[pauldix]

This is purpose built for this use case. The MapReduce system you talk about is part of what's built in. Each aggregate function in the query language is represented as a MapReduce job that gets run on the fly.

The other part of it is that we're optimized for this use case. I've built "time series databases" on top of Cassandra before. It requires a great deal of application level code and hacking to get things like retention policies and continuous queries, which are built into InfluxDB.