Escape the cloud database trap with serverless | Hacker News Reader

Escape the cloud database trap with serverless

fauna.com

28 pointsjchrisa14 comments

The pay-go cloud model makes total sense for orgs of all sizes. Even the term "serverless database" is a mind bender for me, but if it can scale that quickly, and reliably, and fluidly (it's a word now), it sounds exciting.

The reason the term is trending is because people are excited about not having to worry about servers any more.

What are you talking about? The database agent still runs on servers. All this does is take care of the maintenance and scaling and run multi-DC clusters, if I'm not mistaken? Just because you don't "see" the servers in the big picture anymore you can't just flat out call it serverless. Call it autoscaling or fluid or something..

We call it serverless because it fits the serverless model like a glove. You wouldn't fault Amazon for calling Lambda serverless...

Does this terminology make sense to anyone else? So far as I can see, the article suggests that one should "escape the cloud database trap" by using a "serverless cloud database". "Cloud", so far as I've ever been aware, just means "somebody else's server"; so in what way can it be meaningful to say that such an architecture could ever be "serverless"?

Author here, let me know if you have any questions about the content.

What are some lessons you learned at Couchbase that you are applying at Fauna?

Importance of a cloud product obviously is one!

Haha. Well I've wanted to focus on the kind of smooth developer experience and on-ramping that you can only get with a cloud product. That's been super fun, basically to work the experience of the product itself directly into the signup flow.

Maybe more important are industry expectations, like knowing what a multinational tech company is looking for when they do a proof-of-concept with a database startup.

The most fun lessons are the ones I'm learning at Fauna, because we are running a different hypothesis than Couchbase, and it lends itself to different goals, processes, and atmosphere.

Looking forward to seeing where you guys go with this. Good luck!

Is it possible to run FaunaDB locally (in development mode)?

Is it open source?

Yes there is a dev jar. We've removed the download for now because we are running our recent release to cloud through the QA process. Check back in a few weeks or contact us via the website and depending on what you are planning to use it for we can give you an early copy.

FaunaDB is closed source (with plenty of open source companion code on Github) https://github.com/fauna

Thanks for answering.

This means any application based on FaunaDB is locked forever with FaunaDB (because the query language is proprietary and the software is closed source)?

We see database migrations all the time in my line of work. Lock-in is rarely the main concern. Usually reliability, operability, and performance are what close deals. Even big enterprises regularly move between radically different architectures. Don't overestimate the cost of lock-in.

But yes, you'd have a hard time duplicating FaunaDB's query capability with something else.

We have bulk export tools coming soon, so you can take your data anywhere that accepts JSON.

Fauna is a major player in the market. Their knowledge base is second to none.

Reply on news.ycombinator.com