Wednesday, Nov 23 2011

Future IT: Its All About The Applications

Written by
A post by James Urquhart titled “What cloud boils down to for the enterprise” prompted some passionate exchanges of opinions. The most interesting between James, Andi Mann (CA) and JP Morgenthal and then I put in AUS$0.02 worth.

The crux of the discussion was around the statement “cloud is an applications centric operations model”. The discussion focused on two different issues;

“Cloud is actually about simplification of the infrastructure lifecycle” (JP Morgenthal)

So I think JP is right in the context of infrastructure cloud today. For most enterprise clouds, elasticity, ubiquity have less focus and self-provisioning, instrumentation, resource sharing are more important. A high degree of operations automation is being developed and deployed under the umbrella of cloud. You can look at all the major infrastructure SW vendors and find a cloud product that is the combination of a CMDB, a workflow tool and a service catalog (see cloud washing).

Conclusion: JP you are right, but that more a natural evolution of enterprise infrastructure and less a paradigm shift on any operations model. Therefore automation and simplicity are not exclusive to the cloud model.

“focus on the app helps achieve these, but I think that misses the bigger picture” (Andi Mann)

I have posted a few times about the shift in focus from Infrastructure to Service (IaaS) to Platform as a Service (PaaS) as the real cloud. In my opinion, for the mature business processes (eg. CRM, Finance, HR) will be acquired from a SaaS vendor, or a COTS vendor. The rest (eg. custom apps, integration, customizations) will be developed in a public or private Platform as a Service solution. This abstraction away from infrastructure creates the paradigm a shift in the traditional operations and services model for IT.

So is the future model one about applications, or is it about business services?

I first wrote a story to explain how I arrived at my conclusion, but decided that no one has the time or motivation to read lengthy works of non-fiction. If you happen to be around for a beer sometime, then I can be easily encouraged to expand on the topics. Instead, here is the abbreviated text;

- IBM had the perfect operations model with systems analysts and systems programmers, just too expensive and didn’t scale (the service model, not the mainframe) to meet the explosion of business use cases

- The PC and distributed computing put the centralized IT organization in its infancy on the back foot and forced a strategy of 80/20 (deal with the 20% of technology used by the 80% of users) or lowest common denominator service model. COTS applications lower the cost of application deployment.

- Open systems, The Internet and open source lowered the barrier to entry for application development so business unit rolled their own custom apps

itil fw- As the void between central IT and business users grew, accounting firms who rebadged themselves as “IT consultants”, attempted to help central IT take back influence through fluffy concepts (read frameworks, and methodologies) of Compliance, Service Management, Business Process Management and Business Services. The air was full of words like Business Alignment and Customer Centric. Unfortunately it is rich in logic and totally devoid of execution. The frameworks almost always ignore the application development/deployment lifecycle.

We have reached the next step.

Today’s Landscape

It is true that a huge amount of cost and resources are going to the support and maintenance of infrastructure and desktop computing. It is a specialization in itself, and in many cases is insulated from the rest of IT. Many of the customer centric methods and frameworks (e.g. SLA, ITSM) grew up and out of this function. This should not be ignored, but be sure that this is in the most part not value creating. It is not the leading edge of the operations model. It is an exercise in cost management and quality refinement.

The majority of value creation comes from the development/deployment of applications that enhance business processes or allow completely new business process to be executed. The result of business users maturity is they are very specific about what they want. The are looking to enterprise IT organizations, consultants and vendors to provide people who have very specific knowledge on how to implement applications. They want improved time to market and elimination of all costs (time/money) that create barriers to building this functionality. The Service Management world is grabbing onto Agile, Scrum and other methods to extend their reach into this demand.

Business cannot cop out!

When people start talking about “IT alignment with business” I start getting a little uneasy. If you are talking about IT getting better at understanding business process, then all power to you. If the conversation extends to IT treating business units like a customer, then the conversation is starting to go downhill. The reason ? It is too simplistic a interface. IT is too complex and when forced to fit within a buyer/seller context. For the same reason, I am not a big fan of internal chargeback mechanisms, you end up with a one dimensional relationship. If you can reduce the relationship between IT and business users to this extent, then you should outsource the whole box and dice. This is why the application centric model is ideal and not the business service model.

Business users are ultimately responsible for success of their business process. Part of that success is understanding all aspects of execution that make it a success. Technology is a big part of that, and understanding and empathy with the strengths and limitations of IT is crucial to business success. Equally IT needs to come 50% of the way to understanding the business process to know which technologies and tool are best suited to deliver a solution. Innovation must come from both sides of the fence. If your definition of Business Services as a operations model is about abstracting business users from IT this you are on a slippery slope.

Conclusion: Enterprise IT has a huge opportunity to take back the create of value through an applications centric model. Whatever label you choose (applications centric, service management or business services), reaching too far into the business process creates a far greater risk of under delivery and create a slave master relationship. The delivery of a applications (SaaS) and Platform as a Service (PaaS) is the model of the future. Service Management, Project Management, Applications Management disciplines have the role of supporting the applications centric model to ensure these interactions are professional, efficient and of a high quality.

Contributed by: Brad Vaughan