Ron Hughes' Blog http://www.datacenterdynamics.com/ME2/Console/XmlSyndication/Display/RSS.asp?xsid=60C873D85D2C470DACCB96DD61EA6FAA Ron Hughes' Blog The hidden obsolescence issues with collocation data centers http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=79303F8E647E4073A063F889C689CEC2 I’ve been involved in the design, construction and operation of collocation data centers for the last 15 years. During that time I’ve seen tremendous changes in the way space is marketed and sold. In the last several years, I’ve become increasingly aware of two principal challenges facing collocation data center operators: 1. The age of support equipment in many of the data centers 2. The power and cooling design densities of most of the collocation data centers Most collocation data centers is use today were designed and built from 1997 – 2002.  This was the height of the collocation data center boom.  These data centers are now anywhere from 8 to 12 years old and have UPS systems and cooling systems that reflect the technologies that were in use at the time. Most are not very energy efficient and have levels of redundancy and reliability that are significantly below those of current data center designs. That means that primary infrastructure supporting clients is getting close to the end of its useful life.  Obviously different systems have different useful lives, but most of them are in the 15 to 20 year range.  Upgrading these systems will be very difficult and in some cases impossible.  Changing out UPS systems, batteries and cooling systems while still keeping the clients on line and with the levels of redundancy that colo’s are contractually obligated to maintain is going to be a challenge. To add to the challenge, most collocation data centers from that era were designed for 80 to 100 watts per square foot or less.  This compares to the 200 watts per square foot or more that most data centers today are designed for.  Since most collocation data center space today is sold based on power densities per square foot, these legacy data centers are faced with some tough choices.  Either sell the entire space at lower power densities or leave a portion of the space unused since there is insufficient power and cooling to support it. Upgrading the power and cooling to support higher densities is a logical consideration.  Practically, it’s often extremely difficult if not impossible to do this.  Some of the challenges include: Is there sufficient capacity from the utility – Going from 100 to 200 watts per square foot requires the utility to double the size of the feed into the facility. Is there space for all of the additional support equipment – Doubling the capacity nearly doubles the amount of space required for support equipment depending on the design.  Most data centers don’t have the space either inside or outside the facility for the additional equipment. Can the existing chilled water piping support the increase in density – As you might expect, most of the existing chilled water piping is sized to accommodate the design load and won’t support significant increases in the cooling capacity. Will the raised floor height support the increased cooling loads – Most of the legacy collocation data centers were designed with 24 inch raised flooring.  Will this support a significant increase in cooling loads or do you have to go to supplemental cooling systems? While some owners are proactively upgrading their data centers and adding capacity, many seem to be ignoring the issues.  We have walked through numerous collocation data centers recently in looking for space for a client.  All of them had support equipment that was 6 to 12 years old.  Not one of them had any plans in place to replace or upgrade any of the major support equipment, with the exception of batteries.  The useful life of their support equipment seemed to increase exponentially the more I questioned them about it.  If this trend continues, I could see a two tiered pricing structure for collocation data centers.  One for newer data centers with power densities that support what customers are asking for.  A second tier for legacy data centers whose infrastructure can’t support the rack loads that customers are now looking for.  Another possibility is that data centers would be considered fully leased based on power availability, not data center space. This could result in some data centers that have half of their raised floor space vacant.  Considering that most collocation data centers aren’t particularly energy efficient already, this could make the problem even worse.  I’m not sure what the final outcome will be, but I am confident that this is going to be an issue that comes up more frequently as data centers age.  Before making a decision on leasing data center space, customers should look at the age of the support equipment, the densities the facility can support and the energy efficiency of the support systems.  Leasing space in a data center that will have to be upgraded while you are occupying your space, may not be a wise decision. Fri, 18 Dec 2009 00:00:00 EST 79303F8E647E4073A063F889C689CEC2

1. The age of support equipment in many of the data centers

2. The power and cooling design densities of most of the collocation data centers

Most collocation data centers is use today were designed and built from 1997 – 2002.  This was the height of the collocation data center boom.  These data centers are now anywhere from 8 to 12 years old and have UPS systems and cooling systems that reflect the technologies that were in use at the time. Most are not very energy efficient and have levels of redundancy and reliability that are significantly below those of current data center designs.
That means that primary infrastructure supporting clients is getting close to the end of its useful life.  Obviously different systems have different useful lives, but most of them are in the 15 to 20 year range.  Upgrading these systems will be very difficult and in some cases impossible.  Changing out UPS systems, batteries and cooling systems while still keeping the clients on line and with the levels of redundancy that colo’s are contractually obligated to maintain is going to be a challenge.

To add to the challenge, most collocation data centers from that era were designed for 80 to 100 watts per square foot or less.  This compares to the 200 watts per square foot or more that most data centers today are designed for.  Since most collocation data center space today is sold based on power densities per square foot, these legacy data centers are faced with some tough choices.  Either sell the entire space at lower power densities or leave a portion of the space unused since there is insufficient power and cooling to support it.

Upgrading the power and cooling to support higher densities is a logical consideration.  Practically, it’s often extremely difficult if not impossible to do this.  Some of the challenges include:

Is there sufficient capacity from the utility – Going from 100 to 200 watts per square foot requires the utility to double the size of the feed into the facility.

Is there space for all of the additional support equipment – Doubling the capacity nearly doubles the amount of space required for support equipment depending on the design.  Most data centers don’t have the space either inside or outside the facility for the additional equipment.

Can the existing chilled water piping support the increase in density – As you might expect, most of the existing chilled water piping is sized to accommodate the design load and won’t support significant increases in the cooling capacity.

Will the raised floor height support the increased cooling loads – Most of the legacy collocation data centers were designed with 24 inch raised flooring.  Will this support a significant increase in cooling loads or do you have to go to supplemental cooling systems?

While some owners are proactively upgrading their data centers and adding capacity, many seem to be ignoring the issues.  We have walked through numerous collocation data centers recently in looking for space for a client.  All of them had support equipment that was 6 to 12 years old.  Not one of them had any plans in place to replace or upgrade any of the major support equipment, with the exception of batteries.  The useful life of their support equipment seemed to increase exponentially the more I questioned them about it. 

If this trend continues, I could see a two tiered pricing structure for collocation data centers.  One for newer data centers with power densities that support what customers are asking for.  A second tier for legacy data centers whose infrastructure can’t support the rack loads that customers are now looking for. 

Another possibility is that data centers would be considered fully leased based on power availability, not data center space. This could result in some data centers that have half of their raised floor space vacant.  Considering that most collocation data centers aren’t particularly energy efficient already, this could make the problem even worse. 

I’m not sure what the final outcome will be, but I am confident that this is going to be an issue that comes up more frequently as data centers age.  Before making a decision on leasing data center space, customers should look at the age of the support equipment, the densities the facility can support and the energy efficiency of the support systems.  Leasing space in a data center that will have to be upgraded while you are occupying your space, may not be a wise decision.

]]> Continuous cooling – Reality versus assumptions http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=75A18AC0A7F64A20BB2024A0F81B42BD One of the great things about our job at CDCDG is that we not only get to design data centers, but we get to test them once they are up and running. As a result, we get to see how they actually perform and compare that to the assumptions we made during the design process. The other day, one of my partners, Richard Greco and I were discussing the requirement for continuous cooling. Both of us were once proponents of it. I even wrote an article entitled “High density data centers require continuous cooling.” I’m not so sure anymore. Providing continuous cooling for a data center is fairly expensive and fairly complex given the need to provide thermal storage and put pumps and air handlers on UPS power. Most designs set up a separate UPS system just for the continuous cooling loads. All this comes at a pretty hefty price. The question is if it is really necessary? As densities have increased, we have seen more and more articles about the requirement for continuous cooling in data centers. If you go back 5 years, people were saying that it was a requirement once you got above 5kW per rack. That proved to not be true, but the general consensus came to be that if you got above 10kW it was a requirement. The thought was that if the cooling was out for 30 to 60 seconds, the temperature would skyrocket and you could never catch up in time to prevent equipment damage. That has proven not to be true either. We have tested numerous data centers with racks loads of between 8 to 12kW and can tell you that not only did the temperatures not skyrocket during short outages, but the only noticeable change was a lack of air movement. While the server fans continued to move air through the servers (since they were on a UPS), the major air handlers or CRAC units being off line for 30 to 60 seconds did not seem to have much of an impact. Even with a 2 to 3 minute interruption in cooling, the temperature rises were between 2 to 3 degrees. While the temperatures do indeed go up, they are nowhere near the thermal runaway conditions that had been predicted and are certainly within the allowable rate of change. Now these are recent designs, so they are all hot aisle/cold aisle arrangements and most used the ceiling as a return plenum. Some of them had hot aisle containment, which really helps to minimize the issue. If you are isolating the hot air from the data center and routing it into a plenum, the temperature in the data center would remain fairly constant. Especially if we are only talking about the 15 to 30 seconds it takes for a backup generator system to come on line. I’ve heard some theories about why the continuous cooling assumption has proven to be invalid in all but a few instances (super-computer centers being one of them). The thermal mass in a data center is pretty significant, raised flooring, concrete slabs, walls, racks themselves provide lot of thermal mass, and it’s all at 72 to 75 degrees. It takes awhile for those surfaces to heat up. Another is that the fans in the servers themselves provide enough air movement to keep them cool for a short period of time it takes for the generators to start and the cooling to come back on line. The new energy efficient designs should help as well. Although there are a number of different designs, some of them provide for very low velocity air and rely on the server fans to distribute the air. Most of them also dump the hot air to the exterior and use 100% outside air for cooling a large percentage of the time. Certainly these designs should be even less prone to the thermal runaway scenarios. So at what point does continuous cooling become a requirement? I’m not really sure that we have an answer. There are a lot of variables that are site and design specific. One thing we can say is that a blanket statement that all data centers with rack loads of 5 – 10 kW and above require continuous cooling is simply not reality based on our experiences. I would love to get some feedback on what others have seen at their data centers. Tue, 03 Nov 2009 00:00:00 EST 75A18AC0A7F64A20BB2024A0F81B42BD
Providing continuous cooling for a data center is fairly expensive and fairly complex given the need to provide thermal storage and put pumps and air handlers on UPS power. Most designs set up a separate UPS system just for the continuous cooling loads. All this comes at a pretty hefty price. The question is if it is really necessary?

As densities have increased, we have seen more and more articles about the requirement for continuous cooling in data centers. If you go back 5 years, people were saying that it was a requirement once you got above 5kW per rack. That proved to not be true, but the general consensus came to be that if you got above 10kW it was a requirement. The thought was that if the cooling was out for 30 to 60 seconds, the temperature would skyrocket and you could never catch up in time to prevent equipment damage. That has proven not to be true either.

We have tested numerous data centers with racks loads of between 8 to 12kW and can tell you that not only did the temperatures not skyrocket during short outages, but the only noticeable change was a lack of air movement. While the server fans continued to move air through the servers (since they were on a UPS), the major air handlers or CRAC units being off line for 30 to 60 seconds did not seem to have much of an impact. Even with a 2 to 3 minute interruption in cooling, the temperature rises were between 2 to 3 degrees. While the temperatures do indeed go up, they are nowhere near the thermal runaway conditions that had been predicted and are certainly within the allowable rate of change.

Now these are recent designs, so they are all hot aisle/cold aisle arrangements and most used the ceiling as a return plenum. Some of them had hot aisle containment, which really helps to minimize the issue. If you are isolating the hot air from the data center and routing it into a plenum, the temperature in the data center would remain fairly constant. Especially if we are only talking about the 15 to 30 seconds it takes for a backup generator system to come on line.

I’ve heard some theories about why the continuous cooling assumption has proven to be invalid in all but a few instances (super-computer centers being one of them). The thermal mass in a data center is pretty significant, raised flooring, concrete slabs, walls, racks themselves provide lot of thermal mass, and it’s all at 72 to 75 degrees. It takes awhile for those surfaces to heat up. Another is that the fans in the servers themselves provide enough air movement to keep them cool for a short period of time it takes for the generators to start and the cooling to come back on line.

The new energy efficient designs should help as well. Although there are a number of different designs, some of them provide for very low velocity air and rely on the server fans to distribute the air. Most of them also dump the hot air to the exterior and use 100% outside air for cooling a large percentage of the time. Certainly these designs should be even less prone to the thermal runaway scenarios.

So at what point does continuous cooling become a requirement? I’m not really sure that we have an answer. There are a lot of variables that are site and design specific. One thing we can say is that a blanket statement that all data centers with rack loads of 5 – 10 kW and above require continuous cooling is simply not reality based on our experiences. I would love to get some feedback on what others have seen at their data centers. ]]> Update from the Road http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=A89B2CFADED9479FBED99B12440BD274 It’s been a while since my last blog. I blame it on the travel. I have been out of town for 5 of the last 6 weeks. Mostly travel for work, but some of it was for fun. I just got back from a week in Bonneville racing a motorcycle in the World Finals. Now that was fun! We were able to set records in two classes. Of course I was just “the rider”, my friend Mike Bungay supplied the bikes and the knowledge. All I had to do was hold the throttle wide open, stay on the course and make sure I didn’t’ crash. I added a picture of me at the starting line to my blog. I’ve also done some fun things at work. The great thing about this job is that we get an opportunity to work on some fascinating designs. Some of our current projects include: Designing a modular 10,000 square foot data center that could be installed at multiple sites for a client of ours. They want a Tier 4 data center designed for 200 watts per square foot. It could be modular support systems that could be installed in an existing building or could be a totally new modular design. We are currently working on several options that include fully certified Tier 4 data centers as well as Tier 3 plus designs. The great thing about this project is that we have a blank slate. We can design it with any variety of support systems as long as they meet the Tier 4 overall rating. I’ll name the client after we get a little further along on the design. We prepared a conceptual design for a new Supercomputer ata center for the National Center for Atmospheric Research. This extremely energy efficient data center will have one of the lowest PUE’s around when it’s up and running. Given my previous rant against misleading estimated PUE’s, I’m not giving specific numbers. Among the highlights will be a 10 foot raised floor height. We also have a client that we are designing a data center for that is very high density. Unlike the design we did for National Center for Atmospheric Research, this data center is not a supercomputer center. However, the design calls for up to 40kW per rack, the use of flywheels in lieu of batteries and an extremely sophisticated monitoring and controls system. This client wishes to remain anonymous. I also got to tour the massive new data center in Salt Lake City for an online auction firm. While I can’t name the client, I can tell you that it is a very impressive site. Still under construction, it is designed so that it could be quadrupled in size, to an estimated 200,000 square feet or more. I’m off to finish commissioning another 10,000 square foot expansion of the KIO Networks Data center in Queretaro, Mexico and then will be home for a week. After that it’s the Data Center Dynamics Conference in Mexico City. I’ll check in with another blog sometime before that. Tue, 13 Oct 2009 00:00:00 EST A89B2CFADED9479FBED99B12440BD274

I’ve also done some fun things at work. The great thing about this job is that we get an opportunity to work on some fascinating designs. Some of our current projects include:

Designing a modular 10,000 square foot data center that could be installed at multiple sites for a client of ours. They want a Tier 4 data center designed for 200 watts per square foot. It could be modular support systems that could be installed in an existing building or could be a totally new modular design. We are currently working on several options that include fully certified Tier 4 data centers as well as Tier 3 plus designs. The great thing about this project is that we have a blank slate. We can design it with any variety of support systems as long as they meet the Tier 4 overall rating. I’ll name the client after we get a little further along on the design.

We prepared a conceptual design for a new Supercomputer ata center for the National Center for Atmospheric Research. This extremely energy efficient data center will have one of the lowest PUE’s around when it’s up and running. Given my previous rant against misleading estimated PUE’s, I’m not giving specific numbers. Among the highlights will be a 10 foot raised floor height.

We also have a client that we are designing a data center for that is very high density. Unlike the design we did for National Center for Atmospheric Research, this data center is not a supercomputer center. However, the design calls for up to 40kW per rack, the use of flywheels in lieu of batteries and an extremely sophisticated monitoring and controls system. This client wishes to remain anonymous.

I also got to tour the massive new data center in Salt Lake City for an online auction firm. While I can’t name the client, I can tell you that it is a very impressive site. Still under construction, it is designed so that it could be quadrupled in size, to an estimated 200,000 square feet or more.

I’m off to finish commissioning another 10,000 square foot expansion of the KIO Networks Data center in Queretaro, Mexico and then will be home for a week. After that it’s the Data Center Dynamics Conference in Mexico City. I’ll check in with another blog sometime before that.

]]> Estimated PUE – Great expectations unconfirmed http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=819F956A244142DB96C13F1A0FD92B9D I have been watching with amusement as the data center owners and engineering firms fall over themselves claiming to have the “Lowest PUE/Most energy efficient” data center. It started with hardware manufacturers and software companies claiming they were designing their data centers for PUE’s below 1.3. This was largely scoffed at since none of the details were provided to substantiate their claims. When they were finally provided, a number of very competent engineers questioned the validity of the numbers. I won’t get into the arguments about how they were calculated, but I do think that this is a perfect example of why calculated or estimated PUE numbers are meaningless for anything more than a design goal. Estimated PUE’s are usually based on best case operating conditions and are significantly impacted by the levels of redundancy built into the design. They are typically calculated based on the design load of the data center. In actuality, it might be 3 to 4 years before the loads in the data center reach the design loads. During this time the efficiency of the electrical and mechanical systems might be significantly lower than they would be at full design load. A UPS system that is 94% efficient at 90% load might be 80% efficient at 20%. The same concept is true for chillers as well. Yet in reality, the first several years of operation, 20 to 30% loads might be closer to the truth than 90%. If you are using outside air for a large part of the cooling, the calculations typically assume that you will be on outside air 100% of the time that it is available. That is probably not even close to being achievable, since the realities of operating a data center may prevent you from always using the most energy efficient means of cooling. There may be extreme weather conditions that require you to operate with the outside air dampers closed. Since almost all of the low PUE data centers use some form of air side or water side economizer to achieve their numbers, not operating in that mode can have a significant impact on the actual PUE numbers. A more meaningful calculation would be to measure the PUE on an annualized basis after the facility is on line and at its design load. This measured PUE would be much more real world than the estimated PUE that was calculated during the design process. This is not always politically correct, since this may not occur for years. Can you imagine the conversation you would have with a marketing person about waiting for 3 to 4 years to announce that you have the “worlds most energy efficient data center”. Having said that, I don’t think that estimated PUE’s aren’t valuable. Having an idea of how your new data center might perform could affect some of the decisions made in the design process. An important part of establishing the owner’s project requirements document is to have a design goal for PUE. I’m just saying that we should be realistic about how achievable calculated PUE’s actually are. If we use them, we should always qualify them by adding that this is an estimate and that the actual PUE achieved would depend on a number of operating factors that might be beyond the control of the design engineers. Doing this might also cut down on the number of debates about the who really has the most energy efficient design. Wed, 19 Aug 2009 00:00:00 EST 819F956A244142DB96C13F1A0FD92B9D
Estimated PUE’s are usually based on best case operating conditions and are significantly impacted by the levels of redundancy built into the design. They are typically calculated based on the design load of the data center. In actuality, it might be 3 to 4 years before the loads in the data center reach the design loads. During this time the efficiency of the electrical and mechanical systems might be significantly lower than they would be at full design load. A UPS system that is 94% efficient at 90% load might be 80% efficient at 20%. The same concept is true for chillers as well. Yet in reality, the first several years of operation, 20 to 30% loads might be closer to the truth than 90%.

If you are using outside air for a large part of the cooling, the calculations typically assume that you will be on outside air 100% of the time that it is available. That is probably not even close to being achievable, since the realities of operating a data center may prevent you from always using the most energy efficient means of cooling. There may be extreme weather conditions that require you to operate with the outside air dampers closed. Since almost all of the low PUE data centers use some form of air side or water side economizer to achieve their numbers, not operating in that mode can have a significant impact on the actual PUE numbers.

A more meaningful calculation would be to measure the PUE on an annualized basis after the facility is on line and at its design load. This measured PUE would be much more real world than the estimated PUE that was calculated during the design process. This is not always politically correct, since this may not occur for years. Can you imagine the conversation you would have with a marketing person about waiting for 3 to 4 years to announce that you have the “worlds most energy efficient data center”.

Having said that, I don’t think that estimated PUE’s aren’t valuable. Having an idea of how your new data center might perform could affect some of the decisions made in the design process. An important part of establishing the owner’s project requirements document is to have a design goal for PUE. I’m just saying that we should be realistic about how achievable calculated PUE’s actually are. If we use them, we should always qualify them by adding that this is an estimate and that the actual PUE achieved would depend on a number of operating factors that might be beyond the control of the design engineers. Doing this might also cut down on the number of debates about the who really has the most energy efficient design.]]> Things not to do for a sucessful data center project http://www.datacenterdynamics.com/ME2/Audiences/dirmod.asp?sid=&nm=&AudID=E5BD2FF22AF74DF3A0D5F4E519A61511&type=Blog&mod=View+Topic&mid=67D6564029914AD3B204AD35D8F5F780&tier=7&id=F51CCCA464D74DDFBEAD9E0028ED3283 Make equipment selections based on what you were told by the vendor selling you the equipment – I know this will come as a shock to you, but the vendor is actually trying to make money off of you. Have someone with experience working with multiple competing vendors review what is being proposed. While their proposed solution might be the best one for you, look at all of the alternatives before making a decision. Confuse calculated PUE with the PUE you will actually achieve – I’ll buy anyone dinner who can show me a project where the “estimated” PUE was not significantly lower than the actual PUE measured when a facility was up and running. PUE calculations are always made using best case scenarios. Base your design loads on the newest, highest density piece of equipment in your data center or that you have heard might be coming out shortly – The design load for your data center will have one of the biggest impacts on the cost of your project than anything other than redundancy. Base the design on your current loads, plus your long range technology plans. If you don’t have any, start the planning process. If you significantly over or underestimate your requirements for power or raised floor square footage, you could cost yourself millions of dollars. Base your budget on your ten year old assumptions as to what a data center should cost – The primary factors that determine the cost of your data center is the power density and redundancy requirements. Your 15 year old data center designed as a Tier 2, 50 watts per square foot facility is not a good basis for what your Tier 3 or 4, 250 watts per square foot data center will cost. Ignore the experts you hired and rely instead on in house “expertise” for critical design decisions – while we don’t know everything, those of us in the data center design business have designed dozens of data centers since your in-house experts did their last one. Believe me, experience does matter and recent experience matters even more. You would not believe some of the things we’ve heard from in house experts about how a data center should be designed. Establish your owners project requirements after your project is completed and it is too late to implement any changes – The time to determine how you want your data center to function is before you have it designed. The first thing you should do is determine what your requirements are for your data center for things like power density, redundancy, reliability, Tier rating, energy efficiency, LEED certification level, etc. You can’t design a data center if you don’t know how the client wants it to perform. Chose a site for your data center based on cost alone – Data center site selection is based on a large number of factors. The cost of land is only one of them. Lack of natural threats, (earthquakes, floods, fires, tornadoes, hurricanes), the cost of power, the environmental conditions that allow the use of outside air for cooling, the availability of fiber and power are all factors that go into an informed decision. Too many clients have their real estate department find a less expensive piece of property and then try to make it work instead of finding the right piece of property. Specify a new technology that is not in use anywhere else – This is my favorite. We have clients that go into a project with the idea that they are going to do it differently than anyone else. Innovation can be a great thing. Being different just for the sake of saying you do it differently is not. If a technology has not been used anywhere else try it as a pilot project before you implement it in your data center. With a $50 to $100 million dollar investment or more, your job may just depend on it. Try to save money by reducing the cost and complexity of the testing and commissioning – You are making a huge investment in your new data center. Don’t you want to make sure that it is functioning properly before you move in. That is the biggest difference between an office building and a data center. You can fix things in an office building after you move in, you can’t in a data center. Don’t be like the “un-named” airport that cut out electrical testing to save $100k only to have a piece of switchgear blow up. The result was a 6 week delay in opening up the new terminal at a cost of millions of dollars. Design your data center around a specific type of data processing hardware - Dwight Yoakum does a song called “Things change” . That is a perfect description of why you need to design a data center to accommodate any hardware currently available or that might be available in the future. Design in flexibility. While today’s hardware might be air cooled, it could be liquid cooled in the future. While today’s server power density might be 5kW per rack, it might be 20 in the future. Designing in the flexibility to change as the technology in use changes is a critical part of best practice data center design. Thu, 13 Aug 2009 00:00:00 EST F51CCCA464D74DDFBEAD9E0028ED3283 Make equipment selections based on what you were told by the vendor selling you the equipment – I know this will come as a shock to you, but the vendor is actually trying to make money off of you. Have someone with experience working with multiple competing vendors review what is being proposed. While their proposed solution might be the best one for you, look at all of the alternatives before making a decision.

Confuse calculated PUE with the PUE you will actually achieve – I’ll buy anyone dinner who can show me a project where the “estimated” PUE was not significantly lower than the actual PUE measured when a facility was up and running. PUE calculations are always made using best case scenarios.

Base your design loads on the newest, highest density piece of equipment in your data center or that you have heard might be coming out shortly – The design load for your data center will have one of the biggest impacts on the cost of your project than anything other than redundancy. Base the design on your current loads, plus your long range technology plans. If you don’t have any, start the planning process. If you significantly over or underestimate your requirements for power or raised floor square footage, you could cost yourself millions of dollars.

Base your budget on your ten year old assumptions as to what a data center should cost – The primary factors that determine the cost of your data center is the power density and redundancy requirements. Your 15 year old data center designed as a Tier 2, 50 watts per square foot facility is not a good basis for what your Tier 3 or 4, 250 watts per square foot data center will cost.

Ignore the experts you hired and rely instead on in house “expertise” for critical design decisions – while we don’t know everything, those of us in the data center design business have designed dozens of data centers since your in-house experts did their last one. Believe me, experience does matter and recent experience matters even more. You would not believe some of the things we’ve heard from in house experts about how a data center should be designed.

Establish your owners project requirements after your project is completed and it is too late to implement any changes – The time to determine how you want your data center to function is before you have it designed. The first thing you should do is determine what your requirements are for your data center for things like power density, redundancy, reliability, Tier rating, energy efficiency, LEED certification level, etc. You can’t design a data center if you don’t know how the client wants it to perform.

Chose a site for your data center based on cost alone – Data center site selection is based on a large number of factors. The cost of land is only one of them. Lack of natural threats, (earthquakes, floods, fires, tornadoes, hurricanes), the cost of power, the environmental conditions that allow the use of outside air for cooling, the availability of fiber and power are all factors that go into an informed decision. Too many clients have their real estate department find a less expensive piece of property and then try to make it work instead of finding the right piece of property.

Specify a new technology that is not in use anywhere else – This is my favorite. We have clients that go into a project with the idea that they are going to do it differently than anyone else. Innovation can be a great thing. Being different just for the sake of saying you do it differently is not. If a technology has not been used anywhere else try it as a pilot project before you implement it in your data center. With a $50 to $100 million dollar investment or more, your job may just depend on it.

Try to save money by reducing the cost and complexity of the testing and commissioning – You are making a huge investment in your new data center. Don’t you want to make sure that it is functioning properly before you move in. That is the biggest difference between an office building and a data center. You can fix things in an office building after you move in, you can’t in a data center. Don’t be like the “un-named” airport that cut out electrical testing to save $100k only to have a piece of switchgear blow up. The result was a 6 week delay in opening up the new terminal at a cost of millions of dollars.

Design your data center around a specific type of data processing hardware - Dwight Yoakum does a song called “Things change” . That is a perfect description of why you need to design a data center to accommodate any hardware currently available or that might be available in the future. Design in flexibility. While today’s hardware might be air cooled, it could be liquid cooled in the future. While today’s server power density might be 5kW per rack, it might be 20 in the future. Designing in the flexibility to change as the technology in use changes is a critical part of best practice data center design.]]>