Today, I accidently found an interesting stuff, which may help us operating large scales of data sets for redpoll. This is a matrix computational library based on hadoop hbase.http://code.google.com/p/hama/