I read this and think "Former OpenStack supporter with skin in the game over SDN and Kubernetes ignores Jevons' paradox, constraints on underlying components, serverless and speed effects to make a case for cloud repatriation at scale" - not buying it - a16z.com/2021/05/27/cos…
On Jevons' paradox alone, as I've said for a decade ... AWS will be constrained by the speed it can build data centres i.e. it can't drop prices too quickly because that'll force up demand beyond their ability to supply. Eventually they will overcome. Expect prices to tumble.
Getting a handle on expense however is a great point made in the article which is why serverless, billing per function, capital flow inside applications will become and are becoming such critical practices of the future.
So, overall, it's a good article but I take the repatriation argument with a big pinch of salt. Instead expect margins to fall and industry to see a shift into serverless with those new practices.
The problem is, regardless of any merits, those who want to believe the future is more hybrid than not will jump on it (already happening) as "proof" of their views and justification for spending vast sums on home grown efforts.
There are very niche cases where the arguments for home grown hold true but ... they are niche.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
