Wednesday, July 29, 2009

Google Cools Data Center Without Chillers; Data Center Pros Weigh-in

Google’s chiller-less data center in Belgium has received a lot of buzz. The facility relies upon free air cooling to keep its servers cool and will shift the computing load to other data centers when the weather gets too hot.

It's an approach that stands to greatly improve energy efficiency. However, as e-shelter explained to Techworld, there are some risks. For instance, it's possible that airborne particulates could cause havoc with hard disk drives and dampness from heavy humidity could cause electrical problems. To see what other data center professionals think of this cooling strategy, I posed the following question to the Computer Room Design Group on LinkedIn:

Is Google's Chiller-less Data Center the wave of the future, or is this approach too risky for most businesses to accept?

Here’s what some of our group members had to say…

Mark Schwedel, Senior Project Manager at Commonwealth of Massachusetts:

Please note that Google is doing many thing that are not available in current data centers they do not have UPS - They are doing battery backup on each server with 12 volt battery - SO will this be the future? Only when the rest of world can delivery the same aspect as Google.

Sean Conner, Datacenter Professional Services Consultant:

Google's design is well suited for an expansion of their cloud environment. However, it's clear that the facility in question does not run as the same level of criticality as most dedicated or hardened sites. This works well in an environment that can tolerate minor equipment loss and failure.

However, most dedicated sites host applications and data that would suffer, should similar equipment loss occur. So, the two approaches cannot be truly compared. It's like trying to compare the heart to the left hand. Both are useful. But if the left hand fails, you probably don't die.

Perhaps a larger question to ask is: What applications, data, or entire enterprises could migrate to a cloud environment? Those that can stand to gain huge savings from Google's approach.

Dennis Cronin, Principal at Gilbane Mission Critical:

This entire dialog is moot because the way of the future is back to DIRECT WATER COOLED PROCESSORs. All these sites chasing the elusive "FREE" cooling will soon find out that they cannot support the next generation of technology.I suspect that there will be a lot of finger pointing when that occurs with even more adhoc solutions.We need to stick to quality solutions that will support today's AND tomorrow's technology requirements.

David Ibarra, Project Director at DPR Construction:

There is a tremendous pressure on large enterprise customers ( social, search,etc) to use the same fleet of servers for all of their applications. The IT architects behind the scene are now been asked to stop been "geeks" and changing hardware every 3 years and try to make use of what we have or improve with systems that are lower cost. The recession is also amplifying this trend. A lot of water cooled servers and demonstrations held last year have gone silent due to cost and also standardization on hardware for the next 5 years. A lot of large DC customers understand the water cooling technology and are early adopters; however realities have driven the effort elsewhere within their organizations. Customer are pushing high densities ( +300W/sqft) using best of class techniques: containments, free cooling,etc. Plus large scale operators are understanding that the building needs to suit the server needs so there is a shift on how a building is configured. Chiller less data centers have existed since 2006 in countries such as Canada, Ireland, Germany, Norway. Data centers will be coming online at the end of this year in the US that are chiller less and cooling tower less and with a extraordinary reduction of air moving equipment.

Nitin Bhatt, Sr. Engineer at (n)Code Solutions:

Every single Data Center is unique in its own set-up. To adopt some technology which is suiting to one geographical location could not be a wise decision. It is wise to be "Orthodox" rather than lossing the business. If someone can afford the outage / shifting of the work load to DR site or to some other sites as a result of the thermal events, yes they can look into FREE COOLING w/o Chillers. We can save the energy used by chillers having VFDs and room temperature based response to chillers. It is good to have chillers as backup to the Free Cooling.

So what do you think? Please share your experience by posting a comment here, or by continuing the discussion in the Computer Room Design Group on LinkedIn.

Wednesday, July 22, 2009

LinkedIn Discussion on Power Usage Effectiveness (PUE)

Last week I posted the following discussion question in our Computer Room Design networking group at I’m really impressed with the response from group members, so I’d like to share their thoughts with you here:

How can the industry address problems with the reporting of Power Usage Effectiveness (PUE) without undermining the usefulness of the metric?

In a recent post in Data Center Knowledge, Rich Miller points out that the value of Power Usage Effectiveness (PUE) as the leading 'green data center' metric "has become fuzzy due to a disconnect between companies’ desire to market their energy efficiency and the industry’s historic caution about disclosure." [Source:]

What are your thoughts on redefining PUE? Are additional refinements the answer? Or does increasing the complexity of PUE undermine the usefulness of the metric?


• Gordon Lane, Facilities Coordinator at Petro Canada, explained:
I don't see a real value in PUE.

If you leave unused servers powered on you can keep your PUE low.

Assume you have a PUE of 2
2MW total power consumption gives you 1 MW for servers.
If you can reduce your server consumption to 0.75MW by turning off comatose servers total consumption reduces to 1.75MW and gives you a PUE of 2.33

I know there would be some reduction in a/c power usage due to less heat output from the turned off servers but if you are using legacy a/c units with no VFD style control then you will not get a corresponding electrical consumption reduction.

• Scot Heath, Data Center Specialist, weighed in with:
PUE is difficult to measure in mixed facilities, is muddied by configurations such as the Google every-server-has-a-battery and varies widely with Tier level. A universal measurement that combines both IT capability (total Specmarks for example) and availability with respect to energy consumption would be most useful. PUE does have the advantage of being quite easily understood and for controlled comparisons (like tier level, etc.) is very useful.

• Dave Cole, Manager of Data Center Maintenance Management and Education Services at PTS, responded:
Gordon and Scot bring up very good points. I have mixed feelings about PUE. The concept is easily understood - we want to maximize the power that is actually used for IT work. The interpretation of the value is easy to understand - lower is better (or higher is better in the case of DCiE). The problem I see is that it's almost been made too simplistic. You still have to know your data center and the impact of the decisions you make in regards to design and operation. You can actually raise your PUE by virtualizing or by turning off ghost servers as Gordon pointed out. What needs to be understood is that when you lower the demand side, you should also be making corresponding changes to the supply side. At the end of the day, PUE can be valuable as long as you are also looking at what impacts the value. You need to be able to answer the question of WHY your PUE is changing.

What are your thoughts on the value of Power Usage Effectiveness (PUE) as a metric? Please share your experience by posting a comment here, or by continuing the discussion in the Computer Room Design Group on LinkedIn.

Monday, July 20, 2009

LinkedIn Discussion on Eliminating the Battery String

Thanks to everyone who’s participated in our Computer Room Design networking group at so far! We’re off to a great start, with over 200+ members joining in the first two weeks. I’d like to share highlights from one of our recent discussions…

Kevin Woods, Director of Business Development and Sales at i2i, asked:

Eliminating the Battery String? Does anyone have experience/opinion on the viability of the UPS/CPS systems? They incorporate a flywheel in between the generator and engine and in cases of power interruption, the flywheel uses kinetic energy to power the generator for up to 30 seconds while the engine is engaged.


• Mark Schwedel, business partner at EMC and advisor for Green Rack Systems, recommended taking a look at the patent for an improved UPS/CPS system, which employs a high-efficiency uninterrupted power supply function integrated with an engine-generator set that combines both short term protection against momentary power interruptions with longer term power generation.

• Gordon Lane, Facilities Coordinator at Petro Canada, shared his experience:
Not a direct comparison to gen/engine set up but I have a flywheel UPS system that has been in service for 23 years. Very reliable, change the bearings every 50000 hours - about 6 years - and we have just about completed a program of taking the MGs out for cleaning and re-insulation.

Obviously coming to end of life, 20 yrs was estimated life, but the serviceability has been phenomenal.

Certainly looking to replace with a similar system and I believe Caterpillar has a flywheel UPS solution that they integrate into their diesel offerings.

• Jason Schafer, Senior Analyst at Tier1 Research, explained in part:
My personal issue with flywheel solutions, aside from the reliability that both sides will argue, is that 30 seconds simply isn't enough time when you are talking about the criticality most datacenters need. The most common argument relates to allowing time to manually start a generator; and flywheel advocates will say "if a generator doesn't start in 30 seconds it's not very likely that it's going to start in 20 minutes" - I disagree with this. I've seen, on more than one occasion, where generator maintenance was being performed and through human error the EPO switch on the generator was mistakenly left pushed in. There's no way anyone is going to identify the problem and fix it in 30 seconds - I'd be surprised if anyone even got to the generator house in 30 seconds after a power outage. Minutes, however, are a different story.

I'm not saying that flywheels and CPSs don't have their place - I think they do, or rather will in large scale in datacenters, but we're not quite there yet. When virtualization plays a part in the redundancy and fault tolerance of a datacenter, where ride-through in the event of a power outage is more of a convenience than a necessity (a-la Google's datacenters - they can lose an entire facility and continue on for the most part), you'll see flywheels gain more traction.

What are your thoughts on the viability of the UPS/CPS systems? Please share your experience by posting a comment here, or by continuing the discussion in the Computer Room Design Group on LinkedIn.

Thursday, July 16, 2009

Introducing PTS’ Data Center Education Series

How extensive is your knowledge about all aspects of your data center? With our newly launched Data Center Education Series, you will never look at your IT and support infrastructure the same way again.

PTS’ Data Center Education Series will help you better assess problems in your data center by providing you with substantive knowledge that you can take back to your data center to improve operations, availability, and efficiency - ultimately reducing operating cost and improving service delivery to your users.

The education series provides students with comprehensive, vendor-neutral, module based training led by the data center design experts from PTS. We discuss the most pertinent topics in the data center industry, tying in case studies and real world situations to provide the knowledge you need to understand, operate, manage, and improve your data center.

The Standard Training Series is a three (3) day class held multiple times per year at major cities across the United States, Canada, and Europe. Our next session will take place in Midtown NYC from September 15-17th -- visit our site to view the agenda. Can’t make it to NYC? We'll also be coming to Chicago (October 21-23) and Dallas (December 7-9). I encourage you to reserve your seat today, as space is limited.

The education series will cover the following topics:

• Fundamentals of Data Center Cooling
• Fundamentals of Data Center Management
• Fundamentals of Physical Security
• Fundamentals of Fire Protection
• Fundamentals of Data Center Power
• Fundamentals of Data Center Maintenance
• Fundamentals of Designing a Floor Plan
• Fundamentals of Data Center Cabling
• Fundamentals of Energy Efficiency

Priced at only $1,795 per student, the training includes all course materials in addition to a continental breakfast and lunch each day. Additionally if you attend with other colleagues from work, you'll all receive a 10% discount. You'll realize an ROI quickly from this invaluable and intimate knowledge, in which straight from data center experts in this in-depth, intimate training series.

Data Center Education Series – Customized for your needs!

We also offer education programs customized to your IT team’s needs. If you have a large group and need training, we can come to you and present those topics of most interest to you! Choose your desired location (typically your own facility). Choose the topics you want to see, including any or all of the available topics from the standard 3-day training class.

In addition, if you have a topic in mind you don't see currently listed in our offerings, we'll build it for you for only a nominal fee to cover time and material costs.

The Customized Training Series is priced at $15,000 for 2 days or $20,000 for 3 days plus travel expenses. In addition to the training, you have option to purchase a one-day data center site assessment for $5,000. This assessment will be performed prior to the training in order to allow the training to address issues found in the assessment.

Please join us on LinkedIn & Twitter

PTS is excited to provide our peers with a new online forum in which to discuss the planning, design, engineering, and construction of data centers and computer rooms.

If you’ve been reading our blog for a while, you may already be aware of our Facebook Page at (A big ‘thank you’ to everyone who’s added themselves as fans!) Today, I’m happy to announce that PTS is further expanding our online presence with the goal of facilitating the open exchange of ideas among small-to-medium sized data center and computer room operators.

At the forefront of this effort is the newly created Computer Room Design networking group on You can check it out by visiting Hosted by the consultants and engineers at PTS Data Center Solutions, the group is an open forum in which professionals can share industry-related news, ideas, issues and experiences.

Membership is free and open to all professionals and vendors in the computer room and data center industry. We hope that industry leaders will look at this as an opportunity to share knowledge, discover new services and opportunities, and expand their networks.

So far, our networking group on has attracted broad interest, gaining more than 100 members in the first week alone. Featured discussions include best practices for consolidation strategies, how to combat downtime in the data center, and industry concerns regarding the Power Usage Effectiveness (PUE) metric.

This thought leadership is further supported on PTS’ Twitter profile ( which features the latest industry news, highlights from the LinkedIn networking group, and insights from our engineers. If you’re on Twitter, please send us a message and we’ll be sure to follow you back!