Knowledge Library

Here, we make the Myth-of-the-Month video on data center metrics available via Podcast.

• 
• Click to play or Right-Click | Save As to download: The Myth of Data Center Metrics

I have gotten a lot of response from the post on my thoughts on Data Center regulation.   Many of the comments in a response to an Infoworld article focused on the disbelief of regulations particularly targeting data centers.  A Greener Computing article felt that because the current administration is very tech-savvy they wouldn’t do anything to hurt data centers.  In fact the exact quote was:

I can understand Manos’ concerns, but I think he’s on the wrong track. The federal government is very unlikely to issue strict green regulations related to data centers. And if they do regulate them in some way, the regulations will no doubt be reasonable. The current administration is very technology-savvy — after all, the current Secretary of Energy Steven Chu was recently the director of the Lawrence Berkeley National Laboratory, whose work was heavily dependent on its data center. Chu did some great work related to Green IT when at the labs. He knows what can and can’t be done — and will make sure that data centers aren’t hamstrung with unnecessary regulation.

I guess for clarity sake I should state unequivocally that I do not believe that Data Centers will specifically be targeted or singled out for regulation.   Domestically here in the United States the EPA has kicked off its Energy Star Data Center evaluation which looks to study data centers as a sector, and something may come out of that, but in all honesty that wont be for some time.  I think the more immediate threat is in the efforts around Carbon Cap and Trade.  As the Greener Computing Article calls out, it was front and center at the G8 meetings.   With the UK leading the charge and the only real legislation on the books in this space, it would be hard for the other countries not to use it as the base for their programs.   My previous post focuses specifically on the fact that Data Centers will end up being significant contributing factors to Carbon metrics for companies.  Data Center Managers just aren’t thinking about it, and wont be until its far too late. » Continue

First let me start out with the fact that the need for datacenter measurement is paramount if this industry is to be able to manage itself effectively.   When I give a talk I usually begin by asking the crowd three basic questions

1) How many in attendance are monitoring and tracking electrical usage?

2) How many in attendance measure datacenter efficiency?

3) How many work for organizations in which the CIO looks at the power bills?

The response to these questions has been abysmally low for years, but I have been delighted by the fact that slowly but surely, the numbers have been rising.  Not in great numbers mind you, but incrementing.  We are approaching a critical time in the development of the data center industry and where it (and the technologies involved) will go. » Continue

As an industry, we have been talking about it for some time.  Some claimed it would never come and it was just a bunch of fear mongering. Others like me said it was the inevitable outcome of the intensifying focus on energy consumption.   Whether you view this to be a good thing or bad thing its something that you and your company are going to have to start planning for very shortly.  This is no longer a drill.

CRC – its not just a cycle redundancy check

I have been tracking the energy efficiency work being done in the United Kingdom for quite some time and developments in the Carbon Reduction Commitment (CRC).  My recent trip to London afforded me the opportunity to drive significantly harder into the draft and discuss it with a user community (at the Digital Realty Round table event) who will likely be the first impacted by such legislation. For those of you unfamiliar with the initiative let me give a quick overview of the CRC and how it will work. 

The main purpose of the CRC is a mandatory carbon reduction and energy efficiency scheme aimed at changing energy use behaviors and further incent the adoption of technology and infrastructure.  While not specifically aimed at Data Centers (its aimed at everyone) you can see that by its definition Data Centers will be significantly affected.  It was introduced as part of the Climate Change Act 2008.

In effect it is an auction based carbon emissions trading scheme designed to operate under a Cap and Trade mechanism.  While its base claim says that it will be revenue neutral to the government (except of course for penalties resulting from non-compliance), it provides a very handy vehicle for future taxation and revenue.  This is important, because as data center managers you are now placed in a position where you have primary regulatory reporting responsibilities for your company.  No more hiding under the radar, your roles will now be front and center.                             

All organizations including governmental agencies who consume more than 6000 MWh in 2008 are required to participate.  The mechanism is expected to go live in April 2010.  Please keep in mind that this consumption requirement is called out as MWh and not Megawatts.  What’s the difference? Its energy use over time for your whole company.  If you as a data center manager run a 500 kilowatt facility you account for almost 11% of the total energy consumption.  You can bet you will be front and center on that issue. Especially when the proposed introductory price is £12/tCO2 (or $19.48/tCO2).  Its real money.  Again, while not specifically focused on data centers you can see that they will be an active contributor and participant in the process.  For those firms with larger facilities, lets say 5MW of data center space – dont forget to add in your annual average PUE – the data centers will qualify all to themselves.

For more information of the CRC you can check out the links below:

• A Good Deck with a high level overview can be found here
• The wikipedia definition
• For alot of additional detail the CRC user guide is a great tool

While many of you may be reading this and feel poorly for your brothers and sisters in Great Britain while sighing in relief that its not you, keep in mind that there are already other mechanisms being put in place.  The EU has the ETS, and the Obama Administration has been very public about a similar cap and trade program here in the United States.  You can bet that the US and other countries will be closely watching the success and performance of the CRC initiative in the UK. They are likely to model their own versions after the CRC (why invent the wheel over again, when you can just localize to your country or region).  SO it might be a good idea to read through it and start preparing how you and your organization will respond and/or collect.

I would bet that you as a Data Center Manager have not been thinking of this, that your CIO has not thought about this, the head of your facilities group has not thought about this.  First you need to start driving awareness to this issue.    Next we should heed to a call to arms.

One of the items that came out during the Roundtable discussions was how generally disconnected government regulators are to the complexities of the data center.   They want to view Data Centers as big bad energy using boxes that are all the same.  When the differences in what is achievable from small data centers to mega-scale facilities are great.  Achieving PUEs of 1.2x might be achievable for large scale Internet firms who control the entire stack from physical cabling to application development,  banks and financial insitutions are mandated to redundancy requirements which force them to maintain scores of 2.0. 

Someone once decried to me that data centers are actually extremely efficient as they have to integrate themselves into the grid, they generally purchase and procure the most energy efficient technologies, and are incented from an operating budget perspective to keep costs low.  Why would the government go after them before they went after the end users who typically do not have the most energy efficient servers or perhaps the OEMs that manufacture them.  The simple answer is that data centers are easy high energy concentration targets.   Politically going after users is a dicey affair and as such DCs will bear the initial brunt.

As an industry we need to start involving ourselves in educating and representing  the government  and regulatory agencies in our space.   While the Green Grid charter specifically forbids this kind of activity, having a Data Center industry lobby group to ensure dumb things wont happen is a must in my opinion.  

Would love to get your thoughts on that.

/Mm

This post is a portion of my previous post that I broke into two.   It more clearly defines where I think the market is evolving to and why companies like Digital Trust Realty will be at the heart of the change in our industry.

The Birth of Utilities and why a Century ago will matter today and forward…

I normally kick off my vision of the future talk with a mention first to history (my former Microsoft folks are probably groaning at this moment if they are reading this).   I am a huge history buff.  In December of 1879, Thomas Edison harnessed the power of electricity for the first time to light a light bulb.  What’s not apparent is that this “invention” was in itself not complete.  To get this invention from this point to large scale, commercial application required a host of other things to be invented as well.   While much ado is made about the successful kind of filament used to ensure a consistent light source, there were no less than at least seven other inventions to make electric light (and ultimately the electric utility) practical for everyone.  Invention of things like the parallel circuit, an actual durable light bulb, an improved dynamo, underground conductor networks, devices to maintain constant voltage, insulting materials and safety fuses, the light socket, the on/off switch, and a bunch of other minor things.   Once all these things were solved, the creation of the first public electricity utility was created.  On September of 1882, the first commercial power station, located on Pearl Street in lower Manhattan opened its doors and began providing light and electrical power to all customers within the “massive” area of one square mile.   This substation was a marvel of technology staffed with 10s of technicians, maintaining the complex machinery to exacting standards.   The ensuing battle between Direct Current and Alternating Current was then created and in some areas still continues today. More on this in a bit.

A few years earlier a host of people were working on what would eventually become known as the telephone.   In the United States this work is attributed to Alexander Graham Bell and its that story I will focus on here for a second.  Through trial and error Bell and his compatriot Watson, accidently stumbled across a system to transfer sound in June of 1875.  After considerable work on refinement the product launched (There is an incredibly interesting history of this at SCRIBD), and after ever more additional trial and error the first telephonic public utility was created with the very first central office coming online in January of 1878 in New Haven, Connecticut.  This first central office was a marvel to behold.  Again the extremely high tech equipment with a host of people ensuring that telephonic utility was always available and calls were transferred appropriately.  Interestingly by 1881 only 9 cities with populations above 10,000 were without access to the telephone utility and only 1 above 15,000!  That is an adoption rate that remains boggling even by today’s standards. 

These are significant moments in time that truly changed the world in the way we live every day.   Today we are the birth of another such utility.  The Information Utility.   Many people I have spoken to claim this “Information Utility” is something different. It’s more of a product, because it uses existing utility services. Some maintain that its truly not revolutionary because its not leveraging new concepts.   But the same can be said of those utilities as well.   The communications infrastructure we use today whether telephone or data has its very roots in the telegraph.  The power utilities have a lot to thank the gas-lamp utilities of the past for solving early issues as well.  Everything old is new again and everything gets refined into something newer and better.  Some call this new Information Utility the “cloud”, others the Information-sphere, others just call it the Internet.  Regardless what you call it, access to information is going to be at your finger tips more today and tomorrow than it has ever been before.

Even though this utility is built upon existing services, this utility too will have its infrastructure.  Just as the electric utility has its sub-stations and distribution yards, and the communication utilities have central offices, so too will data centers become the distribution mechanism for the the Information Utility.   We still have a lot of progress to make as well.   Not everything is invented or understood yet.  Just as Edison had to invent a host of other items to make electricity practical, and Bell and Watson have to develop the telephone, the telephone ringer (or more correctly, thumper or buzzer), so to does our information Utility have a long way to go.  In some respects its even more complicated than its predecessors as their was not burdened with legislation and government involvement that would affect its early development.  The “Cloud” does.

And that innovation does not always come from a select few.  Westinghouse and his alternating current eventually won out over direct current because it found its killer app and business case.   Alternating current was clearly the technically superior and better for distribution. They had even demonstrated generating power at Niagara Falls and successfully transferred that power all the way to Buffalo, New York! Something direct current was unable to do.  In the end, Westinghouse worked with appliance manufacturers to create devices that used alternating current.  By driving his killer app (things like refrigerators), Edison eventually lost out.  So too will the cloud have its killer apps.  The pending software and services battle will be interesting to note.  However what is interesting to me is that it was the business case that drove adoption and evolution here.  This also modified how the utility was used and designed.   DC substations gave way to AC substations, what used to take scores of people to support has dwindled to occasional visitations and pre-scheduled preventative maintenance.   At the data center level, we cannot afford to think that these killer applications will not change our world.    Our killers applications are coming and it will forever change how our world does business.  Data Centers and their evolution are at the heart of our future.

On Fogs, Mist, and the Clouds Ahead . . .

After living in Seattle for close to 10 years, you learn you become an expert in three things.  Clouds, rain, and more clouds.   Unlike the utilities of the past, this new Information Utility is going to be made up of lots of independent cloudlets full of services.  The Microsoft’s, Google’s and Amazon’s of the world will certainly play a large part of the common platforms used by everyone, but the applications, products, content, customer information, and key development components will continue to have a life in facilities and infrastructure owned or controlled by companies providing those services.    In addition, external factors are already beginning to have a huge influence on cloud infrastructure.  Despite the growing political trend of trans-nationalism where countries are giving up some of their sovereign rights to participate in more regionally-aware economics and like-minded political agendas, that same effect does not seem to be taking place in the area of taxation and regulation of cloud and information infrastructure.  Specifically as it relates to electronic or intellectual property entities that derive revenue from infrastructure housed in those particular countries or do so (drive revenue) off of online activity of citizens of those nations.

There are numerous countries today that have, or are, seriously engaged in establishing and managing their national boundaries digitally online.  What do I mean by that?  There is a host of legislation across the globe that is the beginning to govern the protection and online management of their their citizens through legislation and mandates in accordance with their own laws.   This is having (and will continue to have) a dramatic impact on how infrastructure and electronic products and services will be deployed, where that data is stored,  and how revenue from that activity can and will be taxed by the local country.  This level of state exercised control can be economically, politically, or socially motivated and cloud services providers need to pay attention to it.   A great example of this is Canada which has passed legislation in response to the U.S. Patriot Act.   This legislation forbids personally identifiable information (PII) of Canadian citizens to be housed outside of the boundaries of Canada or perhaps more correctly, forbids its storage in the United States.    There are numerous laws and legislation making their way across Europe and Asia as well.   That puts an interesting kink in the idea of a world wide federated cloud user-base where information will be stored “in the cloud”.  From an infrastructure perspective it will mandate that there are facilities in each country to house that data.  While the data storage and retention challenge is an interesting software problem to solve the physical fact that the data will need to remain in a local geography will require data centers and components of cloud infrastructure to be present.  I expect this to continue as governments become more technically savvy and understand the impact of the rate of change being caused by this technology evolution. Given the fact that data centers are extremely capital intensive only a few players will be able to deploy private global infrastructures.  This means that the “information sub-station” providers will have an even more significant role in driving the future standards of this new Information Utility. One might think that this could be a service that is ultimately provided by the large cloud providers as a service.   That could be a valid assumption however, there is an interesting wrinkle developing around taxation or more correctly exposure to double taxation or multiple-country taxation that those large providers will face.   In my opinion the federation of “information substation” providers will provide the best balance of off-setting taxation issues and still providing a very granular and regionally acceptable way to service customers. That is where companies like Digital Realty Trust are going to come in and drive significant value and business protection.

I watch a lot of these geo-political and economic developments pretty closely as it relates to Data Center and infrastructure legislation and will continue to do so.  But even outside of these issues, the “cloud” or whatever term you like will continue to evolve and the “channels” created by this paradigm will continue to drive innovation at the products and services level.  Its at this level where the data center story will continue to evolve as well.   To start we need to think about the business version of the IT “server-hugging” phenomena. For the uninitiated, “Server Huggers” are those folks in an IT department who believe that the servers have to be geographically close  in order to work on them. In some cases its the right mentality, in others, where the server is located truly doesn’t matter.   It’s as much a psychological experiment as a technical one.   At a business level, there is a general reluctance to house the company jewels outside of corporate controlled space.  Sometimes this is regulated (like banks and financial institutions), most often its because those resources (proprietary applications, data sets, information stores, etc) are crucial to the success of the company, and in many instances ARE the company. Not something you necessarily want to outsource to others for control.  Therefore wholesale adoption of cloud resources is still a very very long way off.  That is not to say that this infrastructure wont get adopted into solutions that companies ultimately use to grow their own businesses.  This is going to drive tons of innovation where businesses evolve their applications , create new business models, and join together in mutually beneficial alliances that will change the shape, color, and feel of the cloud.  In fact, the cloud or “Information Utility” becomes the ultimate channel distribution mechanism.

The first grouping I can see evolving is fraternal operating groups or FOGs.  This is essentially a conglomeration of like minded or related industry players coming together to build shared electronic compute exchanges or product and service exchanges.  These applications and services will be highly customized to that particular industry. They will never be sated by the solutions that the big players will be putting into play, they are too specialized.  This infrastructure will likely not sit within individual company data centers but are likely to be located in common ground facilities or leased facilities with some structure for joint ownership.   Whether large or small, business to business, or business to consumer, I see this as an evolving sector.  There will be definitely companies looking to do this on their behalf, but given the general capital requirements to get into this type of business these FOG Agreements may be just the answer to find a great trade off between capital investment and return on the compute/service.

The next grouping builds off of the “company jewels” mindset and how it could blend with cloud infrastructure.  To continue the overly used metaphor of clouds,I will call them Managed Instances Stationed Territorially or MISTs.   There will likely be a host of companies that want to take advantage of the potential savings of cloud managed infrastructure, but want the warm and fuzzy feeling knowing its literally right in their backyard.  Imagine servers and infrastructure deployed at each customer data center, but centrally managed from cloud service providers.   Perhaps its owned by the cloud provider, perhaps the infrastructure has been purchased by the end-user company.   One can imagine the container-based server solutions being dropped into container-ready facilities or jury-rigged in the parking lot of a corporate owned or leased facility.  This gives companies the ability to structure their use of cloud technologies and map them into their own use case scenarios.  What makes the most sense for them.  The recent McKinsey paper talked about how certain elements of the cloud are more expensive than managing the resources through traditional means.  This is potentially a great hybrid scenario where companies can integrate as they need to using those services.  One could even see Misty FOGs or Foggy Mists.  I know the analogy is getting old at this point, but hopefully you can see that the future isn’t as static as some would have you believe.  This ability to channelize the technologies of the cloud will have a huge impact on business costs, operations, and technology.   It also suggests that mission critical infrastructure is not going to go away but become even more important and potentially more varied.  This is why I think that the biggest infrastructure impact will occur in the “information substation provider” level.  Data Centers aren’t going away, they might actually be growing in terms of demand, and the one thing definitely for sure is that they are evolving today and will continue to evolve as this space matures.  Does your current facility allow for this level of interconnectivity?  Do you have the ability to have a mixed solution management providers in your facility?  Lots of questions lots of opportunities to develop answers.

The last grouping is potentially an evolution of modern content delivery infrastructure or edge computing capabilities.  I will quit with the cutesy cloud names and call this generically Near Cloud Content Objects.   Given that products, services, and data will become the domain of those entities owning them, and a general reluctance to wholesale store them in someone else’s infrastructure, one could see that this proprietary content could leverage the global cloud infrastructures through regional gateways where they will be able to maintain ownership and control of their asset.  This becomes even more important when you balance into this the economic and geo-political aspects emerging in cloud compute.

In the end the cloud approach is going to significantly drive data center demand and cause it to evolve even further.  It will  not as some would like to project end the need for corporate data centers.  Then there is that not so little issue of the IT Applications and internal company services we use everyday.  This leads me into my next point . . .

The Continued and Increasing Importance of Enterprise Data Centers

This post has concentrated a lot on the future of cloud computing, so I will probably tick off a bunch of cloud-fan-folk with this next bit, but the need for the corporate data centers is not going away.  They may change in size, shape, efficiency, and the like, but there is a need to continue to maintain a home for those company jewels and to serve internal business communities.  The value of any company is the information and intellectual property developed, maintained, and driven by their employees.   Concepts like FOGs and MISTs and such still require ultimate homes or locations for that work to be terminated into or results sent to.  Additionally look at the suite of software each company may have in its facilities today supporting their business.  We are at least a decade or more away before those could be migrated to a distributed cloud based infrastructure.  Think about the migration costs of any particular application you have, then compound that with having the complexity of having your data stored in those cloud environments as well.  Are you then locked into a single cloud provider forever? It obviously requires cloud interoperability, which doesn’t exist today with exception of half-hearted non-binding efforts that don’t actually include any of the existing cloud providers.   If you believe as I do that the “cloud”  will actually be many little and large channelized solution cloudlets, you have to believe that the corporate data center is here to stay.  The mix of applications and products in your facilities may differ in the future, but you will still have them.  That’s not to say the facilities themselves will not have to evolve.  They will.  With changing requirements around energy efficiency and green reporting, along with the geo-political and other regulations coming through the pipeline, the enterprise data center will still be an area full of innovation as well. 

/Mm