I'm looking to setup a Hadoop cluster to take advantage of its parallel architecture. I'd like to use R, SAS, or equivalent to run datamining or analytics models. Since these analytics will by and large be serial, will there be no benefit to using Hadoop over a single server? That is, does the software have to be coded for parallel distribution in order for it to gain any of the distinctive Hadoop advantages?
Thanks for any advice, kbrownk -- View this message in context: http://old.nabble.com/Can-serially-code-software-utilize-Hadoop-benefits-tp34192373p34192373.html Sent from the HBase User mailing list archive at Nabble.com.
