Project 2012: Day 111

We’ve already spoken about “Vendor Contestability” and this is certainly the nirvana we’d like to get to in the future for computing delivery. The idea that we can seamlessly shift between providers at close to zero expense.

Kind’ve like you do with a mobile phone. Although there may be contract penalties (not with pre-paid) to pay the incumbent provider, the actual porting of the number, de-provisioning one service, and provisioning another, is close to immediate, and free.

Also, it doesn’t matter which device you own. 3G service to a Nokia feature phone works just as well with an Apple smart phone.

So there’s the holy grail. But until then, when you consider the economic benefits of the cloud, you don’t only have to consider getting on the bus, but getting off at some stage in the future. Perhaps you’ll want to get on another bus, perhaps you’ll hire a car, or buy one.

In fact rather than transport, or mobile phones, which despite allowing us to understand these concepts, do a disservice because they over-simplify the case; let’s consider property.

An enterprise will deliver property facilities in a number of ways, from furnishing a home-office, leasing perhaps offices, renting project accommodation, and owning land. As we move from owning to renting, then we need to consider moving all of the people and equipment to the new facilities. Also moving them out again should we want to rent other facilities.

There are a couple of points to note when considering getting off cloud:

  • Prior to signing contracts (or accepting EULA’s & cloud T’s&C’s) ensure there are no termination penalties.
  • When architecting your systems in the cloud (IaaS & PaaS) or determining which SaaS applications to use, work to the principle of zero dependencies. E.g. don’t choose a CRM application if this ties you in to a particular inflexible or expensive authentication system. Of course complex enterprise applications all have dependencies (authentication, security, database, BUR etc) so consider carefully the dependent systems, licencing & support agreements, as well as any future costs to transform
  • Consider capacity management trends. The cost of transforming current workloads may well be less than shifting workloads affected by future growth.

There is a balance to strike between the agility and standardisation of architectures enforced by cloud computing, and the strict governance and flexibility provided by traditional modes of delivery.

So when putting together the business case for a particular workload, remember to consider the potential or opportunity cost to de-provision this workload as well.

Thanks for reading, I'd love to hear your thoughts and have a chat...

%d bloggers like this: