When I’ve just been hired in the company, 10 years ago, the infrastructure just started to be outsourced. This was a big change and complex challenge, as the company lost servers, data centers and associated skills. It has to be transformed. The governance has changed. With some distance, now, this change was not so huge as the company kept control of its infrastructure. For sure the way decisions are taken was different but at the end the company still in position to take decisions, orient strategy and modify the rules, based on its own choice.
What’s coming nowdays seems something different and more complex to manage, in my point if view cloud computing is a strong challenge on infrastructure control for companies. There are not only the SaaS service coming in the company on which we have no control on how the operations and the hosting is done by the service provider. And for which, due to maturity level of the offer, the quality is generally under the industrial standards.
There is also the “cloud” spirit coming, I mean, in all the mind we can go faster, go instantaneously, with no cost and so on… A a consequence, part of the company, not usually in charge of the IT operations take its own external cloud environment to do faster, cheaper, it’s so easy ! And it’s true, they are doing faster, they are doing cheaper… but they are not doing the same, they are doing lower service, lower backup, lower security … and this is working because the perimeter is limited to a couple of applications they are managing, not thousands . But as a consequence a new piece of the infrastructure is going out of control of the infra teams.
In fact, compare to what can be provided today, the old infrastructure standards are too complex, too long. In the mind of managers and project managers we must be agile by looking to external providers, cloud, IaaS, PaaS, SaaS providers to get the solution. This have for consequence that we should look outside to manage POC and later, why not development. In a way, it’s not false, we have to be more agile particularly in these environments, but the truth is that we keep loosing control on infrastructure and multiply the number of service provider and the diversity of the solutions.
For some other part of business, we directly decide to split it out of the standard IT department to let them progress more efficiently, alone, far away of the industrial standard. Once again, it’s true : it’s more efficient in a short term point of view : less quality, agile / recent technologies. But as a consequence, this split part loose the industrial skills we can have and the mother company loose the opportunity of learning about efficient IT. In a longer terms, it also means that in the future, the main IT department will have to manage some non standard technologies from the later merge of what has been split in the past.
By looking on these facts, growing, month after month I really think Infrastructure teams are actually loosing control. We are already consider as slow, expensive, because we are managing complexity, technology diversity, because we are driven by contract with cost constraints, because we are constantly challenged on SLA and because we are inheriting a weight silo organization. The truth is, all of this is due to the age of this kind of organization, we can also name it maturity. I’m pretty sure all the cloud like solution will become more expensive, less agile after each progress (security, availability…). Don’t think I’m against the cloud solution, I’m fully convinced of the advantage of these technologies, I just consider them with reserve due to the actual maturity and think that all these technologies can take place in the current environment and help them to change the current silo organization to a more agile one. I’m just afraid by the future impact of this loss of control and the future multiplicity of actors we will have to manage. If in parallel no investment are made on the current infrastructure because they are made on outsourcing to the external cloud, I’m pretty sure we are going to hit the wall. This means being in front of availability problem we can’t manage, data loss we can’t recover, service provider closing we can’t replace… when in parallel we will be challenge more and more on managing an internal infra environment where only old, fat, really complex applications will be kept.
I’m also convinced that this situation is due to the new arrival of these technologies, coming fast and growing faster than our infra organizations can adopt them and we will soon react. I hope.