|
Ilove the emulation, but it will never come from the hyper scalers. One issue is that local emulation runs into some big political rocks as soon as it gets good. To start with, the emulator is good enough and covers a tiny surface of what people want, eg k8s and s3. Resistance here is about customers experiencing issues caused by gaps in fidelity vs the real environment and subsequent pain for the emulator product team. ok, fine. But then you get customers who take your emulator and use it in places where AWS cant go, eg, airgapped environments. They start asking for more serious features. But wait! another team in the hyperscaler was already trying to solve this, for far more than zero dollars. Azure Stack. Azure Local. AWS Snowball. Now there are VPs shooting at you because you are, in their view, cannibalizing their revenue. You might try to avoid this war by emphasizing the dev sandbox aspect, selling to developers only and making sure that you only talk about APIs and stuff. Problem is, the API surface is 90% of why the cloud is useful (the other ten percent being the assertion that you don't have to think about it, which is an increasingly untrue proposition, as the reams of SREs will tell you). So now you have an emulator for the most valuable part of Cloud, in the hands of people who know how to use it and are strongly incentivized and capable of making it better, all running locally. It's a very small step to making that commercial and wiping huge chunks of revenue out, as your VP will tell you as they sign your pink slip. Talking to devs, the most common thing I hear re emulation is a desire to be able to let rip on any service and not fear a giant bill. Since all clouds have budget tools I wonder why this isnt possible today? Maybe theres a weakness in the planning tools rather thanthe post-use budgeting ones? |