In my last post I discussed how the cloud is enabling consumerisation for enterprise applications, in this post we look at how this plays a role in this and what does this mean for IT moving forward.
On the supply side of IT, it means applications are best delivered as easily consumerable services to employees and others (partners, customers, suppliers). One large European multinational is already delivering all their new applications as internet (so not intranet) applications. Meaning any application can be accessed from anywhere by simply entering a URL and doing proper authentication. Choosing which applications to provide internally is based on whether there are outside parties willing and capable to provide these services or whether the company can gain a distinct advantage by providing the service themselves.
When speaking about consuming services, one should try and think broader than just IT services. The head of distribution may be looking for a parcel tracking system, but when asking the CEO or the COO they are more likely to think of a service in terms of something a DHL or TNT delivers. Services such as warehousing, distribution, but also complaint tracking, returns and repairs, or even accounting, marketing and reselling, all including the associate IT parts of those services. It is the idea of everything as a services, but on steroids (XaaSoS). Please note that even when an organization decides to provide one of these services internally, they can still source the underlying infrastructure and even applications "as a service" externally (this last scenario strangely enough is what many an IT person seems to think of exclusively when discussing cloud computing).
On the demand side of IT the issue is an altogether other one. How do we warrant continuity, efficiency and compliance, in such a consumption oriented IT World. If it is every man (or department) for themself, how do we prevent suboptimisation, In fact , how do we even know what is going on in the first place. How do we know what services are being consumed. This is the new challenge, and it is very similar to what companies faced when they decided to not manufacture everything themselves anymore, abandoning vertical integration where it made sense and taking a "supply chain" approach. Cloud computing is in many aspects a similar movement, and also here a supply chain approach looks like the way to go.
Such a supply chain approach means thoroughly understanding both demand and supply, matching the two and making sure that the goods - or in this case services - reach the right audience at the right time (on demand). IT has invested a fair amount of time and effort in better ways and methodologies to understand demand. On the supply side, IT till now assumed they were the supplier. In that role they used industry analysts to classify the components required, such as hardware and software. In this new world they need to start thoroughly understanding the full services that are available on the market. An interesting effort worth mentioning here is the SMI (Service Measurement Index) an approach to classify cloud services co-initiated by my employer, CA technologies and lead by Carnegie Mellon University.
After having gained an understanding of both demand and supply, the remaining task is "connecting the dots". This sounds trivial but is an activity that analysts estimate becoming a multi-billion industry within just a few years. It includes non-trivial tasks like identifying which users are allowed to do which tasks in this now open environment and optimizing the processes by picking resources that have the lowest utilization and thus cost. Because going forward scarcity will determine price especially in the new cloud world (which resembles Adam Smith's idea of a perfect open market a lot closer than any internal IT department ever did or will do).
Now off course all of the above won't happen overnight. Many a reader (and with a little luck the author) will have retired by the time today's vertically integrated systems - many of which are several decades old and based on solid, reliable mainframes - will have become services that are brokered in an open cloud market. A couple of high profile outages may even prolong this a generation or two more. But long term I see no other way. Other markets (electricity, electronics, publishing and even healthcare) have taken or are taking the same path. It is the era of consumption.
PS Short term, however, the thing we (IT) probably need most is a new diagraming technique. Why? From the above it will be clear that - in such a consumerised world - architecture diagrams are a thing of the past. And an IT person without a diagram is like a fish without water . We need something that allows us to evolve our IT fins into feet and our IT chews into lungs, so we can transition from water to land and not become extinct in the process. One essential aspect will be that unlike pictures of clouds and very much like real clouds, the diagrams will need to be able to change dynamically, much like pictures in a Harry Potter movie (it's magic). Who has a suggestion for such a technique ?