Thanks for your opinion. > I think we should do that, as BSP will change the entire architecture. My > only suggestion would be to have an architecture diagram as to how we would > approach the BSP model.
Fortunately interfaces won't change. ;) Yes, we should organize our plan on the wiki. On Wed, Oct 7, 2009 at 9:43 PM, Vaijanath N. RAo <[email protected]> wrote: > Hi Edward, > > I think we should do that, as BSP will change the entire architecture. My > only suggestion would be to have an architecture diagram as to how we would > approach the BSP model. > > --Thanks and Regards > Vaijanath N. Rao > > ----- Original Message ----- > From: "Edward J. Yoon" <[email protected]> > To: [email protected] > Sent: Wednesday, October 7, 2009 6:05:26 PM GMT +05:30 Chennai, Kolkata, > Mumbai, New Delhi > Subject: About 0.19 branch and release > > Hello all, > > IMO, Hama-0.19 (based on Hadoop/Hbase 0.19 and Map/Reduce) doesn't > have a chance to improve functionality and performance. The > pre-computation (e.g., blocking,..) is too heavy to avoid locality > problem. So, I'd like to release it as a simple version and focus on > new developments based on BSP model and Map/Reduce. > > WDYT? > > -- > Best Regards, Edward J. Yoon @ NHN, corp. > [email protected] > http://blog.udanax.org > -- Best Regards, Edward J. Yoon @ NHN, corp. [email protected] http://blog.udanax.org
