|
Here we have a strong move towards what is possibly the next phase in the seemingly eternal tension between concentration/core convergence of resources versus distribution/meshing. The last swing was strongly in favor of concentration, with a new word: "datacenters" became "clouds", though the distinction seems more marketing than ontology, if we stretch it more CS (abstract) than actual specs. Now we're seeing a move back towards "on-prems", "self-hosted", "distributed"; towards the "edge" as we call it now, away from whatever passes for a "core" (the means is always somewhere in the middle, but as any pendulum, we probably cross it too fast to stop there given the momentum; such is the way of cyclic phenomenon driven and bound by natural objects such as human beings, or computers). This ping-pong goes back to the first Turing machine and what came immediately thereafter; we saw big incarnations of each extreme which became popular names, iconic hallmarks of their era — the "mainframe", the "micro computer", the "super computer", the "handheld". Cue IoT (the extreme form of distribution, as far as we can conceive it), or holistic notions of "noosphere" (information "biome" so to speak, the whole of computers seen as a ecosystem, which you may describe as a "species" like we would gases, biological beings or... technological things. It might seem crazy then again internet displays some very biological properties). The unambiguous conclusion that one should make, after observing this cycle first-hand for at least 1.5-ish period's (when you realize it's a cycle), is that we need both to make a world, and probably many variations thereof; and that we will "refresh" all of it, probably upon each half-cycle (these are, what, 30 years approx? so each 15 years give or take, like it too that from "the web" and cell phones to reach the smartphone era; or from then to now, the next few years, 2023-ish, as we cross into this next stage of the pendulum story?) Lots of things are this way in technology, cycles from one extreme to the next, ultimately impossible to fix in a "happy medium" because there's no such thing; rather there are "happy trends" which fit their time, and as many such "trends" as there are "times" — how we moved from mainframes+terminals to the almost equally capable sever-client paradigm will be a vastly different story than the move from clouds to whatever's next. I think HPE is poised to win huge rewards by grabbing this market first, and hopefully their execution is good enough. |