Can CounchDB keep up with massive writes and reads?

Question

I was wondering if CounchDB is able to handle a massive writes and reads to the database. The size of each write and read is small, but the number of writes and reads within a second is a lot. Since the data is changing constantly, can it keep up with the view?

Thank you

Answer 1

I guess that is what it is designed for, and that is their pitch - massive read-write scalability, being a object oriented database.

Main advantage is scale out (to multiple machines) and replication between multiple copies of the same data (read-write). You can check out the documentation at

http://couchdb.apache.org/docs/intro.html

Answer 2

Big Picture

Without numbers I can't guess at what you mean by "massive", so I guess the simple answer is "yes it can". I wouldn't worry about reads and writes keeping up with view generation - your definition of "massive" would have to be really, well, massive. At that point you'd scale horizontally or vertically, which CouchDB is very good at. Plus, hopefully massive load means massive amounts of money for you, so hardware should be much less of a problem.

Less Big Picture

Also, writes don't block reads in CouchDB. And if you don't want the view to update when you read from it (ie., if there have been updates to documents in the view since you last read from it), then specify stale=ok as a query option. If you have a more specific example of what you're worried about, then we could probably try to tailor a more specific answer for you.

A Totally Different Picture

Lastly, take a look at http://books.couchdb.org/relax/reference/high-performance for a more detailed discussion on benchmarking from CouchDB's developers.

Cheers.