Implement cached data statistics for Cassandra

Implement cached data statistics for Cassandra

Implement a stat class for C*: https://github.com/locationtech/geomesa/blob/geomesa_2.11-1.3.1/geomesa-cassandra/geomesa-cassandra-datastore/src/main/scala/org/locationtech/geomesa/cassandra/data/CassandraDataStore.scala#L31

In general, stats need to be synchronized from various writer/reader threads. In Accumulo, we accomplish this without locking by using a Combiner on the table - this way various threads can write their local stats, and they will be merged on read (and eventually merged on disk during compaction). I don't know if anything equivalent exists for C*.

Original JIRA Issue: https://geomesa.atlassian.net/browse/GEOMESA-1887

Key: GEOMESA-1887
Type: Improvement
Priority: Major
Status: To Do
Resolution: Unresolved
Reporter: Emilio Lahr-Vivaz
Created: Tue, 20 Jun 2017 09:28:40 -0400
Updated: Tue, 20 Jun 2017 09:28:40 -0400

Start with CassandraDataStore.scala at the linked statistics definition, then compare the existing Accumulo statistics implementation and its synchronization approach. Determine how Cassandra can maintain cached statistics across concurrent readers and writers; done means Cassandra has the requested stat class with consistent updates and reads.