Opex Only IT

With cloud, SaaS catching up, IT's capex is bound to shrink. With the utility metered model, and with enterprises buying data center services rather than data center, IT expenditure will move to very high Opex and a very low with not nil Capex.

This is what would get CFOs interested into this phenomenon. It will make it easy for businesses to be more flexible and agile. It would be easy to predict expenses and pay higher when you need more (say quarter ends and holiday sale etc.) and not pay for excess capacity when you don't need it. The cloud provider will be able to even out usage spurts across customers and give that advantage.

The only enterprises which will be having high capex will be cloud providers who will finally be the only one buying infrastructure to build and maintain capacities.

IT Enterprise 2.0

I mentioned the following in my previous post :

"In its ultimate and extreme form which may take many years to materialize hardware will finally be bought and managed by large service providers. Most enterprises will then have a link to the cloud. Think of it like the electricity supply to consumers and businesses. Now we don't need to have generators to produce our electricity but rather depend on a central agency which generates electricity. Replace electricity with computing and storage and you see how these service providers will build and sell computing services. Thus the complexity is pushed to the periphery or the user is distanced from the core."

Data Centers will move to the periphery or away from the enterprises to service providers. As that happens all that goes into the data centers (computing infra, storage infra, systems administration, system management et al). So what will IT Managers do? What would be in the realm of the CIO and the IT Managers once the cloud becomes staple (and stable)?

The focus will then shift on those that are relevant to the enterprise users and not handled by the cloud. All that relates to cloud's performance will be factored in the QoS SLA of the cloud. Here are things that IT departments will then be working on:

• Focus on applications, their relevance, efficiency and architecture
• Performance of infrastructure that links to the cloud. The network elements linking to the cloud will ultimately decide the quality of service that would be realized
• Focus on end user experience of applications. This is something which seems to be long missing out its rightful share.
• Enterprise management will be reduced to managing performance of network elements which in all likelyhood will be outsourced to the network provider. Centerstage will be Business Service Management where the ultimate impact and performance at the business performance level will manage. The market for bulky enterprise management tools will suddenly shrink to the cloud providers who will be servicing hundreds and thousands of customers

The Cloud

There is probably more being written about the cloud than it is probably known about. At a high level it feeds the need for some buzz which sounds logical and true but try explaining it to a CFO and you see why it is still in the err.. air!

Different people seem to have a different view but it is surely one buzz that is difficult to ignore. How it finally plays out is difficult to predict but investments by the majors around surely show that many are betting their dollars on this. And sometimes even if a technology is not the best, if there is so much of hype and dollars it may actually evolve to be something which would worthwhile. If biggies put in their best resources and biggest dollars into this, it could surely turn out better than anything else which may be equally or more promising but does not have this sponsorship.

It means some of these to some:

• A virtualized environment run at a very large scale available on-demand
• Throwing out the traditional data center out of the window for a completely outsourced model unless you are talking "private clouds" for large enterprises
• A move towards a utility model - "pay as you use" model
• On-demand provisioning closely linked to the current immediate term dream of virtualized environments

How it will impact technology and business models is to be seen, but here are things that may happen:

• Greater focus on impact of network for computing that ever before. The quality link to the cloud would define the quality of service from the cloud.
• Emergence of totally new era of providers who are probably not even considered in this space. Google, Amazon who are current first generation implementors of this technology may lead the show directly or indirectly.
• Compute technology may move to more simpler blocks than those complicated ones being worked on today. The building blocks of the cloud would be simple commotized pieces.
• Technology to improve compute, its availability, its crunching power would move from hardware to software. The differntiators would be in the software and not the hardware as it is today.
• Storage would move to the next level as compute is commotized and all the date crunched will need to be stored in various forms.
• In its ultimate and extreme form which may take many years to materialize hardware will finally be bought and managed by large service providers. Most enterprises will then have a link to the cloud. Think of it like the electricity supply to consumers and businesses. Now we don't need to have generators to produce our electricity but rather depend on a central agency which generates electricity. Replace electricity with computing and storage and you see how these service providers will build and sell computing services. Thus the complexity is pushed to the periphery or the user is distanced from the core.
• Applications (the reason why we need data centers and computing power) will reside in the cloud and be increasingly componentized. The IP will rest with companies who could deliver required functionality by integrating various applications. Bespoke will probably continue but would also be hosted on the cloud.