Fauna is a serverless database where you store data as JSON. Also, you have build in a HTTP GraphQL interface with a full authentication & authorization layer. That means you can skip your Backend and call it directly from the Frontend. With the power, that you can write data transformation function within Fauna with her own language called FQL, we're getting a blazing fast application.

Also, Fauna takes care about scaling and backups (All data are sharded on three different locations on the globe). That means we can fully focus on writing business logic and don't have to worry anymore about infrastructure.

I’m newbie I was developing a pouchdb and couchdb app cause if the sync. Lots of learning very little code available. I dropped the project cause it consumed my life. Yeats later I’m back into it. I researched other db and came across rethinkdb and mongo for the subscription features. With socketio I should be able to create and similar sync feature. Attempted to use mongo. I attempted to use rethink. Rethink for the win. Super clear l. I had it running in minutes on my local machine and I believe it’s supposed to scale easy. Mongo wasn’t as easy and there free online db is so slow what’s the point. Very easy to find mongo code examples and use rethink code in its place. I wish I went this route years ago. All that corporate google Amazon crap get bent. The reason they have so much power in the world is cause you guys are giving it to them.

So, we started using foundationDB for an OLAP system although the inbuilt tools for some core things like aggregation and filtering were negligible, with the high through put of the DB, we were able to handle it on the application. The system has been running pretty well for the past 6 months, although the data load isn’t very high yet, the performance is fairly promising

Our application data all goes in SQL. We will use something like Cosmos or Couch DB if one or both of these conditions are true: * We need to ingest a large amount of bulk data from a third party, and integrating it straight into an RDBMS with referential integrity checks would create a performance hit * We need to ingest a large amount of data that does not have a clearly defined, or consistent schema. In either case, we will have a process that migrates the data from Cosmos/Couch to SQL in a way that doesn't create a noticeable performance hit and ensures that we are not introducing bad data to the system. Because of this, there is a third condition that must be met: the data that is coming in must be something that the users will not need immediately, i.e. stock ticker information, real-time telemetry from other systems for performance/safety monitoring, etc.

We implemented our first large scale EPR application from naologic.com using CouchDB .

Very fast, replication works great, doesn't consume much RAM, queries are blazing fast but we found a problem: the queries were very hard to write, it took a long time to figure out the API, we had to go and write our own @nodejs library to make it work properly.

It lost most of its support. Since then, we migrated to Couchbase and the learning curve was steep but all worth it. Memcached indexing out of the box, full text search works great.