If you want to be strict about definitions, everywhere is a hybrid cloud. Originally, the principal proponents of cloud computing, Ray Ozzie at Microsoft and semi-rival Paul Maritz at VMWare, came out with a set of definitions that have not entirely been taken up by the industry, even today.
Maritz in particular foresaw the problems of being trapped within a single relationship by total dependence on a specific type of cloud.
Ozzie, by contrast, was too busy building one of those traps to want to argue the definition. Nonetheless, they saw clouds as being the only way to solve the growing problems that companies run into with information technology, particularly the problems of slow development cycles and hugely variable customer-facing loads.
This had very little in common with the consumer perception of cloud, which smoothly mutated from the preceding notions of large, free, public-access websites being somehow “out there”. Cloud would be better, everyone seemed to think, because it was in some way intangible, as well as fulfilling Ozzie and Maritz’s original notion of the vast mass of identical servers, ready to be used by whichever client in whatever way was required of the moment.
Clouds and virtualisation
A key component of that flexibility was that whole servers ought to be movable, which in turn required that their software was virtualised – a buzzword 10 years ago and still hot these days, though applied to the field of making nerds seasick while showing them wraparound videos, if YouTube is to be believed.
The whole process of cloud evolution is a bit like one of those passages in the Old Testament with long strings of “begats” to delineate a line of succession: virtualisation begats portability, which begats the twins of scale-up and scale-out, which begats the prospect of coupling available capacity to required demand. All those cloud positive architectures are so persuasive that it didn’t take very long for sales pitches to start talking about “your journey to the cloud” as if there was simply no other choice to be made.
This came as a shock to the people who already had carrier-grade connectivity and a large estate of servers inside their businesses, and for whom the various overblown fantasies of the spectre of variable load were just not a big deal. That’s quite a lot of businesses, no matter who you believe when it comes to digital disruption, the evolution of large corporate IT investment patterns has meant that actually, cloud propositions have to take a rather different style. And that’s hybrid; the idea that you can link your on-premise kit to the net via a privileged, secure channel and then pick and choose between local and global resources.
Payments and the hybrid cloud
Once in that position, the smart CIO starts to divide up resources and, perhaps, oddly enough, payment systems are an easy thought experiment in how such a division works in your favour. Let’s say you have a cloud-based payment management app, with an up to the minute response time, with transactions appearing the instant that a card slips into a reader anywhere in the world. That’s an operational priority (or to be less polite, it sounds really cool in board meetings). It’s not the only priority, though. Long term analysis of spending patterns by card, by department, et cetera, can be done locally, by downloading monthly statements in a traditional old-school banking style.
It can be done remotely, too, however, there are frequent motivations for being local and taking advantage of a hybrid link. I am thinking in particular of the regulations around merger and acquisitions work, where analysing a long run of expenses by linking them together according to the code-name of the target company in an aggressive takeover would not be a smart thing to do. Not because the consequent security breach is realistic, but because the regulators love to find this kind of sloppy approach.
So the argument for hybrid cloud being just a periscope into the wider internet, with all your systems inside your four walls, is considerably weakened by the need for payment systems to get managed and updated centrally, and to present carefully-vetted industrial-strength interfaces to payment processors and partners.
That’s a lot of work you don’t have to do, which leaves the work you can manage locally, the data lakes and the deep reporting, the regulatory retention and the legal disclosure, all where they can be managed strictly by the rules that keep your information safe.