Hi, Are you talking about this result?
http://wiki.apache.org/hama/Benchmarks#Test_1_.28many_small_messages_vs._few_large_messages.29 It's hard to say that the result is an accurate analysis. According to logP model[1], "many small messages" will be more sensitive to network latency and kill performance. 1. http://en.wikipedia.org/wiki/LogP_machine On Wed, Jan 2, 2013 at 5:09 PM, wang fangfang <[email protected]> wrote: > hi, > Why HAMA is good at processing mang small messages other than small > large messages? > Thack you very much! -- Best Regards, Edward J. Yoon @eddieyoon
