Reading this article definitely shows we have a new Architectural strategy kid on the block. Potentially migrating to an open source technology like Openflow would appear to the previous generation of IT architects as being risky business. Look to other big players in the new breed of computing and such as Instagram’s engineering blog and you can see yet more adoption of non proprietary technologies so its quite obviously not, and the financial results are showing it isn’t.
It’s no wonder why this could be perceived as being a risky business, in contrast to the strategy adopted in previous proprietary dominated worlds of Mainframe and Server/Client computing. However with the early adopters of open, extensible technologies now being the biggest fish in the pond and not the smallest first it is becoming quite clear that this architectural strategy is going to continue to dominate and be a defacto strategy to adopt for any new breed player in the world of consumer IT today, and then to likely merge into enterprise strategy.
Comparing the older strategies to the new generation and you couldn’t be far enough apart between what technologies have been used. In the past proprietary technologies have been used to ensure that they have;
- Assurance with the external vendor support when things go well…tits up
- A throat to choke on the end of a phone line when things again go wrong with the risk being applied to the vendor
- Product integration and support within an ecosystem which is capable of partial compatibility with other toolsets and services
To add to this, the previous adoption of proprietary technology, has also made building a support capability to manage and grow that Infrastructure much easier with the educational programs allowing both old and new dogs to be taught tricks.. but again at a cost.
Granted, the difference here is that the likes of Google whom are adopting new strategies have a completely different business model and risk factor in certain areas of business operation but the majority adopting new breed strategy still serve enterprises or indirectly serve enterprise customers. So in my limited wisdom here some thoughts on this new generation of Architectural strategy and where it will go (or not);
How long will it last?
Personally I think this type of strategy is likely to never stall as the momentum grows and the new generations “growing up” in a cloud world are all oblivious to the past strategies, as lets face it the Client/Server world of computing is never going to the inheiritance that the Mainframe has been.
What dependencies are there?
A big one is the community that keeps it ticking and innovative, and I’m also sure this is never going to be an issue for as long as Proprietary technologies are around and people want freedom of openness. To add the likes of Facebook and Google are even building a community themselves that will be respected and continual, for both the organisations own business and the benefit of the open community.
Where does this leave proprietary tech vendors?
In a vulnerable position I think, for years organisations have shelled massive volumes of investment into technology solutions with very little revolutionary development in product sets. We’ve of course seen new technologies arise (at a cost) but they are still lumbered with the original technology that is depreciating meaning less competitive edge and less ability to manoeuvre.
Will it fail?
As i’ve said, we will see yet more emergence of the types of strategy google are adopting and the failure is going to be in the hands of the adopter not the hands of a vendors R&D program. Openness will provide more choice on what is and what isn’t adopted and more importantly at a cheaper cost. If this cost saving against risk is accepted by the key stakeholders then I can’t see it failing.
Do I need to change?
Well I think you and I will need to change the way we approach architecture. I am and I expect you are still caught in an enterprise world of Client/Server legacy but I see this legacy dissolving in the 3-5 year time frame with more emphasis on knowing how to utilise the inner workings of an open platform and not just know how to read a vendors Readme or PDF in order to architect the solution.