We Left the Cloud (2023) | Hacker News Reader

We Left the Cloud (2023)

basecamp.com

38 pointstalboren7 comments

I'm always curious what RoI analysis goes into this kind of decision - whether to leave on-prem and join the cloud, or vice versa. The numbers always seem huge, and in opposite directions. "Moving from onsite datacenters to AWS saved us $2m/year!"

Has something changed with AWS' pricing recently, have their business needs changed over the years, or were the calculations (to use AWS) just wrong to begin with?

AWS can get very expensive if a company scales and things have not been setup properly

The “you’re doing it wrong” argument doesn’t float for me here. I see this as kind of scammy and it’s got that vibe of producing complexity to obscure in an area that is really meant to be the simpler solution to rolling your own hardware.

The cloud has never been about absolute cheapness, but value for money and time-value in starting up.

The classic Enterprise model is to fire up a business operation using cloud providers to extend geographically, and quickly validate the business. Then you either pull the plug or start going on or near-perm for operational cost control. The finance math works out well and you free up capital for other investments.

Tech startups are in a bit of a bubble where a huge AWS integration might be a product differentiator, but also have an orthodoxy built around massive growth capabilities that investors want even if pricey (startups mean dreams of scale and cashing in).

The cloud is pricey, but has value in its services that’s hard to replicate. On prem is better but now you’re in the data centre business want to or not.

And, then, if you’re evil you’d probably look at hybrid solutions, where each part is being milked for money and redundancy, and vendors are being played off one another since you’re perpetually moving away and towards already in-prod solutions from competitors. “Are you saying we have to move that on-prem? Those dudes’ll love the bonuses I guess.” “Huh, well the OtherCloud instance price we’re getting is way better, we were gonna harmonize anyways...” “You know what, forget it, BigCloud solves that for pennies, unless…”

The cloud is neither cheap nor expensive, it’s how you play with it that costs.

(2023)

Some discussion then: https://news.ycombinator.com/item?id=38644550

These decisions always depend on the lifecycle of the product. I assume that at Basecamp’s level of maturity, where it has reached a certain saturation point and growth and usage are fairly predictable, it makes perfect sense to make a strategic decision like this and commit to a long-term bet.

Regardless, kudos to DHH and team for being so vocal about it, it's a great case study for product teams in similar lifecycle.

The cloud always seemed foolish to me. A company's computers and the data they contain are one of its most critical resources. For an IT company it's everything. To put that anywhere other than one's own secured datacenter is crazy. I'm a firm believer in "If you don't hold it, you don't own it." Also: "Two is one, and one is none."

Companies might complain that they can't find the talent to manage all this. Well then they need to hire some bright teenagers and teach them instead of expecting to get the perfect plug and play candidate. Those employees will be more loyal over the long term for a variety of reasons. But the owners don't care about providing for the future, only this quarter's sales.

Companies should invest in their own computing infrastructure even if it costs more, just so they can continue doing business as usual while the rest of the world's hair is on fire about the latest Windows rootkit or Microsoft/Github/AWS outage or whatever.

Reply on news.ycombinator.com