Why should an outsourcer be entrusted with what already seems to be natively managed? This is the question that organizations are asking themselves when they are about to switch large parts of their information systems to the cloud. A legitimate question: whether it’s Microsoft Azure, Amazon Web Services, Google Cloud Platform or Alibaba Cloud, all these infrastructures are managed by their eponymous suppliers. In these conditions, what added value brings an outsourcer? Explanations.
The cloud offers can be associated with SLA (Service Level Agreement) or to a best effort model concerning offers made available free of charge. For Azure, Microsoft details, product by product (storage, container, databases), all service commitments that can flirt with the 99.99% monthly uptime. Reassuring, at least at first sight. Except that these figures only describe the availability of resources taken one by one and not the final performance (seen from a user’s point of view) of an application that will use many of these services.
The outsourcer is primarily a cloud architect
Is adding services with a 99.99% promise enough to guarantee the final performance? Everyone knows that an information system is complex enough to refrain from answering in the affirmative. And even more in a cloud environment. The coherence of the architecture – the application breakdown, the retained cloud services, their interaction modalities, their redundancy – take the lead on the final performance. Using the cloud does not preclude designing an upstream architecture based on business expectations.
On the contrary, the abundance of services eligible for answering need makes it complicated. Let’s illustrate: on AWS, opting for a adapted storage, its needs supposes to choose among nine potential products. And to answer, it will be necessary to look at the nature of the data, the uses, the needs of safeguarding of the managed environments. All this without forgetting that, in the cloud, the sizing is different from those of the physical world already known.
To decide, an intimate knowledge of these services, but also the limits of the cloud infrastructure (because the backbones of cloud infrastructures also have their own limits) is needed. And that’s where the expertise of the outsourcer comes in. With certified experts for the different clouds – which involves significant ongoing investment and skill management – the outsourcer can recommend a combination of cloud services and an architecture to deliver an expected performance. The outsourcer therefore does not commit to the availability of infrastructure resources – the scope of the cloud provider – but to the availability and ultimate performance of an application. What is managed from design.
The outsourcer is organized to manage the pace of cloud updates
The design issues will be even more numerous in the case of migration from an existing application to the cloud. Which strategy to choose? A simple rehosting (also referred to as “Lift & Shift”)? A re-platforming to take advantage of some features of the cloud? Or a refactoring, or a much deeper redesign to couple the application with cloud services?
Not surprisingly, there are no good answers by default. It all depends on the nature of the existing, business issues and, of course, what the cloud infrastructure has to offer. One certainty however: to use an outsourcer will be more relevant if he has an integrated consulting service. Objective: to avoid a discontinuity of information between architects and operators. The type of discontinuity that leads (too often) to re-auditing migrated applications in the cloud in order to properly specify the operating documents …
The exploitation precisely. Here too, a question emerges: what is the contribution of the outsourcer at the time when we speak not only about DevOps, but also of NoOps? In other words, a sufficiently automated environment to do without operators. In practice, this vision of automation in the cloud faces a serious limitation: the fast pace of cloud infrastructure updates that interweave new functions and deprecate existing functions. On the Azure alone, there are 600 new features that appeared in less than a year.
The outsourcer is equipped to keep budget control of the cloud
Good news, on this playground, the interests of the customers join those of the outsourcer. For the latter, each incident results in a loss of profitability. That’s why resources are invested to support continuous improvement in operations, with measurable gains and savings. This long-term work, which every client in his corner could hardly finance, directly benefits everyone.
The control of the cloud budget is also another subject on which the contribution of the outsourcer is decisive. If the cloud carries a promise of consumer billing, in practice it is difficult to modelize (to anticipate) the cloud budget. The different calculators made available by the actors are of a relative help. Evidenced by the proliferation of customers who seek to “uncloud” applications as a result of overflows budget …
Here too, the experience of the outsourcer, as well as the tools he developed, will be invaluable in order to better assess the cloud budget and master the services at risk. This competence will be even more essential in an international context where it is to play with different regions of a cloud infrastructure, or even with different clouds.