking that on board and starting to become aware of those issues and while they can’t change US law, they can start providing storage mechanisms that they are guaranteed to be in region.

So that’s cloud computing and the next question we get to is, is this really energy efficient because lots of people say it is and even Nancy alluded to that report from the Carbon Disclosure Project which I’ll blow apart in a minute. They aren’t the only ones Microsoft, Accenture and WSP environment brought out this story in November of last year. And this is the actual title of the story, where they say it shows significant energy in carbon emissions reduction potential from cloud computing and again the link to the report is down there at the bottom.

The difficulties I have with that are several, first is Microsoft are a cloud computing provider so they kind of skin in the game. The second is that, they don’t actually use any hard data, it’s all imputed. And the third is that after months and months of work from all these people the best they could come up with is they could say it has potential. Yeah it has potential to end world hunger and bring on world peace and fix the euro, anything kind of potential. So that’s a non-report.

Cloud computing has phenomenal advantages, don’t get me wrong, I am a big fan. So if you are into traditional IT, you know well that if you are deploying a new application or a new server it’s pain staking, you have to go through an RFP process, a tender process, PO process. You have to put, you have to go to tender and you have to get that — when you have to place the order, the order then can take several weeks from the supplier. When it comes in, it goes into the logistics area, if you got to get the guys in warehousing to tell you where the server is, you have to get the server, you have got to put the company image on the server, you got to install the applications, you got to do testing, you got to patch the server, the list goes on and on. Basically you want to deploy a new server, it’s a process that can take weeks to months.

You deploy a cloud application, there is usually no RFP and no PO process because there is — the capital cost is minimal. So typically the time to deploy for a cloud application shrinks from weeks to months to hours to minutes depending on what you are deploying, so phenomenal, cloud is fantastic for streamlining that kind of stuff.

It’s also great for what’s called dynamic provisioning. So this is the Alexa graph, the website traffic of a website for the Australian Open. The Australian Open is a big tennis competition happens in Australia every January. So you’ll notice 11 months of the year no traffic to the site, come December, January vroom, spike, that’s 2006, 2007, larger spike 2008, larger spike and the spikes keep getting bigger as you go in that direction.

So if you were the website owner for the AustralianOpen.com website you would need to have — if there were no cloud computing options you would need to have servers that could hit and deal with the traffic at this growing spike for 12 months of the year when the traffic is only there one month of the year. But with dynamic provisioning and cloud computing you can use the elasticity of the cloud to turn up the resources assigned to that site as the traffic starts to build up in December and January and then as the traffic falls off, you turn it back down again.

So in that respect cloud computing is fantastic as well, you are not using resources needlessly. You’ve also got the idea of multi-tenancy and if you can’t see what’s in this picture it’s actually a Mini Cooper with 26 people inside in her, EMC sponsored it as the world record attempt to fit people into a Mini Cooper and they fit 26 people into it. So they stuff people into it with multi-tenancy in cloud computing it means you are sharing applications across companies, lots of different companies often competitors are using the same single version of the application.

And that’s fantastic, that adds greater value. You know, you have only one instance of the application which is great as well for updates, updates of the application are instantly deployed. You know, you don’t have to download the latest update and apply it to the test server and make sure it works in the environment, the whole thing, you know, it’s just instantly on.

This is the issue of server utilization which again Nancy referred to, Nancy you stole my talk, come on. So this is a typical graph of server utilization and you can see this the memory part, but this is the server utilization and it’s at zero percent here. And well that’s a bit of a outlier, you’ll often and get in normal server, you’ll often get utilizations in single digits 7, 8% server utilization for traditional servers in data center environments. But with the advent of virtualization and cloud computing you can ramp that up significantly. So that should be quite energy efficient.

Then you have got this kind of outlier thing called chasing the moon, which you may or may not have heard off. It’s one I am kind of found of as an idea, but not many people have deployed it yet. People are kind of talking about it as out there, and what it is, is with cloud computing if you’ve got data centers in say, US, West Coast, another in Northern Europe or Southern Europe, Northern European typically because it’s cooler there and cooler I mean colder not more ‘hip’. And you’ve got another data center say somewhere in Asia or Eastern Russia. Then you’ve got the time zones covered about eight hours apart. So if you have an application in those three centers, you can move the compute to where energy is cheapest at any particular point in time. So if you are doing that typically energy is cheapest when it’s in highest supply, when it’s in highest supply and it’s cheapest, its actually, this is on the wholesale markets, it’s actually greenest as well.

So when electricity is at its cheapest, it’s actually also at its greenest that’s – it’s kind of counter intuitive but I can explain that if any one who is interested later.

So if you move your compute to where the energy and the compute is cheapest at any point in time, it’s typically night time when wind is blowing and at that time you are chasing the moon, you are putting your applications wherever the moon is out, it’s called chasing the moon.

And so it’s something you could only do — something that’s only made possible by the likes of cloud computing. Your information is ubiquitous, it’s wherever you have an internet connection, so your road warriors, your sales people on the road, can access the application while sitting up in the beach.

It also enables a lot more home working, homeshoring, teleworking whatever you want to call it. And people like ATT, IBM, lots of big companies are huge fans of this. IBM reported a couple of years back that 25% of their employees did teleworking and those 25% were saving IBM $700 million a year. That’s significant savings and a lot of that savings comes from a lower real estate footprint and a lower energy footprint because of the lower real estate footprints.

So is it energy efficient or lot of those savings coming from less commuting or from less building stock or are they from offsetting your energy? So if you are working from home you are still burning energy, it’s just not in your company’s building, your company isn’t accounting for it anymore. These are kind of questions we are not sure of, there hasn’t been any definitive studies either way, and it’s difficult anyway because it differs in every company and every geography.

One huge problem I have with cloud computing and people saying that cloud computing is energy efficient is that none of the cloud providers are publishing data around their energy utilization, not one of them. So I often do a kind of a hands up exercise at this point and I don’t know if it’ll work here, because very few people admitted that they were going to be putting stuff in the cloud, but let’s raise hands again. Hands up everyone who has or plan to deploy applications to the cloud? Okay, so keep your hands up, keep your hands up. Now keep your hands up if you know the current energy utilization of the applications you are going to deploy to the cloud or the energy applications you have already deployed to the cloud, if you know how much energy your applications burn, keep your hands up. Okay we got one, anyone else just the one? Good. Okay, keep your hand up, we are not finished. Okay keep your hand up if you know the energy utilization of that application in the cloud. You do, is it a private cloud?

And they are giving you the energy utilization of that?

Okay, I am interested in that because I do a lot of work for SAP and they can never tell me the energy utilization of any of their cloud infrastructure.

I think, okay I must get back to you on that because I — if that is the case it must be very new, because it’s not something they’ve ever shown me before and if they do it it’s often kind of what we call humbligated it’s often — they will give you an average but they won’t tell you exactly what your application and what your users are utilizing which is what you need to know.

So as I say with the possible exception of SAP, most cloud providers do not provide the data. So without that data, we have no way of knowing if our applications are in fact energy efficient in the cloud. Even if they are, energy efficiency is not the same as being green. Just because something is energy efficient does not mean it’s green, and this is a very common mistake that people make.

So what is green? Well I’m also going to quote from the CDC report that Nancy mentioned. And this CDC report, it’s called Cloud Computing – The IT Solution for the 21st Century and again there is a link to it there.

One of the quotes that the CDC put into that was that that a typical food and beverage firm transitioning it’s HR application from dedicated IT to public cloud can reduce CO2 emissions by 30,000 metric tons over five years. That sounds good, I buy that, that sounds quite green actually. They also went on though in their executive summary to say allowing companies to maximize performance, drive down costs, reduce inefficiency, and minimize energy use and therefore carbon emissions. So they have made the classic fundamental mistake of thinking that reducing or making something energy efficient makes it green or reduces emissions. There is nothing that says that is the case. I will tell you why?

And by the way Nancy and this is the blog post I wrote on GreenMonk where I explain in significant detail why the CDC report is completely flawed. So there is a link there you can go and have a look.

I will give it to you later, I will give you the link later. So two reasons why that report is deeply flawed. One of them is it’s based on assumptions and they say so in their model, they say their model is based on assumptions and the second is based on metric called PUE. So I’ll get the PUE in a second. PUE is Power Usage Efficiency. Anyone here familiar with the term PUE? Couple are, okay. So I’ll get to in a second. They based their model on this PUE metric, which is a widely used metric. But they have based it on a average PUE across the entire United States which is put out by the EPA, so that’s not really very indicative firstly.

But secondly, PUE the measurement itself, it’s a ratio of the total amount of power used by data center compared to the power deliver to the computer equipment. So the total power delivered to a data center is the power delivered literally to the door of the data center, which goes to power of the lighting, it goes to power the cooling, it goes to power of the UPSs, the whole thing, that’s the total power. And it’s a ratio of that to the amount of power which actually makes to the IT equipment. So the closer your PUE figure is to 1.0, the better.

Now couple of problems with PUE as a metric. The first is there hasn’t traditionally been any standard about where you measure the power. It’s measured at the meter, the power meter, the electricity company power meter but in data centers that is often been measured on the high voltage site, the medium voltage site, or the low voltage site depending on the data center, depending on where they put the meter. And of course if you are measuring out at these different place one data center measures at one place, another at another place you can’t cross compare because you are leaving out the loses that occur in the conversion from high to medium to low voltage.

So right there the lack of standardization that’s been worked on at the moment, and it is getting better, they are start to standardize but traditionally it hasn’t been standardized. So that’s one issue straightaway.

But even more important is, quick look at this graph here, if I have a data center which takes in two megawatts of power and of those two megawatts of power, one megawatt goes to powering the servers, then I have a PUE of 2.0, very simple. However if I realize that my one megawatt isn’t being used very efficiently, and I realize that some of my servers are being under utilized, may be I’ll virtualize some of them and I shut some of them down. Then my power draw from my IT equipment drops to 0.75 a megawatt. My total draw drops to 1.74 and my PUE goes up to 2.33. Remember I said PUE closer to 1.0 means you are more efficient. I’ve actually made it more efficient and my PUE has gone from 2 to 2.3, so it’s a huge problem right there with the PUE metric.

Another issue and I am sorry this is a little bit complex, but I’ll walk it though it again another issue with PUE is it takes no account of carbon. So top line is a typical data center, typical data center in a European country, a European country where the supply of power causes CO2 per kWh that’s not unusual that’s a pretty average figure. So if that typical data center has a PUE of 1.5, just quiet good, then we get 1.5 by 0.5 we get 0.75 kg CO2 per kWh is the IT carbon intensity, is how much the IT equipment is producing in terms of CO2.

So if we have a data center with the a good PUE 1.2 and that good data center is drawing mostly from coal-fired power, so it’s got a supply carbon intensity of 0.8 kilos, the you are — 0.8 by 1.2 means you are producing — even though you have got a good PUE in your data center, you are producing 0.96 kilo CO2 per kWh.

On the other hand if you have got a really bad data center with a PUE of 3, very inefficient data center, PUE of 3 but is fired mostly by renewables. So your carbon intensity on your supply is 0.2, but still it’s not 0 it’s 0.2, it’s significant, there is still carbon being produced per kWh. 0.2 by 3 gives you 0.6, which is significantly less. It’s, you know, 60% of the carbon intensity of the data center with a PUE of 1.2.

So PUE is no indicator whatsoever of how green some thing is, it’s a very bad metric. There is other metric called a CUE the Carbon Usage Efficiency but it is not widely used by the industry unfortunately.

So given that, the report that the CDC produced could just as easily have said that your typical food and beverage firm transitioning HR to a dedicated public cloud could increase CO2 emissions by 30,000, just as valid. They pick the number out of the air and they base it on a flawed metric, so it’s just as valid to say it could increase CO2 as decreased CO2.

And a good example of this in fact is Facebook, Facebook build this lovely new data center in Prineville, Oregon, they opened it early this year. Had an unbelievable PUE coming in around 1.08, highly efficient, they have open sourced it. If you go to opencompute.org you can get the blueprints for building the data center and the list of suppliers, so you can go on build that same data center yourself. They reduced their energy consumption per unit of computing by 38%, that’s fantastic, that’s really good.

Unfortunately, their energy supplier is a company called PacifiCorp. PacifiCorp produces 58% of its power from coal, this is their website, they say it themselves right there. They produce 9.6 million tons of coal from their own mines every year. They also produce another 12% of their electricity from natural gas, so that’s 70% they produce from fossil fuels, plus they also buy 22.5% of their power from other suppliers. Now they don’t say how those other suppliers generate theirs. So they get at least 70 and possibly significantly more of their energy from fossil fuels. So right there Facebook’s really impressive 1.02, sorry 1.08 PUE is blown out of the water by the fact that it’s all produced using carbon or almost all.

To contrast that this is the graph of power production on the Spanish grid. I picked Spain because I happen to live there and you can see that this was taken last Friday, the snap shot and again you can get information from the Red Eléctrica de España. The red bar here, or bar whatever it is area is coal, comes in around 20%, green one here is wind, the other is, the rest of the renewables put together except water, which is down here.

So on the Spanish grid, carbon intensity comes in around 20% at the moment based on that plus the gases and other 15% sorry. And that’s actually bad compared to the same grid two years ago. The same grid two years ago, coal was coming in single digits here on 9% and the reasons gone from the 9% two years ago to 20% today is because Zapatero, the now former President of the parliament comes from Castile and León which is a coal producing part of Spain and he went to the EU and he petitioned for years to be allowed to give subsidies to the coal miners and the coal production in that part of Spain. And last February, sorry February 2010 he succeeded and they got 4.5 billion Euro in subsidies for producing coal in Spain and that’s a direct result, it’s scandalous.

Anyway that’s either here or there let’s move on. Dublin, I am originally from Cork, so I’ll talk of Ireland for little bit. Dublin has become a European, key European datacenter hub. Most of the big data center providers have significant operations in Dublin. Microsoft have their largest datacenter outside of US in Dublin. It provides the Microsoft Live and Azure services for EMEA. Similarly Amazon are there, Google are there, they all have significant centers there and they are all expanding. Google have announced new expansion as of Microsoft representatives.

Unfortunately, Ireland produces 84% of its electricity from Fossil fuels, that number is falling as it rolls out wind, it was a 87% a couple of years ago, but now its 84%. So that’s not really very green. On the other hand Apple have announced their iCloud service and their iCloud service is currently housed in this datacenter in North Carolina, it’s a photograph of it under construction. Unfortunately this datacenter gets 78% of its energy from Duke Energy who get it from, no they get all of their energy from Duke Energy, Duke Energy get their energy from 78% coal and nuclear.

Coal is obviously really bad, nuclear is also really bad, not just because of the Fukushima reasons, but you got to think as well, nuclear power plants have an enormous water footprint. And that is, it can often be devastating to the environment around them. But Apple have said in fairness to them that they have got a 121 acres site behind them that they are now clearing for scrub to install a solar facility, it will mitigate a little their emissions, a 121 acre solar facility is not going to complete power the plant and it certainly won’t power at night time but it will have at least.

Google are the really interesting player in this field, because Google have gone to extreme lengths to get their carbon footprint down to zero. They have signed power purchase agreements with Wind farms. A power purchase agreement is where you sign on the dotted line with a wind farm or any one in this case it’s wind farms in Google’s case, you sign on the dotted line and you say, for the next X years and in Google’s case its next 20 Years I will take all of your power, all of it.

For wind farms that’s a huge win because that means they can go to investors and say, we’ve got checks rolling in from Google for the next 20 years guaranteed and here is the contract, do you want to invest in us now. So obviously it will help them get investment inward investment. It’s a great win for Google on two fronts, one is they’ve got guaranteed renewable energy and the second is they’ve got guaranteed pricing for their energy for the next 20 years. How many people can say they know what the price of their electricity is going to be in 20 years?

So they have done it in a number of wind farms, this one is in Iowa, 114 megawatts, this one is in Oklahoma, another 100 megawatts. And they’ve gone to incredible lengths and this is the slide I mentioned earlier with all the bullet points and I’m really sorry this slide is more bullet points than I normally have in entire talks. So there is nine bullet points in this but it gives you an idea of the investments that Google have made in renewable energy.

The Potter Drilling one is for geothermal, Makani Power is for high altitude wind power. Solar City, the Atlantic Wind connection, that’s a — I can’t remember how long it is, it’s a massive offshore wind farm off the east coast of the US and they’ve invested in the infrastructure of that, they own a percentage of the infrastructure in that, in total Google have invested $850 million on renewables. So if anyone could be slightly green I guess in cloud it would be Google.

Having said that though, there is lot of having said that in this presentation I am sorry about that, it’s a complex area. Jevons paradox, William Stanley Jevons is an economist in the 19th century in England. And he realized that as steam engines were being made more efficient in burning coal, as they were getting more efficient at burning coal the amount of coal being burnt was actually going up not down. It was kind of counter intuitive but how it worked was as stream engines become more efficient the price dropped, more people bought them, more coal was burnt. And it’s similar with cloud computing. As cloud computing goes and starts taking off in adoption the resources that are used actually go up not down.

So cloud computing leads to an increase in consumption and this also then Parkinson’s Law and a curare of Parkinson’s Law which says data expands to fill the space available for storage, the more storage you have the quicker it fills up, you all know that. And a good example of that is when Gmail started in 2004 Hotmail had a 2 meg limit on the size of your inbox. Yahoo at 4 meg and gmail said, I am offering one gigabyte, blew them out of the water and suddenly they had catch up and go to one gigabyte, Google is now at 7.5 giga byte and they allow you to send emails of 25 megabyte, single email 25 megabytes.

So, also a nice quote I got from Infochimps’ Flip Kromer and this really characterizes how cloud computing can promote consumption. And Flip said it very well when he said, EC2, which is one of Amazon’s offerings in cloud means anyone with a $10 bill can rent a 10 machine cluster with a terabyte of distributed storage for eight hours and because it costs virtually nothing to do it and because anyone can do it — it happens all the time. So that’s not very green.

You’re confused yet? So what if ultimate irony you had cloud delivered green software, Hara, Nootrol, SAP’s Carbon Impact OnDemand, these are all carbon management applications which are cloud delivered. Now, I am really confused I don’t know if they are green or not. Microsoft Iron Earth, cloud delivered using the Azure platform for managing air quality, water quality, noise pollution, that kind of thing, I think my head just exploded.

So my conclusion from all this is that cloud computing has a significant number of advantages, but being green isn’t one of them. One last thing if you are deploying stuff to the cloud, this is not really how you want to do it. Thanks very much.

Follow @TomRaftery

Back in July of this year (2011), the Carbon Disclosure Project (CDP), in conjunction with Verdantix, released a report titled Cloud Computing – The IT Solution for the 21st Century [PDF warning] which erroneously claims Cloud Computing is Green. Shortly after it was released, I wrote a long post outlining exactly where the report was flawed. I also contacted the CDP directly outlining my concerns to them and pointing them to the blog post.

Then, a couple of weeks back, when preparing my slides for my Cloud Computing’s Green Potential talk for the Cepis and Hepis Green IT conference in Athens, I discovered that Verdantix and the CDP had published

a new report [PDF] on the business and environmental benefits of cloud computing in France and the UK

Unfortunately, not only does the new report make the same mistakes as the original one, but it further compounds those errors with an even more fundamental one.

Let me explain.

In the key assumptions section of the report it talks about the metric tons of CO2/kWh in both the UK and French electricity grids (0.000521 tonnes and 0.000088 tonnes respectively). It uses these figures to extrapolate the savings in both France and the UK for companies migrating their applications to cloud computing.

So? You say. Sounds reasonable to me.

Well, the issue is that they didn’t do any work to identify where applications migrated to the cloud would be hosted. The implication being that UK applications migrated to the cloud, will be hosted on UK cloud infrastructure and French IT applications will be migrated to French hosted cloud infrastructure. In fact this would be a highly unusual scenario.

A quick look at where most cloud hosting takes place shows that the vast majority of it is occurring in the US, with quite a lot happening in Singapore with a lesser amount in Europe (and that split between Ireland, Germany, UK, etc. but almost none in France – Ireland is underestimated in the list as it doesn’t include Microsoft which has a significant Cloud hosting facility in Dublin which it is now expanding or Google’s Dublin facility).

Ok, and what about the carbon intensity of electricity generation in these countries? If a cloud application is moved to somewhere with a lower carbon intensity for electricity generation, then there is a possibility of a carbon saving. However, with the vast majority of cloud hosting still being done in the US, that isn’t a likely scenario.

This table of CO2 emissions from electricity generation, by country shows that the US has one of the most carbon intensive electrical grids in the world. France, on the other hand, with its high concentration of nuclear power (78%) has one of the least carbon intensive electricity grids in the world. While the UK grid’s carbon intensiveness at 557kg CO2/mWeh sits just above the world average of 548kg CO2/mWeh.

While it is possible (though not probable) that UK IT applications outsourced to the cloud would be hosted in a country with a lower carbon intensity than the UK, the chances of a French IT application being hosted in a country with a lower carbon intensity than France are virtually nil.

Given this, the assertion by the CDP report that

large businesses in France and the UK can reduce CO2 emissions from their IT estate by 50% compared to a scenario where there was no cloud computing.

seems, at best, extremely improbable.

One problem with coming up with reports like this is the lack of transparency from cloud providers on their locations, their energy and carbon footprints. If all cloud providers reported these metrics, it would be a far simpler matter to decide whether cloud computing is green, or not. Without these data, there is absolutely no way to say whether moving to the cloud increases or decreases CO2 emissions.

If you are wondering why the Carbon Disclosure Project and Verdantix are so bullish in their assertions that Cloud Computing is Green – if you scroll to the bottom of the report, you’ll see this:

This study was supported by AT&T
For more information on AT&T Cloud Solutions go to …

The report was paid for by the Cloud Solutions division of AT&T. Enough said.

Photo credit fotdmike

Follow @TomRaftery

Google has had an impressive record in renewable energy. They invested over $850m dollars in renewable energy projects to do with geothermal, solar and wind energy. They entered into 20 year power purchase agreements with wind farm producers guaranteeing to buy their energy at an agreed price for twenty years giving the wind farms an income stream with which to approach investors about further investment and giving Google certainty about the price of their energy for the next twenty years – a definite win-win.

Google also set up RE < C – an ambitious research project looking at ways to make renewable energy cheaper than coal (unfortunately this project was shelved recently).

And Google set up a company called Google Energy to trade energy on the wholesale market. Google Energy buys renewable energy from renewable producers and when it has an excess over Google’s requirements, it sells this energy and gets Renewable Energy Certificates for it.

All hugely innovative stuff and all instituted under the stewardship of Google’s Green Energy Czar, Bill Weihl (on the right in the photo above).

However Bill, who left Google in November, is now set to start working for Facebook this coming January.

Facebook’s commitment to renewable energy has not been particularly inspiring to-date. They drew criticism for the placement of their Prineville data center because, although it is highly energy efficient, it sources its electricity from PacificCorp, a utility which mines 9.6 million tons of coal every year! Greenpeace mounted a highly visible campaign calling on Facebook to unfriend coal using Facebook’s own platform.

The campaign appears to have been quite successful – Facebook’s latest data center announcement has been about the opening of their latest facility in Lulea, Sweden. The data center, when it opens in 2012, will source most of its energy from renewable sources and the northerly latitudes in Lulea means it will have significant free cooling at its disposal.

Then in December of this year (2011) Facebook and Greenpeace issued a joint statement [PDF] where they say:

Facebook is committed to supporting the development of clean and renewable sources of energy, and our goal is to power all of our operations with clean and renewable energy.

In the statement Facebook commits to adopting a data center siting policy which states a preference for clean and renewable energy and crucially, they also commit to

Engaging in a dialogue with our utility providers about increasing the supply of clean energy that power Facebook data centers

So, not alone will Facebook decide where their future data centers will be located, based on the availability of renewable energy, but Facebook will encourage its existing utility providers to increase the amount of renewables in their mix. This is a seriously big deal as it increases the demand for renewable energy from utilities. As more and more people and companies demand renewable energy, utilities will need to source more renewable generation to meet this demand.

And all of this is before Google’s former Green Energy Czar officially joins Facebook this coming January.

If Bill Weihl can bring the amount of innovation and enthusiasm to Facebook that he engendered in Google, we could see some fascinating energy announcements coming from Facebook in the coming year.

Photo credit Jaymi Heimbuch

Follow @TomRaftery

I attended the 2011 DataCenterDynamics Converged conference in London recently and at it I chatted to a number of people in the data center industry about where the industry is going.

One of these was Active Power‘s Graham Evans. Active Power make flywheel UPS’s so we talked about the technology behind these and how they are now becoming a more mainstream option for data centers.

Tom Raftery: Hi everyone, welcome to GreenMonk TV, we are at the DCD Converge Conference in London and with me I have Graham Evans from Active Power. Graham you guys, you make the spinning UPSs.

Graham Evans: That’s right yes the flywheel UPSs, kinetic energy. So behind us here we have our powerhouse. So what we found with the flywheel UPS is because of its high density environment, the fact it doesn’t need cooling the fact that it is really suited to a containerized environment we’ve put it in a powerhouse to show the guys in DCD to show the benefits that we can provide from a systems perspective.

Tom Raftery: So what the flywheel UPS does is it takes in electricity, while the electricity is running, spins wheels really, really fast and then if there is a break it uses that kinetic energy to keep the system up.

Graham Evans: Not quite, so the flywheel itself is spinning all the time as an energy storage device. The UPS system is primarily conditioning the power. So as the power comes through it’s a parallel online system, all of the kilowatts flow through to the load and our converters regulate that power to make sure you get UPS grade output through to your critical load. At the same time the flywheel is spinning it’s sat there as a kinetic energy store ready to bridge the gap when its required to do so.

An Active Power flywheel UPS

So voltage or mains fails on the input, the flywheel itself changes state instantaneously from a motor to a generator and we extrapolate that kinetic energy through some converters to support the load, start our diesel engine, and that then becomes the primary power source through to the system.

Tom Raftery: And you got the diesel engine in fact built into the system here, it’s on the right hand side as we are looking at here. So there is a reason that you have your own diesel engine kind of fitted into in there.

Graham Evans: Yes, so we are not holding into one particular diesel engine manufacturer so what we do as a complete system is designed as a critical power solution. So the thought really is from a client point of view we can be flexible in terms of their requirements. We can size the engine to support the UPS load only or maybe we can pick up some mechanical loads as well. We make some enhancements to the diesel so we have our own diesel controller to start the diesel quickly. We have our own product we call GenSTART,which allows us to have a UPS backed starter mechanism to the system so we can use that UPS power to start it.

Tom Raftery: And that’s important because the flywheel don’t stay up as long as say a battery bank.

Graham Evans: Its important because this type of loads that we are supporting need that quick power restorations, so from a UPS point of view we need to restore or keep power instantaneously that’s the job of a UPS, no break power supply, but we also find with mechanical loads certainly in high density datacenter environments we need to restore the short break mechanical loads very quickly. So the system you see here is able to do that. We continuously support the UPS load and we can bring on the cooling load ten seconds afterwards. So very fast starting, very robust system.

Tom Raftery: And the whole flywheel UPS idea is relatively new comer to the datacenter environment?

Graham Evans: Not especially I think it feels like that sometimes but we have been around for 15 years as a business, we have 3000 plus installations worldwide, but certainly we are not as common place as some other technologies but we are probably one of the fastest growing companies globally. So, yeah not brand new 15 years in business, but yeah the concept’s really taken off and it’s been really successful for us.

Tom Raftery: Cool. Graham that’s been fantastic, thanks for coming to the show.

Graham Evans: No problem, thank you, cheers.

Intelligent water is the latest addition to the IBM Smarter Cities portfolio.

in the world’s cities are growing at an astounding rate. For the 1st time in history, over 50% of the world’s population now lives in urban areas. Over 1 million people are moving into cities every week, and it is estimated that by 2050 70% of people on the planet will live in cities.

This unprecedented growth in urban populations makes the provision of basic services like transport, security, water etc. increasingly complex. This is irrespective of whether the city is a mature city or a developing one. It was against this backdrop that IBM launched its Smarter Cities product.

At the core of this offering is its Intelligent Operation Center (IOC) which takes inputs from systems throughout the city and depending on the input, raises alerts, kicks off workflows, or displays the information on any of a number of dashboards which can be configured to display differing information based on a user’s login.

The newly announced Intelligent Water offering is yet another module capable of working with the IOC. In their briefing call, IBM were at pains to point out that their experience working with custom projects like the ones in Galway Bay, the Washington DC, and the Dubuque, Iowa all helped shape this new product.

Water issues are global in their reach, even if the difficulties differ from region to region (i.e. drought in Texas, flooding in Thailand, water quality in India).

IBM’s Intelligent Water helps organise water-related work. It drives proactive maintenance and schedules the maintenance so that the majority of the time is spent actually doing the maintenance, as opposed to driving between destinations. It allows water managers to see where the water is going – is it being delivered to customers, or disappearing through leaks in the pipework? Are customers using it effectively or not?

It integrates with geospatial packages and has map-based views, there are analytics for optimized scheduling, work order reporting, water usage reporting and display dashboards with roles-based information display. It is possible therefore to create views for the public (to display on the municipality’s website, for example), views for the Mayor’s office and others screens for the water planners and water operators.

According to IBM, Intelligent Water is available today. It comes with IBM’s business intelligence reporting tools as part of the solution and is available in standalone, cloud or hybrid versions.

I’m open to correction, but I’m not aware of any other company offering a comprehensive city management software solution like this. With cities growing at the rates they are, and most resources being finite, management solutions like this are going to be in greater and greater demand.

Photo credit Tom Raftery

Follow @TomRaftery

I attended the 2011 DataCenterDynamics Converged conference in London recently and at it I chatted to a number of people in the data center industry about where the industry is going.

The first of these was Paul-François Cattier of Schneider Electric who talked about the need for standardisation of infrastructure in the data centers to speed up the time to build.

Tom Raftery: Hi everyone, welcome to GreenMonk TV. We’re at DCD London Converge and with me I have Paul-François Cattier, who is VP Data Center Solutions for Schneider Electric. Paul welcome to the show.

Paul-François Cattier: Thank you Tom.

Tom Raftery: So Paul you guys made a bit of an announcement at the show here, can you talk to us a little bit about that?

Paul-François Cattier: Yes we announced what we call the way to bring the standardization and modular IT into data center to bring energy efficiency in the data center. So basically what we think is today’s data center industry is still very immature in its infancy and we need to bring this with stage of maturity to be better efficient.

Tom Raftery: So tell me why do you think a modular infrastructure is better for data centers?

Paul-François Cattier: It’s not really the modular infrastructure that is better for energy efficiency in data center, it is really the standardization that allow this modular IT. In fact what we need to bring into the data center, I think to bring it to maturity is the standardization and to be able to standardize the data center, you need to find a point of granularity of modularity, where you bring the standardization and after a lot of these different data centers have been serving the different business peoples but this level of standardization will allow a lot of CapEx and OpEx efficiency in your data center and you know that most of the OpEx of the data center are in the energy.

Tom Raftery: So talk to me about standardization. What exactly are you talking about when you say standardization, standardization of… ?

Paul-François Cattier: Most of the data center today are designed as unique design. So each time it’s very long process to design data center because it’s a unique design so you have 24 months, 20 months before you decide to do a data center when you are coming to the completion of the data center. So with the standardization you are using subsystems that are completely standardized, manufacture build, manufacture interested, respectable performance and you are using these bricks or these blocks or these Lego if you want of subsystems, standardized subsystems to build your data center.

Schneider Electric Power and Cooling modules

If doing this, you can really build your data center in three to four months with optimized performance and ensure due to this standardization that the management that is needed to really tie the data center physical infrastructure and it’s energy consumption to the effective IT of your data center to be enabled in the data center because as the data center is very much in standardized module, it’s very easy to require this type of management system. So like if you want — if you would like to develop yourself, your GPS in your car, it will — you will spend maybe 20 years before being able to use your own GPS designed by you.

What you do is you share the R&D to have an excellent GPS system that is sold to many, many customer to spread out the cost. So this is what standardization and modularity will bring into the data center world.

Tom Raftery: How far away do you reckon we are from that becoming the norm? I mean you talk about the data center industry currently being quite immature. When do you reckon we will be that much further along that this will be the norm?

Paul-François Cattier: Well we have a long way to go. Today 80% of the new built data center are still built in a very traditional way, that is totally inefficient in terms of CapEx use, in terms of OpEx use and in terms of energy efficiency, and you know that we are working in all the market, we are in Schneider Electric to bring energy efficiency into this market. And really we believe that the standardization in the modularity enables the management aspect of the challenge to be enabled, and — enabled to be enabled of course and allow this energy efficiency and when you save 1 kilowatt at the plug, you save most of the time 3 kilowatt as a generation plant.

Tom Raftery: Great. Well François that’s been fantastic, thanks for coming on the show.

Paul-François Cattier: Thank you very much Tom.

Power Usage Effectiveness (PUE) is a widely used metric which is supposed to measure how efficient data centers are. It is the unit of data center efficiency regularly quoted by all the industry players (Facebook, Google, Microsoft, etc.).
However, despite it’s widespread usage, it is a very poor measure of data center energy efficiency or of a data center’s Green credentials.

Consider the example above (which I first saw espoused here) – in the first row, a typical data center has a total draw of 2MW of electricity for the entire facility. Of which 1MW goes to the IT equipment (servers, storage and networking equipment). This results in a PUE of 2.0.

If the data center owner then goes on an efficiency drive and reduces the IT equipment energy draw by 0.25MW (by turning off old servers, virtualising, etc.), then the total draw drops to 1.75MW (ignoring any reduced requirement for cooling from the lower IT draw). This causes the PUE to increase to 2.33.

When lower PUE’s are considered better (1.0 is the theoretical max), this is a ludicrous situation.

Then, consider that not alone is PUE a poor indicator of an data center’s energy efficiency, it is also a terrible indicator of how Green a data center is as Romonet’s Liam Newcombe points out.

Consider the example above – in the first row, a typical data center with a PUE of 1.5 uses an average energy supplier with a carbon intensity of 0.5kg CO2/kWh resulting in carbon emissions of 0.75kg CO2/kWh for the IT equipment.

Now look at the situation with a data center with a low PUE of 1.2 but sourcing energy from a supplier who burns a lot of coal, for example. Their carbon intensity of supply is 0.8kg CO2/kWh resulting in an IT equipment carbon intensity of 0.96kg CO2/kWh.

On the other hand look at the situation with a data center with a poor PUE of 3.0. If their energy supplier uses a lot of renewables (and/or nuclear) in their generation mix they could easily have a carbon intensity of 0.2kg CO2/kWh or lower. With 0.2 the IT equipment’s carbon emissions are 0.6kg CO2/kWh.

So, the data center with the lowest PUE by a long shot has the highest carbon footprint. While the data center with the ridiculously high PUE of 3.0 has by far the lowest carbon footprint. And that takes no consideration of the water footprint of the data center (nuclear power has an enormous water footprint) or its energy supplier.

The Green Grid is doing its best to address these deficiencies coming up with other useful metrics such as, Carbon Usage Effectiveness (CUE) and Water Usage Effectiveness (WUE).

Now, how to make these the standard measures for all data centers?

The images above are from the slides I used in the recent talk I gave on Cloud Computing’s Green Potential at a Green IT conference in Athens.

Follow @TomRaftery

And we’re back this week with another instalment in our Data Center War Stories series (sponsored by Sentilla).

In this episode I talked to the Green Grid’s EMEA Tech Chair, Harqs (also known as Harkeeret Singh).

The Green Grid recently published a study on the RoI of energy efficiency upgrades for data center cooling systems [PDF warning]. I asked Harqs to come on the show to discuss the practical implications of this study for data center practitioners.

Here’s the transcript of our conversation:

Tom Raftery: Hi everyone, welcome to GreenMonk TV, this is the Data Centers War Stories series sponsored by Sentilla. The guest on the show today is Harkeeret Singh aka Harqs. And Harqs is the EMEA Tech Chair of Green Grid. Harqs welcome to the show.

Harqs: Thank you Tom.

Tom Raftery: So Harqs as the Tech Chair of the Green Grid in the EMEA region, you get a larger overview of what goes on in data centers than some of the previous interviewees we’ve had on the show.

We have been talking about kind of the war stories, the issues that data centers have come across and ways they have resolved them. I imagine you would have some interesting stories to tell about what can go on in data centers globally. Do you want to talk a bit about some of that stuff?

Harqs: The Green Grid undertook a study which implements a lot of good practices that we all talk about in terms of improving air flow management and increasing temperature, putting in variable frequency drives. And what we did is after each initiative, we measure the benefit of — in terms of — from an energy consumption perspective.

And Kerry Hazelrigg from the Walt Disney Company led the study in a data center in the southeast of the US. And we believe that is representative of the most data centers, probably pre-2007, so we think there is a lot of good knowledge in here which others can learn from.

So I am going to take you through some of the changes that were made and some of the expectations but also some of the findings some of which we weren’t expecting. So starting with — there was five different initiatives, the first initiative was implementing the variable speed drives. And what we found was, they installed new CRAC units and CRAH units which they put in variable frequency drives in the standard, there is 14 of them, and then they retrofitted 24 existing CRAHs. And took out the fixed speed drives and put in the variable speed drives.

The expectation was we would find a reduction in energy consumption and fan horsepower. And also there was a potential of maybe looking always the providing of coolant to the right place in the data center. And once we put those in, we found out that they didn’t actually introduce any, any hotspots, which was a positive thing. But some of the things that were a little different from what we expected and the PUE didn’t reflect the savings. That was because there was external factors, things like the external weather which impacted the PUE figure as well. So you need to bear that in mind as you make changes. You need to look at the average across the year.

The other issue that we found was by putting in variable speed drives they found it introduced harmonics to the power systems. And that came through the monitoring tools and so they are putting — they put in filtering to help resolve those harmonics.

The last issue was also around skills, so they had to train the data center staff on using variable frequency drives and actually maintain them. This was the biggest power saving, it was a third of the overall saving and the saving in total was 9.1% of energy consumption and that’s saved some thing in the order of $300,000 in terms of real cash and the PUE went down from 1.87 down to 1.64 by doing these five initiatives.

The second issue was actually putting in the air flow management. So things like the blanking panels and the floor grommets, putting in the cold tiles where they are supposed to be, and that was for around 7 inch cabinets and the findings were that that reduced the cold aisle temperature because you have less mixing, and also increase the temperate on the hot isle in terms of the temperatures going back to the CRAH. So that was interesting.

We saw that being a key enabler to actually increase in temperature, so you have cold to cold aisles and hot to hot aisles because of those mixing. There wasn’t any energy savings for this piece in itself, but it was in this — airflow management activity is an enabler in that it allows you to then do some optimization and also to increase temperature without risk.

The third activity was relocating the sensors that the CRAHs worked off from, way from the return to CRAC and return to CRAH which is what most data centers use today to actually aligning that to sensors on the front of the cabinets. So actually moving from return air to supply and that’s the specification that ASHRAE provides, that’s what we should be controlling, the temperature and the humidity of the air going into the servers. They themselves say they don’t really care about what the temperature is coming out of back of the servers. Well the rest of us do from a — making sure that it’s not too hot for our data center operators.

So what we did was move those sensors to the front of cabinets and what that did was that optimized the fan speeds and actually started to raise the temperature and the cold air that was required by the servers. It did take them a little awhile getting the locations right, so making sure that they have them moving them around as much as possible, looking the CFD to make sure they are optimizing and putting it in the right place eventually. And that was a small improvement, but it was the — again another enabler for increasing temperate. So there was only a few percent improvement by doing that, but what it does is when you start look at increasing temperature you are increasing temperature at the right point.

Tom Raftery: So how much did it increase temperature by — was it like from 20 to 25 or… –

Harqs: That’s a good question as the next initiative they did was, they were increasing the temperature, so I was just about to — so they went from 18° C, which was what their original set point was and they took it up to 22° C. Now obviously that’s still in the middle of the ASHRAE standard. So there is still more scope there to become better. But it wasn’t just increasing the temperature in the room but it was actually increasing the temperature of the chiller plant which — where the biggest savings were, so if you increase the temperature of the room that then allows you to increase the temperature of your chiller plant.

And that’s — they increase the set point of their chiller plant from 6.7° C to just under 8. And what they found was, there was significant savings due to the reduction in compressor and condensor fan power. And what they found was for each and I’m going to do this in degree F because they calculate degree F. So they went from 44 to 46° F. For every degree F they increased the set point of the chiller, they found out that reduced just over 50 kilowatts of chiller energy consumption.

Now in terms of other people’s data centers, they are also — your mileage may vary depending on the configuration and where you are, but that’s what their significant saving was. By doing that what they found was — by doing it this way, where they put the air flow management in place and then they increased temperature in the room, increased the set points of the chiller plant they found that actually there was — that made no significant impact on the data center in terms of hot spots or anything like that. So there is no detrimental impact to the data center by doing this. Obviously the saving of the energy was a positive and saved real money.

Tom Raftery: Alright, Harqs that was great, thanks a million for coming on the show.

Follow @TomRaftery

And we’re back this week with the third instalment in our Data Center War Stories series (sponsored by Sentilla).

In this episode of the series I am talking to Jerry Sweeney. Jerry is Managing Director of Cork Internet eXchange (CIX). CIX is a small, currently co-lo, data centre located in Cork, Ireland (and full disclosure – I was a co-founder of CIX).

I love Jerry’s story about the chiller compressors coming on for the first time after 12 weeks – free cooling rocks! (watch the video, or see the transcript below!).

Here’s the transcript of our conversation:

Tom Raftery: Hi everyone and welcome to GreenMonk’s DataCenter War Stories sponsored by Sentilla. The guest in the show is CIX’s Jerry Sweeney. Jerry is Director of Cork Internet eXchange. Jerry welcome to the show.

Jerry Sweeney: Thank you for having me Tom.

Tom Raftery: Jerry can you tell me a little bit about Cork Internet eXchange, how old it is, what kind of size you are talking about?

Jerry Sweeney: Cork Internet eXchange was conceived in 2006, in September of 2006, construction occurred in 2007, and it opened for business in March 2008. So it’s 3-years-old now.

We have two rooms on the technical floor area, one of them is kitted out, it’s 3,000 square feet and the other one is available for expansion and that’s also 3,000 square feet, as well as there is approximately 7,000 or 8,000 square feet for the services, offices, call center and so on.

So, whatever that works out at 12 and seven, so it’s about 19,000 square feet in total. Eventually it will be a 240 rack facility. At the moment we have about 75 occupied racks. To date it’s exclusively a collocation facility, but we are now getting into the infrastructure as a service and platform as a service business.

Tom Raftery: In the building of a facility of that size what kind of — what are the most pressing kind of issues you come across typically day to day?

Jerry Sweeney: I suppose your question had the concept of size and — so we are a very small data center, and I suppose trying to scale the expenses against our revenue stream is probably an issue with a company this size. So running 24/7 shifts, so I would say scale is probably our biggest single problem, and having people with the right resources, and having the facility occupied. If you have a 1000 racks okay, then you can spread those costs over a greater number of customers and a greater number of racks.

Tom Raftery: Any interesting things that you — any interesting problems you happened to cross and solutions you came up with to solve them?

Jerry Sweeney: We live in a city Tom with 160,000, 170,000 people. We — all of the data centers in Ireland are basically clustered around Dublin, all of the connectivity that comes into Ireland is located or lands in Dublin.

So, remoteness and scale okay were huge problems for us when we started off. And one of the big issues for us okay was to get adequate connectivity into the building so that we would be taken seriously. And we came up with a strategy very early on and the strategy was to — initially before we focused on being a data center that we focused on being a regional internet connectivity center.

And the name of the business is very interesting; the name of the business is Cork Internet eXchange. We registered the URL which was the Cork Data Center, but we never used it, and the reason for that is because Cork Internet eXchange was more vital to us at start up then the Cork Data Center.

So, in order to justify gigabit connectivity in the back-haul costs around that, we had to get serious volumes of IP transit through the building first. And we have a 30 meter, it was 24 meters initially, but we just added six meters to it this year, our address is Hollyhill and that’s a clue, we are on top of a hill. That enabled us to sign up every single wireless internet service provider in the region.

So, all of the non-incumbent supply broadband homes and businesses in Cork take their connectivity out of here and we see that as being about 20,000 homes and businesses. So that was a huge win for us in the 2000 — in early 2009. By the time we got to say March 2009, which would be a year after we opened, we had our IP transit up in the gigabits and that made cost effective procurement of transit sensible.

And it was at that time that we noticed a growth in the — people took us more seriously as a data center, because of the connectivity. We had the resilience from design in, what we didn’t have is, we didn’t have connectivity at a price okay, and at a quality level that made us attractive.

So, I think that probably was the… and if we hadn’t been successful of getting that connectivity issue; then, I don’t think we would have been able to scale as a data center.

Tom Raftery: Can you talk a little bit about some of the interesting concepts that went into the design of the data center?

Jerry Sweeney: The concept of building the data center was started in September 2006, and we made a decision in 2006 to go for cold aisle containment and today that seems like a really kind of standard idea — the argument is now do you go for hot aisle or cold aisle containment. But in 2006, it was actually even alternate hot and cold aisles were considered novel at that time.

So it seemed like a remarkable unusual thing. So we built it from the ground up with the cold aisle containment as a strategy. Also because we are located in Cork, which is a mild – neither hot or cold climate, we have 11 degrees as an ambient temperate, average for the year and the difference between summer and winter is not enormous, so we are able to take advantage of an awful lot of free cooling.

Even in the summer at night time we can usually do free cooling here and for much of the winter okay, our chillers never start. We know that our chillers did not start from the — from November of 2010 until a warm sunny afternoon in February. So free cooling okay, took us for whatever number of weeks, that is six and six — about 12 weeks, without ever starting a compressor.

We were shocked when the compressor cut in, what’s that noise, okay.

Tom Raftery: Jerry it’s been fantastic. Thanks a million for coming on the show.

Jerry Sweeney: Yeah, it’s my pleasure; Tom, thank you.

Follow @TomRaftery