Blog Post

DataStax takes Cassandra in-memory

NoSQL startup DataStax announced on Wednesday that it has added an in-memory option to its commercial version of the Cassandra key-value database. Cassandra is seeing an uptick in adoption right now because of its scalability and ability to span data centers, and the ability to serve data from memory instead of disk will make it a lot faster, too. If the approaches of startups like DataStax, MemSQL and others are any indication, it looks like databases of the future will feature broad ranges of capabilities, data formats and storage options.

One Response to “DataStax takes Cassandra in-memory”

  1. Robert Greene

    Oracle’s NoSQL Database has been doing this for years. But true, “databases of the future will feature a broad range of capabilities”. This announcement though, it is just the ability to control whether writes return after writing to server memory or disk. In Oracle NoSQL Database’s case, you can choose one level further: server memory, file system buffer cache, disk. All of the rest relies in configurable flushing intervals. I would have expected that the read side was always as described, if the data fits into memory, then it was read from memory instead of disk… I would be surprised if a read always went to disk, at least I would hope not. Anyway, was looking for some real innovation rather than incremental improvement, but hey progress is progress.