What Happens When Central "IT" No Longer Exists?

When we get used to something, we often think it will never change, but it does eventually; who bought a house in 2006 and assumed the value would surely keep going up?

We are working at an architectural inflection point. The signals are all around us – cloud, big data, mobility, smart computing, etc. While each of these appears to be only modestly connected, I think together they signify a major shift in how business gets done and in the architecture that supports it. If true, this means the tried-and-true Business-Data-Applications-Technology model architected and delivered by central IT will not serve us much longer.

Consider the following:

  • Big and complex are here to stay. In the past we strove for simplicity because we did not have the techniques and technology to deal with the world as it is – infinitely complex. Read Chaos: Making a New Science by James Gleick. The cloud has brought the power of distributed, elastic computing to bear on enormous problems, and this trend will continue. Will central IT continue to grow in response to the increasing size and complexity of technology problems, or will a different model arise?
  • The cloud and the App Internet are two sides of the same coin. The cloud is about optimizing the power of centralized data processing, while the App Internet is about exploiting the enormous power of mobile devices on the periphery. What happens when we figure out how these work together? Can we create a smart grid across mobile devices that also leverages cloud resources? What can we accomplish when apps no longer live in central data centers that we own and control?
  • Virtual Everything Services (VXSes?) will continue to separate complexity from functionality. Look around – everything is going virtual. Virtual servers are running virtual apps creating virtual databases that are running on virtual networks. Further abstraction of this virtualized complexity into simple services will allow value delivery while specialists continue to dig ever deeper into more-complex computing techniques. Where will these specialist work, and how will we access their services?
  • Smart integration will finally deliver self-service technology. We have been talking about Smart Computing for a while; however, I believe the next big thing is smart integration. What I mean is the ability for apps (mobile or desktop) to seamlessly discover integration points and protocols and work together. This will allow the business professionals of the future (who grew up tech savvy) to self-service in ways centralized IT simply cannot keep up with.

What does this all add up to? Even the most brick-and-mortar shops cannot survive on trucked-in products and cash in the register – business and technology are inseparable. We need a new architecture that deals with the new reality: Business is going to get more complex, and technology will be simple on the outside and unfathomably complicated under the hood. The operating model realized from the new architecture will first diminish then extinguish centralized IT.

Things don’t stay the same forever . . . home prices can go down.

Comments

Brian, couldn't agree more.

Brian, couldn't agree more. In BI specifically, decentralization (via hub and spoke) and end user self service is a huge trend that we started seeing a couple of years ago.

Great

It always makes me feel better when you agree with me Boris!

All valid points ... I just

All valid points ... I just hope IT folks don't forget the hard lessons of the past. Implement redundancy and high availability where needed. Don't just assume that because the app has been deployed in the cloud that magically one doesn't have make sure an appropriate architecture extends into the cloud to match the business needs for solution availability. The pressure is on EA folks to dig even deeper into cloud providers and service partners to make sure technology is stitched together appropriately.

I agree, the pressure is on EA

There is no magic that's going to happen. My gut feel is that the economics that forced most large businesses that are not in the IT business to have a centralized IT shop will change. Consumerization of IT, simplification through services and the spread of computing from in house to the cloud and mobile devices are likely drivers to change the model. All the needs and lessons from the past are still valid, its just who meets the needs that will change.

I believe in magic

Central IT no longer needed? Seriously? Econ 101 refresher question: What happens when IT services get cheaper and faster (less friction etc)? A) Less demand for IT services? B) More demand for IT services? Cheaper/faster will create more demand.

The notion that a busy (fill in the blank) professional wants to send precious time with IT self service is a bad assumption.

The need won't go away as you point out

Backup, DR, security, integration, testing/QA all will still be valid thigs that need to happen. My point is how that needs get's met will ultimately change. It won't be over night.

In a current state world with complex, expensive systems maintained by central IT, it hard to image a world that's different.

Brain, I agree with you

Brain, I agree with you overall. I just don't see any sign that demand for IT is waning. The faster/better/cheaper it becomes - the more demand there will be. Saturation at some point is inevitable. But for now, even with all the XaaS public cloud models, I don't see how most enterprises are going to significantly reduce central IT personnel.

Here is a great tangential story about how British Generals continued to rely on the cavalry to the end of WWI - even after it was very clear that cavalry no longer won wars. Something about horses not mixing well with barbed wire and machine guns.

http://www.npr.org/2011/05/04/135983570/wwi-a-moral-contest-between-paci...

I use this story with IT executives and ask "What is your IT cavalry?" What "horse" did you ride to power, that no longer helps achieve success?

Cheers.

Demand is not waning

Agree, its not waning now nor will it in the near future. I'm trying to read a few weak signals and project that it will. Like your analogy, BTW.

And

I think Central IT will still exist in many organizations, but it won't be the only place IT lives. We already have a federation of central shared services and business-aligned IT capabilities.

No central governance causes islands of data or systems

Central 'infrastructure' teams may be less likely to be running their own hardware in their own data centres - but at the business, data and application level, there are still significant challenges.

In a large organisation, each unit is likely to 'go its own way' and create independant domains of data and applications - potentially spread across multiple IT/cloud/service providers - which are increasingly difficult to integrate to create a composite view of the organisation or to provide a uniform 'customer experience' as customers move between units.

Architecture needs to focus on the next challenge - better application and customer experience integration. SOA and Portal techniques have their place, as a 'front end' integration technique, but for very high volumes of data, some other technique is often required. This is where integration between applications more and more falls down, and gaps appear in the information available to an organisation.

Big data or central warehouse closes the loop back to a central team, which remains challenging unless embraced by an existing central function.

Its shadow systems all over again, but for services, not data

Interesting thread. The balkanization aspect is apt, as there is now an escape vector out of central IT that will allow people, without governance from a central authority, to escape the clutches of the central IT shop. This is similar to the eruption of shadow systems for information within the enterprise. I suspect there are already many tentacles in many organizations firmly tucked into public clouds and senior management will not be aware of the scale of the movement. http://bit.ly/lnNJCy