Currently Being Moderated

What the cloud means to the corporate data center

 

Companies like Walmart and Seven Eleven famously use IT capabilities to create a comparative advantage.  While automating core processes with Information Technology creates an advantage, it is not the core of their business.  If there were a way to run their business without a data center, you can bet that they would abandon the data center in a heartbeat.  

So why do most companies with more than a few dozen employees carry the cost of a computer room and specialists to operate it?   In a word: Innovation.

Most companies have similar processes and there are generic IT systems sufficient to run a normal business.  Above average companies actually align the structure of their business to IT systems in innovative ways and gain an advantage on the competition: Innovation! 

It is the need to have complete freedom to innovate and control systems and the data therein that drive companies to own data centers.

Fundamentally Seven Eleven and Walmart are in the same business.  They buy merchandise, put it in stores and sell it to whoever walks in.   Yet the systems they have developed are very different.  Seven Eleven wins by ordering up to three times a day for each store.  They are very sensitive to regional events and weather.   Walmart wins by complete control of the supply chain.  They are similar businesses that have very different IT systems and impressive achievements innovating in their specialties.  

So the best reason a business should invest in building and maintaining private data centers seems to be that it creates opportunities to innovate the interface between Enterprise Architecture and IT Architecture.   We’ll discuss Enterprise Architecture and Business Process Transformation in my next blog - “How the Cloud changes the role of IT.”  

Mainframes and Servers and Clients, oh my!

Cloud computing will make the corporate data center smaller.  How and why the corporate data center will get smaller requires a little more thinking.

Let’s first review computing over the last fifty years.

Mainframes were vertically integrated, meaning all the parts came from one supplier who was responsible for the whole solution.  Mainframes were expensive and administrators scheduled time on the systems carefully.  Businesses who could not afford to own a mainframe would rent time on someone else’s.  Mainframes were so expensive that if you had one sitting idle, you would probably be willing to rent out the unused time on the system.  To make this work, these big, vertically integrated systems had facilities for charging back the departments that used them.  Remember “charge back.”

Personal computers made CPU time a commodity.   While convenient, personal computers turned every computer user into a part time administrator.  These Open Systems were often horizontally integrated, meaning you could buy parts from anyone and the consumer was responsible for integrating the system.

Personal computers led to “Client Server” computing.  These smaller systems are horizontally integrated and created a paradigm where even large systems could be horizontally integrated.  One attribute of Client Server computing is low asset utilization.  Small systems need to be sized for peak loads and applications are typically implemented in silos of IT infrastructure.

Virtualization of the computer is a means of potentially gaining benefit from those under-utilized servers. 

Which leads us to Cloud computing; you’ll notice that through virtualization, inexpensive servers can now be shared and scheduled.   Inexpensive servers now attain high asset utilization just like a mainframe. They might be better in many cases than a mainframe because of commoditization and horizontal integration.  The one thing they lack that a mainframe has is charge back.  It’s a key attribute of cloud computing. 

Charge back

Corporate IT practitioners have competition.   Our competitors are organized and they want your budget.    They think they can run a computer infrastructure more efficiently than “Corporate IT” and take the savings as profit.  These people are entrepreneurs and they intend to make money doing what you do.  They are the IaaS providers I mentioned in my first blog.

We IT practitioners need to organize our business as if we were SaaS or IaaS providers if only to demonstrate that we are cost competitive.   When we are able to do this business will be able to make excellent business decisions.   

·         Core systems where innovation provides a business advantage will stay

·         Systems that can be run more economically will stay

·         Systems that are not core to the business, where aaS is more economical, will go

Further, as Cloud practices mature, more and more peripheral parts of core systems will migrate out of the corporate data center.

A word about Data Storage: NetApp is already the storage behind many popular IaaS, PaaS and SaaS providers.   We earned this role because we are a pretty good fit for the first generation of XaaS services, but our work is not done.  We have a compelling roadmap that is designed to ensure that our products are a great fit for private and public clouds going forward.

What this means to the corporate Data Center

The corporate data center will be run like a business that competes with Cloud providers.  This means that the corporate data center must adopt many of the attributes of the cloud in order to compete.  We must be:

·         Scalable- To meet variable business demand without incurring project time.

·         On demand- Service catalog driven, and competitively nimble.

·         Pay as you go- Charge back, we have to describe where the money goes.

·         Multi Tennant- If the data center is to compete, it must be efficient

Since the corporate data center will take on these attributes, we have real hope that as the cloud market matures standards will evolve.  More than OVF and the Amazon API, practical standards will emerge that allow businesses to run 80% of the time on inexpensive infrastructure in the corporate data center, but take peak loads to the cloud the other 20% of the time.  

The corporate data center will be lean and capable of “Cloud Bursting”.

 

Next section:  How the Cloud changes the role of IT.

 

Bonus:  Ever want a score card that keeps track of progress against a roll out plan?  How about one that keeps track of progress for a lot of roll out plans?   Try this spreadsheet out and dazzle your boss J

Attachments:
238 Views Tags: saas, data_center, paas, it, gary_garcia, iaas, chargeback, cloud, netapp, virtualization, netapp, virtualization, saas, paas, iaas, it, chargeback, data_center, gary_garcia, cloud


Jul 17, 2009 5:44 AM raymclean raymclean    says:

There are two perspectives woven into this article which I found a little confusing.

 

On the one hand there is the IT practitioner perspective highlighted in the statement:

Corporate IT practitioners have competition.   Our competitors are organized and they want your budget.

 

Perhaps this is a wake up call to IT practioners but I have a problem with it.  The budget is ultimately the business's and not IT's, and the IT practioner should be aiming to do what is right for the business and not the IT bit of the business.  If letting someone else run part of our infrastructure, say with SaaS or IaaS, is a better choice for the business then we should be ahead of the game and evangelising this rather than adopting a defensive posture.


On the other hand the more general perspective of the article is from the outside observing an industry trend.  With the general tenet being that the flow towards more IaaS and SaaS is inevitible with a destination of an 80%/20% split of cloud/local being likely, and we need to go with it.

 

These two different perspectives in the article may really reflect a divide between the "insiders", entrenched practioners getting on with their locally-hosted infrastructure in a rough business cliamte, and seeing the cloud as hype (of which there is more than a little).  And "outsiders", industry watchers, early adopters and others who can see a major change happening, but it has lots of parties jockeying for position and muddying the waters.

Jul 17, 2009 11:22 AM garcia garcia    says in response to raymclean:

Thanks Ray- Your observations are right on and give me an opportunity to clarify the point.

 

The budget is absolutely the domain of the business.  Which is why IT practioners need to evolve our thinking about how we deliver value to the business.

 

The old way was pretty much project based thinking. The business would ask for a new feature or service and we would evaluate options and do an ROI about how we would satisfy the request.  If we could bring the service in at a price that made sense, the project moved forward.  Otherwise we looked at scope options and lastly we might look to alternative delivery.

 

There are two problems with this approach.

 

First, the cost of simply having an IT department is not factored into the ROI.  Simply having a data center exposes the project to a number of costs that IT does not factor into the traditional project ROI.  For example, Gartner says that 55% of all IT departments have neverf seen the power bill for thier data center.  And power is 33% of the cost of running a data center.  Seems like a big miss.

 

Second, most IT deparments do not have a charge back discipline.  If they are unable to account for utilization, a feature of chargeback, then they are unable to measure the cost of change for delivering any one service.

 

Those two problems are solved by Cloud providers.  Because aaS providers must deliver a simple Service Catalog with predictable prices, the business can easily understand the total cost of TRYING a project in the cloud.

 

Yes, the Cloud will win part of the business IT has traditionally delivered, the question is what part of the pie should they win?

 

I believe that Internal IT departments can be very efficient.  But we need to adopt some of the business practices that the cloud providers have adopted.  Co-Tenancy, the ability to charge back and a simple service catalog.

 

I agree with your statement "If Letting someone else run part of our infrasturuct ... is a better choice for the business then we should be ahead of the game"  .  

 

Hopefully, this dialog can help us to understand how we can help the business make an informed decision.   Which means we IT practitioners need to adopt architectures that compete in an emerging marketplace and we need to know how to be proce competative!

 

Thanks for keeping this interesting!