When comparing Cassandra vs CouchDB, the Slant community recommends Cassandra for most people. In the question“What are the best self-hosted NoSQL distributed databases?” Cassandra is ranked 2nd while CouchDB is ranked 3rd. The most important reason people chose Cassandra is:
Cassandra is very scalable and achieves the highest throughput for the maximum number of nodes compared to other alternatives. Unfortunately this also brings rather high write and read latencies.
Specs
Ranked in these QuestionsQuestion Ranking
Pros
Pro Highly scalable
Cassandra is very scalable and achieves the highest throughput for the maximum number of nodes compared to other alternatives. Unfortunately this also brings rather high write and read latencies.
Pro Familiar to developers used to SQL
The query language that Cassandra uses (CQL) is similar to SQL even though it's a NoSQL database.
Pro Works well between physical network partitions
CouchDB works very well even when the network is physically partitioned
Pro Available
CouchDB is considered an available DMS according to the ACP theory of database management. As such it allows every client to always read and write
Pro Useful for applications where versioning is important
CouchDB is mostly used in applications where a large amount of data needs to be accumulated and where data only changes rarely.
Pro Changes API
You can use the RESTful API to listen for changes in your database, which is something most databases can't do. It makes it really easy for clients to keep their view of data up-to-date.
Pro RESTful API
Since everything is stored as a JSON document and served over HTTP, it's perfectly suited for communicating with client-side javascript, with or without middleware.
Cons
Con Not for newbies
If your dataset is in order of gigabytes then maybe consider a toy database, not a serious one like Cassandra.
Con No JOINS
Cassandra has no support for JOINS.
Con Can only achieve consistency through replication and verification
Since CouchDB is considered an AP (Available, Partition-Tolerant database management system), it is not really consistent (not all clients can have the same view of the data consistently) and the only way to achieve some "eventual consistency" is through replication and verification of data.
Con Uncertain future
After a very promising start, development began to drag after major supporters like Canonical, Selenium and CouchOne either shut down or moved to other tools. Development has begun to pick up again.
