RethinkDB and Compose: Where Next | Hacker News Reader

RethinkDB and Compose: Where Next

compose.com

131 pointsandrewbarba26 comments

Wish RethinkDB could have created a product similar to MongoDB's Atlas[1] which perhaps could have been the revenue stream they needed.

[1] https://www.mongodb.com/cloud

RethinkDB (the company) spent 7 years creating amazing value by building RethinkDB (the database) from scratch.

Compose.io probably only spent 1 or 2 months to incorporate it as part of their 'as-a-service' offerings...

Guess which company actually survives and profits from all the value which was created?!

... And this 'news release' from Compose.io is the business equivalent of a vulture feasting on the carcass of its own mother.

Errr... no it's not. It's reassuring their customers that the service they pay for will continue to operate. As one of those customers, I'm glad to hear it.

I have to assume they'll have a "Migrating To Compose MongoDB" article in a few months...

I'm not sure if they do but how much does Compose actually contribute to the project? Will they contribute in the future?

Are they simply a parasite or will they invest in their business by improving the software they sell? If they won't then I would never use them for anything as they are providing a service for a product they have no control over and no desire to invest in and improve and I'm not interested in tying my business to that.

Gross. I'm fairly sure no one at Compose would ever write a "migrate RethinkDB to MongoDB" article.

We improved most DBs we sold. In fact, I think Compose is valuable to most DB companies ... nothing like real world ops problems to learn about customers.

Well, I definitely won't be writing that article. I'm rather stubborn like that. :)

Lesson: RethinkDB should have been at the forefront of offering their DB as a service.

Lots of OSS-oriented companies do this and it works at least well enough to keep the lights on while you can work to get to higher tiers of the enterprise.

I have to agree here a bit. However, there was nothing stoping Rethink from building their own service and promoting it as the "official" RethinkDB service. They have the unique opportunity of making it better than any other service by:

-Updating the service the day new versions go GA

-Faster turn around on breaking changes

-Ability to provide authentic expert support

-Ability to provide additional functionality power users would pay for and 3rd parties like compose would not implement

They wouldn't capture the entire market but they could be the leader in it.

The open source business model is hard enough. It becomes very difficult when every other Cloud simply wants to offer your product and you capture none of that. And then the project dies and no one wins.

Open source companies need to be the leaders of their tools, provide the best services around them and differentiate their offering from the open source code base. If you can't do that you won't have a company. Asking for donations is not a business model.

RethinkDB was a good database with a bad business.

Eh.. not sure about that. Picking a less-in-the-mainstream database already seems a little bit risky for a business. To also tie yourself to a service strikes me as even more risky.

Consulting we can agree on. I can't imagine RethinkDB didn't have a consulting business.

If they were anything like the other OS infrastructure software companies I've been around then consulting at this stage was the last thing their investors wanted this business to be. Consulting is not a business model you invest in. Anymore than 15-20% of revenue from the consulting practice is a business model that does not scale or have high margins and is not worthy of much valuation.

Consulting makes sense when your software business is driving and you're looking for a bit of fat around the edges. Look at RHT's 10-Q's, for instance.

Running a database service is really, really hard. It's an obvious thing for OSS database companies to try, but the skills and experience necessary to do it are not at all the same as what it takes to build databases. MongoDB took _years_ to launch theirs, and then even more time to get it working well.

I guess the alternative is to not try, put all your resources into making Aphyr cultists happy and then going out of business a few months later.

Built on top of existing cloud architecture and with a few experienced developers a cloud service can be made available in months. mLab didn't take years to build it and get it right, for instance. And those folks were a tiny operation years ago when they started. I'm not discounting the effort but I've been around that effort and have seen DB-As-A-Service go live with a couple smart people in a matter of months. AWS API's, containers, Zookeeper and come Python and you're well on your way. Again - it's hard but so is making a DB of Rethink's quality. It can be done, however. You start small and you go from there. You exploit your advantages as you build a fantastic product over time.

I'm not saying a SAAS would have saved them as I'm pretty sure their biggest issue is adoption in the first place.

But there's something to be said about your business plan when someone hosting your software is doing well why you, the owner, has to find a new job.

> However, there was nothing stoping Rethink from building their own service and promoting it as the "official" RethinkDB service.

Nothing except the facts that writing a database engine and doing DBaaS are two entirely different things to get right, it would have required at least double the money and headcount, it would have caused even more risk on the consumer side, and ultimately diluted RethinkDB's core competency. Other than that, great idea.

sure, but in hindsight would have been a better business move

Exactaly. It would have given them a chance. Some runway. The alternative is obviously to go out of business.

In hindsight, it would have caused this news to happen in 2014, not 2016.

Which should be seen as an advantage, right? I'm guessing their employees and founders who are left with little to nothing would have been better off not squandering the last 2 years of their professional lives.

Fail fast, right?

I can only speak for myself, but as one of the founders, I don't regret a minute of the time I've spent building RethinkDB.

That's a great attitude and you all have so much to be proud of. You did something and made something pretty cool.

Good luck.

This is the fate that many OSS business models are gravitating to - developers have no expectation to pay for OSS libraries which drives the price towards free which can only be made back from large volume to sponsor indirect commercial business models. However since there's an expectation to pay for hosting which requires a direct payment connection via a credit card they're easily able to extract sustainable income without needing to pay for developing the OSS software they're hosting.

So most of the profits from OSS will be around hosting which due to economies of scale and brand confidence will flow to large cloud vendors controlled by a handful of tech conglomerates.

The commercial value derived from OSS Software will shift away from authors to big Tech Co's which is also how they're able to sponsor development of a lot of paid OSS libraries around their hosting Services themselves.

> The commercial value derived from OSS Software will shift away from authors to big Tech Co's

I'm pretty sure that the inevitability of that was predicted by lots of people as early as 1990 or before, for pretty much all the reasons you state except for the word "cloud".

Compose didn't offer only RethinkDB. It offered many other databases as well.

I'm biased (was part of Compose, departed post acquisition), but I think that's a really reductionist take.

We spent 5 years building a platform for operating databases. Putting Rethink on that took a few months, yes.

Hopefully the RethinkDB folks think it was a good move. :) They were by far our favorite DB company to work with.

I have to defend Compose.io here. They spent countless hours working alongside the RethinkDB team to debug problems, provide fixes, and improve the product.

We have always considered them key collaborators and contributors to the RethinkDB ecosystem. There is a conversation to be had about the difference in business models between RethinkDB and Compose.io, but I would never describe it as anything but friendly cooperation and mutual support.

I'm glad to hear that they expressed their dedication to preserving RethinkDB's open-source future. We're working to contact all the folks in our community to decide the next steps together. I'm excited to keep working with them.

Thank you for your kind words on our behalf. Hopefully the product we created together with the community will have a bright future.

Good to hear that they've contributed and that you're cool with it. It's inspiring to see that you're still putting your users first in spite of everything.

> ... And this 'news release' from Compose.io is the business equivalent of a vulture feasting on the carcass of its own mother.

That is brutal, for conjecture.

A fledgling open-source DB winding up on an as-a-service provider can provide mutual benefit; the database gets users running their product on a stable, consistent platform and the provider can profit from the margins on the value-added deployments they provision.

Hardly 'a vulture feasting on the carcass of its own mother'...

Reply on news.ycombinator.com