Probably a good move. I'd looked at Cockroach before for a project - they basically disqualified themselves from the start by nerfing the "core" version so bad it was useless, while Enterprise was some absolutely insane figure for a cash-strapped startup. While it was possible to hotfix the code to get around their restrictions - we eventually just used something else.
This at least gets the full-fledged product in the door at startups. Say what you want about the timing or the BSL but I think this makes sense business-wise.
The enterprise per core is still an insane figure, based on last time I interacted with sales- would be amazing if this was revised, too, to be more competitive with Planetscale, etc.
Would be far easier to recommend CockroachDB if it were more competitive with Planetscale.
That is very interesting. As CRDB user, I priced Spanner (had to do some estimates during load testing), and Spanner came 3 times more expensive includign our eng salary to run CRDB
I'm long gone and never signed anything so have no problem mentioning that we were quoted USD$150/m per vCPU core with a 12-month commitment (from memory).
The project called for a minimum of 4 replicas and likely 8 vCPUs per, so right off the bat that's close to $60k/year before you even do anything. Plus the unwelcome addition of "license anxiety" to the whole development experience, and being forced to make a big decision up front.. it's not how I like to work or what a startup needs.
Oh the joys of "Contact Sales" pricing strategy, where made up rates are no more consistent than "whatever the sales rep thinks they can extract from the business".
From what I remember, the cost per server per year was about 5x to 6x (annually) the hardware cost of a new server, and these were dual 32 core EPYCs. 64 cores per box at per core licensing gets really expensive.
Re: CockroachDB vs Planetscale. It's all about the price per core of the CockroachDB license.
In my understanding, last time I talked to sales it's approximately 3x worse (because Planetscale offers 1 primary + 2 replicas) with CockroachDB you'd have to triple the CockroachDB license fees to even be competitive to achieve the same HA .... on hardware you purchase and run yourself.
Last time I checked, the cockroach serverless pricing model and free tier were cheaper than planet scale for small projects. IIRC, the dedicated cloud product was also cheaper if you kept it utilized. What’s your evidence that planetscale is cheaper?
For example, planetscale charges 3x as much per gb of storage if I read the pricing correctly.
Cockroach is also doing 3x replication of the data, so I don’t think that’s particularly relevant here. Cockroach serverless will dynamically scale up sql serving processes based on load. The storage and compute are separated in the cockroach architecture. My point is that if your query load is relatively low, cockroach serverless is definitely cheaper because the storage costs dominate. I think there’s ambiguity on which product is cheaper for a real-world application with meaningful load and data size.
I remain curious about the perception that cockroach is a meaningfully more expensive product. Where does that idea come from?
through cash strapped startups can now use the "free" enterprise version until they reach 10M$ annual revenue
weather it's a good idea to commit to it if you might not want to afford it once your revenue went up is another matter
and 10M$ annually is not little but also no absurdly huge, I mean a ~80 person company probably will struggle to be profitable with that revenue (if it's 80 good paying jobs like software developer).
For a US startup I would divide annual revenue by aprox 200k for reasonable bootstrapped employee max size. So maybe 50 max? This is assuming standard software startup with most cost being employees.
It's not that much different in the EU. Through due to higher sales/revenue tax etc. a bit less employees. Also the additional cost above neto salary for epmploying someone is higher, but AFIK (especially as a startup) you can get away with a paying a bit less. Through in general it's less viable to scam your employees by doing stuff like goading them with non voting shares and then diluting them massively before selling. Like it's still possible but with much more limits. So this is comparison is limited to ethical company operation.
> they basically disqualified themselves from the start by nerfing the "core" version so bad it was useless
Ran the core version for around 3 years in production for a smart city project. The company I worked for has been running it for around 6 years. Not sure what you are talking about. Of course, we would love to use features like stale replicas for exports. But this isn't something we absolutely need.
It was a data domiciling project so just went with sharding in good old postgres. Cockroach would have been perfect but it was going to cost something like $5k/m just to turn it on..
This at least gets the full-fledged product in the door at startups. Say what you want about the timing or the BSL but I think this makes sense business-wise.