I read an article the other day about how a term that Cisco coined is gaining more steam in the tech community: fog computing. Unlike cloud computing, which describes all kinds of resources, fog computing is specifically around storage and networking. The article (which can be found at http://www.foxbusiness.com/features/2016/04/26/forget-cloud-fog-is-next-big-thing.html) implies that the time of the cloud has already passed and we are once again moving back to decentralized computing.
I don’t really see it that way.
It’s true that the hybrid model (sometimes lumped together with bimodal IT) is the way that a lot of companies are moving forward, but this is not an end to cloud computing, it simply signifies the beginning of real adoption. True paradigm shifts are hard to come by and in most cases are very difficult to adopt. The average company does not have the agility or budget to completely abandon years of investments in a move to the new shiny, nor should they have to. The momentum of cloud adoption we are seeing now is really being driven by two things. First, companies are ready for a tech refresh and the agility that cloud computing offers is very appealing. Second, the services being offered by hyperscale cloud companies are evolving to be more than just extensions of on-premise compute. Services like big data and IoT are much better suited to large installations.
It is important to understand during this move that IT is very rarely black and white (or green) anymore, so the idea of “all-in” on a given technology is nice to strive for, but the reality is that it may simply not work for the business. And that is a point I see a lot of companies and analysts missing. IT is expected to drive business innovation – we are actually the best suited to do so – however IT innovation in isolation leads to things like the .com bust. Unless your business is IT innovation anything that IT attempts must be done in the context of the business. So really understanding what new technology can and cannot do as well as having a deep knowledge of business processes and goals are the two foundations to true business innovation.
One critical component to business innovation success that is often overlooked is the human element. Regardless of what types of systems we use or where they live, humans will always need a way to interact with them. So if we manage to move all of our workloads to a public cloud, saving a bunch of money in the process, but our sales team collectively spends an additional 50 hours a week performing the same tasks because of latency and poor offline availability, have we really been successful?
Cloud adoption is a perfect example of this. The Cloud is often mistakenly seen as a place, as in we move our workloads “out there”. In reality cloud computing is a paradigm – it is a way to offer and provision resources that extends the concept of virtualization to a new level of availability and scalability. It is true that public clouds are “out there”, however private clouds have been around just as long as the public cloud – even longer if you count colocation and hosted services in the private cloud camp. This identity crisis has led to a number of missed or failed opportunities because the workloads were not evaluated ahead of time to understand where they were best suited. Moving them to the cloud model is a valid solution, the question really needs to be which cloud.
That understanding is driving what we are seeing now with the push to hybrid and the coining of terms like fog computing. It is the evolution of the cloud strategy. Most public cloud vendors see their offerings as able to handle any workload. That may be true from a purely IT perspective, but for a business that requires data sovereignty or offline survivability, that is not the case. Microsoft seems to be the only major public cloud company that realizes this and is developing a solution with Azure Stack. Simply put, Azure Stack is a way to extend the cloud management model of Azure into a private cloud, and it is gaining a lot of attention as companies realize they have tools and processes that are just not suited for the public cloud.
During this discovery process, I expect that we are going to see a number of new edge computing offerings to help companies manage the workloads that consume commodity resources in the public cloud while keeping specialized tasks local or hosted in private clouds.
While it would be nice if we could keep away from additional weather related terms, I would hate to have to explain to a customer that they need rainy applications to take advantage of their foggy cloud, I am preparing myself for the worst.