Merriman’s vision was to build a DB that would scale on commodity hardware and in the cloud. Their customers include names like Intuit, Shutterfly and foursquare. Like other NoSQL DBs, they really excel when the data needs and read/write loads are big. For this reason, they are perfectly suited to much of the social networking and big web based apps and networks that we all use today.
So many mischievous remarks would fit in. Getting back to facts though, I haven’t heard any of the companies — at least not those that talk publicly about their large data sets (Google, Yahoo, Amazon, Facebook, Twitter, etc.) — as using MongoDB. They are running either internally developed solutions (BigTable, MegaStore, PNUTS, Dynamo), Cassandra, HBase, and MySQL.
On the other hand, Foursquare has started to use MongoDB replica sets and sharding just recently. While I’ll probably be proved wrong, the largest data set running on MongoDB is Cragslist’s online archive.
My point is that while MongoDB replica sets and sharding are theoretically solutions for scaling, there’s no proof yet — or at least none that I heard of — that MongoDB could handle data sets and workloads like Cassandra and HBase are already doing.
Original title and link: 10gen Aims to Make MongoDB King of the NoSQLs (NoSQL databases © myNoSQL)