Use of public clouds to run Big Compute has turned the corner. There is no market survey or analytics behind the statement; the evidence is purely anecdotal. I call it the “I Got It Transformation.” It is that moment when the majority of customer conversations evolve from “why would someone want to use public cloud for Big Compute” to “how do they do it?” The overwhelming number of conversations we have with potential customers now start with “how do I start?”; as few as six months ago we were spending a large amount of time walking through justifications.
One wonderful thing about cloud is dramatically eval and spin-up time. In a cloud environment, the length of time between deciding to start a pilot and getting meaningful results is typically weeks. In a fixed infrastructure model the period is typically months. In fact, we often run customer pilots in less time than it takes to procure, deploy, and image test infrastructure.
Differences like these do not make cloud better than internal infrastructure, but does highlight one of the key benefits to utilizing a mix of fixed and flexible resources. And when that happens for the teams running those pilots, it generates one of those “I got it” moments.