2011 Data Center Trend Reviews

Sorry for not getting back to you all with the review of the 2011 predictions.  All in all, I think we hit it on the head on two items and missed on the third.

Cloud Computing
The greatest increase in data center, space and utilization was clearly in the cloud space last year.  When the 2011 prediction was made, the inference is that cloud, as a service or as a cornerstone to the technology plan for a business.  In review of where the money flowed last year, the kings of the hill were Amazon, Facebook, Google, Apple, Zynga, Linked In and other service systems.  In 2011, Siri, Apple Cloud, Facebook user growth, the Zynga and LinkedIn expansions simply offer irrefutable evidence that the big growth is in geopgraphically independant, enterprise-level application delivery - the ALL/ANYWHERE/ANYTIME pardigm.  We also take this as a major and continuing shift to SaaS, IaaS and PaaS, depending on your needs.

Modular Data Centers
Don't get me wrong, our modular business was very robust this last 12 months.  We have viewed the modularity of the data center speaks to the larger issue of outsourcing real estate solutions to the wholesale colocation providers.  This is related to scale, cost/SF or cost/kW and the generally efficiencies you see when you scale a business to, say 18M square feet of data center space.

Creativity from Manufacturers
Dead missed this one.  We expected breakthroughs on design or manufactured systems this year.  What we observed was the general failure of new designs when faced with a highly-efficient, capacitive IT load.  While not to count the high efficiency data center dead from an electrical standpoint, it certainly is handicapped by having to serve a heterogeneous IT equipment system.  Where Amazon, Facebook, Apple or Google (herein and now referred to in this blog as the CLOUD GIANTS) can vertically integrate their application, platform and facility solutions to come up with some very efficient models, not everyone can enjoy those savings unless they are willing to make a profound change in their hardware systems.  Everyone's looking at this, but it may take a decade for our traditional enterprise-level users in the banking and manufacturing businesses to be able to churn out older hardware for newer systems current seen in the CLOUD GIANTS.

We also observed limited advancement in critical power and cooling systems outside of Facebook Open Compute.  While we applaud the use of new long-life battery systems, the power systems we have seen are merely derivative in their advancement.

And packaging into a modular building does NOT count!  That's just a modification of a construction technique. 

Server Power Supplies Take an Ugly Turn

Well, it's official.  I have lived long enough to see history repeat itself. 

So this is what they mean when people mention the benefit from the wisdom gained from age.  It's more like the movie "Groundhog Day", where you are stuck in a tragic comedy.

On several occasions in the past six months, our team has run into significant power quality issues with 3rd order harmonics in a few of our data centers, both in the wholesale colo and enterprise markets.  Third order harmonics in data centers is like saying that it's going to be hot in the desert - you know it's true, but you hope it's not. 

We have come to discover that the commodity servers from Dell have been the culprit of some pretty ugly load-generated, third-order harmonics.. 

Power factor correction has been the boogeyman for servers and data center for the past three years.  Stemming from the need to run engine-generators at or below unity power factor, or a lagging power factor.  Power factor improvement was a byproduct of effciency improvements demanded by the marketplace.  This is accomplished by capacitors.  However, the amount of capacitance being injected into critical power system is starting to exhibit some darned strange things.  In solving one problem, the server manufacturers, as well as the OEM power supply vendors, are creating others.  It's like the Will Smith movie, "I Am Legend", where the vaccine for cancer ends up turning everyone into flesh-eating zombies.  Great.

Here's what we are seeing.

• Rectifier walk-in setting up a harmonic feedback on a 10s interval in transformer-less UPS modules.  And this is happening on 100% transistorized IGBT modules, not old 6-pulse units.
• Harmonics in large multiples of the funamental, just like the 1990's and early Dot Com years.  In fact, we dug 15 year old waveforms out of the files, and the new power supplies have a waveform exactly just like the crap that used to be supplied back then.
• Power factor correction in the server power supply is now "swtichable" by the customer.  If the power factor correction and power supply were engineered correctly (as so aptly pointed out by one of our Engineering Directors Clint Summers), why is it switchable in a passive critical power electrical system.

When you peel back all the hysterics back, the bottom line is that the THD and the amplification of the harmonic disturbance into large multiples of what's expected and from the fundamental frequency are well outside of industry-stated standards.

Most data center electrical infrastructure systems can deal with relatively strong harmonics.  We have seen some issues with zig-zag neutral reference operations.  But we have also seen this pop up in PDU operations as well.  Our position here is that the Z-Z would not have an issue if the power supplies worked according to the industry tolerances in the first place.  It's like arresting a home owner for murder when the mob dumped a dead body on the guy's front lawn.  You are blaming the wrong party.

The best of all, your power professionals have sufficient power quality metering that allows complete diagnosis of these issues.  So, don't let that server sales guy pull an ole' on you.  The matador goes home and the bull ends up being steak dinners for everyone!

Make sure that you demand the power spectrum data for the server power supply, both in a waveform format as well as the summary of frequency values, similar to what we're showing below.

And for you reading enjoyment, please see the redacted power survey from the site.  I thought I was back in 1999 when I read it.  Note the PU multiples!  Holy crap! 

Sorry for the inverse order of the pages.  Enjoy!

And Merry Christmas everyone!

The Pros and Cons of Modular Data Centers

I knew I forgot something.  Back in June, Rosendin delivered a presentation at the 7x24 Exchange National meeting in Orlando, Florida.  The purpose of this presentation is to speak to modular and containerized data center solutions, as compated to traditionally built mission critical environments.  Exclusive to this presentation is the cost, schedule and form factors for each, something that has not been revealed before this time anywhere in industry writings.

Enjoy.

Rosendin Electric receives Two IT Innovation Awards for 2011!

Rosendin Electric receives two significant construction technology award for 2011.

For our integrated BIM and Estimating systems, we made the Information Week Top 500 Innovators for 2011:

http://www.informationweek.com/iw500/2011

For our Vela productivity tracking system, we were honored by ConstrucTech with a Gold Vision award in 2011 for our productivity and QA/QC tracking systems:

Well done Rosendin!

This Time the WSJ Gets It Wrong In Data Center Real Estate

On Friday, July 7th, Anton Troianovski wrote an article critical of the growth of the wholesale data center market in the US. Search the Wall Street Journal on line, July 7th, and look for the article:

Storage Wars - Web Sparks Data-Center Boom

And Anton, only civilians hyphenate data center.

While not one to sharply opine on non-technical articles, I will have to say Anton got it flat wrong. I've been design and building data centers for the past 25 years. Yes, I'm that old, but I've been fortunate to watch the information business move from centralized, big-iron (and I love big iron - it's just a HUGE server) cooled by water and powered by 415 Hz motor-generators.

The one indisputable fact this that information, in parallel to Moore's Law, has grown exponentially. In the past 25 years, I've seen really only seen three off years - 1984, 2001 and 2009, all resulting from contractions in the finance markets that supports the capital side of business IT growth or the general shock of recovering from some economic trauma or force majeure. And it continues to do so today.

If you really look at enterprise computing (and the derivative benefit that shows up on your smart phone or desktop system vis a vis capabilities and capacity), it is likely to be the most innovative portion of our society. One of the major changes with information in the past 10 years is that it has moved from a being a business tool to being a societal tool. Think about your life without a smart phone, social networking, the ability to shop on line, stream movies, the ability to streamline and render more efficient any operation, smart homes - the evidence is endless. Have you seen the new HTC phones with the Facebook location button built right into the face of the phone?

On an organic level, you buy that 10 TB HDD for your desktop at home to hold your music, photos and movies. 10 years ago, a 1 TB HDD was unthinkable for your home - what in the heck would I put on that big a drive? And you likely have an online or redundant drive or cloud backing up your personal information. Holy cow - personal disaster recovery! Unthinkable ten years ago, uncommon five years ago, and common today.

And this does not even speak to the drivers of the MC real estate business in the last five years, cloud computing, social networking and content. Your Facebook data, your NetFlix movies and your outsourced processing at Amazon are living somewhere safe, clean, quiet and continuously cooled and powered.

So, all of your virtual needs need to be fulfilled somewhere. And that somewhere is a in data center.

I think the major flaw in that thinking is that the article views data center real estate as any other real estate operation. Agreed on a basic level, as everyone need to live, shop and work somewhere. But using a traditional RE industry observation to the data business would mean you would have to say that I need to build houses for a population that doubles every three years. The math and your sensibilities simply don't work at that rate of growth.

Putting aside the commoditization and packaging of data center space by your leading REITs like Digital Realty Trust (DLR), Vantage or T5 Partners into more homogeneous and appropriate solutions for large business, the fact remains for every Facebook that moves out of a data center, there are several more parties coming right in behind them. The article also missed the fact that data operations move first from the small back-of-house hand-built set ups, to managed services, then to wholesale colocation on a more homogenized platform and then to either widespread colo or bespoke builds. So the service levels vary for the stage of life and IT need your business has.

Older facilities may not be readily available for newer technologies and power densities or cooling requirements without substantial capital outlay and physical renewal. That's fine and no building is relevant forever unless it's an icon. Only the truly dodgy and speculative data center real estate solution stand empty. The better built, better designed and better operated stand the test of time and are refitted on a routine basis.

People may also state that IT efficiencies will overcome the fixed asset limitations of UPS, generator or cooling plant size - that IT can compact faster than it's growing. We've never seen an IT operation be able to "out compress" the data center's physical infrastructure for space, power or cooling. The growth rate simply trumps the compaction rate. We attribute this directly to the further depth and sophistication of both the business and personal use of information, especially storage. In short, with that freed up IT support capacity (power, cooling & space), IT guys will put in more stuff (I know, great technical term). And that stuff tends to be quite a bit more powerful, viewed both in IT capabilities and W/SF.

As a result, we now offer the Mazzetti Algorithm:
The pace of IT system growth will always trump the ability of that IT system to compact or virtualize over a three year period.
And that's why the data center real estate business is in no danger of oversupply.

Thanks for reading. C U soon.

Funny if you are in the business

Thank you Scott Adams for making my day!

Jesus, It Fit In the Virtual Building Model!

I received an interesting article from our CEO last month. It dealt with a successful lawsuit against the design professionals on a design-bid-build project. The design team, as part of their contracted services, produced a virtual building model indicating the location of a majority, if not all MEP building systems. The crux of the complaint is that the MPE design, while indicating the X-Y-Z positions of MPE components such as duct work, cable tray and sprinkler piping, did not organize the location of the systems in a manner where they may be easily installed.

The case was argued in the contention that the design cold fit in a certain volume, but could not be installed commensurate with the typical constructor's practices. It becomes the argument of the Swiss watch. Commercial and industrial building works need sufficient space to allow for the nuances of the builders' approach. This does not mean that the design should be so roomy to be wasteful. It simply means that systems are organized in the Z-axis in a manner to allow for logical, sequential work during the build.

The design professionals lost the case.

This is not an uncommon issue. The design professionals are at a distinct disadvantage in two ways (and it's symptomatic of their roles and not their capabilities):

• The MPE design is based on generic sizes of equipment based on what is typically an open spec for major MPE systems and end items. The design professionals have to design around the worst-case scenario for sizing.


• Construction means and methods have evolved significantly over time. While engineered documents are definitive of scope and sizing of MPE systems, engineered documents do not, as a routine, possess enough coordination between trades to allow the DP to present a set of plans coordinated with sufficient detail or sequence to allow for smooth installation.

How do you get around this on fixed-priced, engineered work? Well, the sinister side of the the issue is that the building gets bigger to make room for the works and to prevent this type of claim. Dismissing alternate contracting methods like IPD or design-assist, the primary focus should be on the Z-axis organization of systems, especially in the substructure and above ceilings, as well as focusing on vertical and horizontal transitions.

Containers - Packaging or Gift

Since I have been asked to present on the metrics for container and modular data center deployments at the 2011 7x24 Exchange Spring Conference (www.7x24exchange.org/spring11/brochure), one of the main issues I have with non-stick built deployments is the metric used to measure success. Here are a few nuggets for thought:

1. How much are you willing to pay for speed of deployment?


2. Who's selling you the solution and what is the motivation for said sale outside of the obvious desire for more business?


3. Seeing the cost for speed of deployment, what premium are you willing to pay for speed to deploy new assets over excess space in existing facilities?


4. Are you happy with the container or modular solution in the long run, operationally speaking?


5. Do you understand the dynamic of costs and who and where the money falls on a container or modular approach?


6. What is your financial metric for the $$/rack position or MIP or TB in the data center?

So, ask yourself this question:

Are you happier with the gift or the wrapping?

Open Source Computing - The Facebook Dynamic & The Changes Coming To Hardware

It's been over a week since Facebook published all of their server, power supply and data center design. They can be found on the Open Compute Project at http://www.opencompute.com/. I've been involved with Facebook since before the patent application, and now for those who follow my blog, you can see where we have been going for the new dyanmic in facilities design. It has been a priviledge to be involved with the server, power supply and facility teams during the past 24 months. It has been a deeply rewarding experience with a group of very capable professionals who, in addition, are great folks to boot. I personally swear to the Jay Park cocktail napkin design story, as I have personally seen said napkin. Someday, it will be placed in the Facebook Musuem or Hall of Fame. Let's be honest. Facebook and the social media and web content businesses who work, write and manage custom code for their businesses have the unique opportunity to integrate their application and server approaches and then spin the critical facility solution to speak and serve those apps and hardware most pointedly. So what do we do about the rest of us that have apps that run on a host of systems. We think the first, best step is hardware standarization to a commodity system. The operating system is a commodity already, whether that's open or closed protocol. There's no difference to this and any app that cold run on a variety of similar platforms provdied by a large set of suppliers. This work was started over a decade ago, at Google. In short - "We buy a lot of servers, we need them cheap and identical and fast". We recall seeing the earliest Google servers in the Equinix facilities in Ashburn, VA, in the early 21st century. The Facebook solution seeks to further this via mass industrialization the process while opening the approach to the marketplace. What's different today is the push from Facebook, and others in that business, to drive hardware to a new dynamic. The Facebook and Google approaches aren't for everyone, today. There's not a data center manager in the world who would not kill to run a homogeneous platform system, where they only have to worry about the apps (the FB facilities possess a very simply power and cooling infrastructure). We all recognize that these business are cloud computing on steroids. While some may dismiss cloud computing in the private sector, there's no doubt that we should look at the open compute solution as an ability to render the hardware side of the business as a commodity. Based on shear volume and magnatude, Google, Microsoft, Apple or Yahoo! simply hit it harder than the rest of us for this type of server. However, some folks can employ this as a transitional step going forward as they migrate and reshape a more mature systems inventory, not unlike blades 10 years ago (another Google idea). We also feel is in the same vein as open source operating systems, like Linux, or widely adopted systms such as Windows and Java. The question remains is how long this will take for the hardware manufacturers' take up the slack - or resist the trends and pay open compute lip service. The sinister alternative is that the open source server initiatives become an adjunct to the hardware manufacturers traditional business. We feel that the traditional server and storage business is better integrated to their cost and manufacturing models and likely to be far more profitable for them for the next several years. We do agree that the Facebook Open Source initiative does seek to place the impetus and control of the data processing business back into the hands of the users. WHAT A CONCEPT! The direct benefit of this is a known electrical and mechanical solution set for the critical facility environment. This means cost, schedule and operations become very well know, by a large community, very, very quickly. What is also means is that server computer power (as viewed in MIPS/kW or TB/kW) should plummet in price. That's not going to make some folks happy.

2011 Prediction Comments - January 20, 2011

I got several comments on my predictions for 2011, specifically on containers and innovations.

Addressing manufacturer innovation, I believe what really started the trend was a drive to differentiate. The Eaton 9395 UPS power module drive many to multi-mode operation with their introduction of their Economy mode. That, and Eaton FINALLY produces a 750+ kVA module, after many of us had waited almost 10 years for it.

This then was taken one step further by Chloride/Emerson with the module that can operate on inverter, economy or line interactive mode. You can argue who started the trend, but none the less, we have newer and sharper tools to use.

For containers, I think it's just as simple as several owners are now using containerized IT packs. While the first systems did not have on board critical power, many of the applications we now see do (we just finished a job of this ilk). So, some of the ME packaging is following the IT trends. And since several firms now produce ME packs as part of their business (Rosendin certainly does!), there are many influential parties that seek to place their newly developed products. We do have some sales "gravity" toward this issues.

What we have not seen is an adequate or prolonged discussion on the long-term viability of containerized utility systems. While there is certainly a shelf life to IT packs and technology, that shelf life is far longer for MEP system. Similarly, I don't think anyone has taken a critical eye to the maintainability and life-cycle cost of the pack versus a stick-built system that is typically around for 10 years, at a minimum.

We will acknowledge that we can build a containerized MEP solution more quickly than a stick-built solution, but that should not be the only criteria for selection.