Writes already update rows that are hot in the cache when they are sent. So in the best case this would be no better, and in the average case (where you push out rows in the cache to make room for cold ones from the memtable) a lot worse.
On Fri, Sep 16, 2011 at 8:40 PM, Yang <teddyyyy...@gmail.com> wrote: > this way we wouldn't need to get a sudden rise in read latency after flush. > or is something similar is already i there? > -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of DataStax, the source for professional Cassandra support http://www.datastax.com
