Mapreduce and BSP are two complemetary computing framework. When facing non-iterative computing, people prefer to mapreduce, when facing iterative computing, BSP has its advantage. I agree with Anastasis, we need to attract people who needs iterative computing to use Hama.
Regards, Yexi 2013/8/28 andronat_asf <[email protected]> > IMHO BSP can be advertised as a different computing model to MapReduce, > which means that applications might use BSP as a backbone and you might > create a different ecosystem around this. You might even get existing big > projects (e.g. Hive :P) start working on top of Hama. I think the first > target should be to convince (some how) the projects that already use BSP > under the hood, to use Hama. I think this is the first step to create a > different ecosystem. > > Kindly, > Anastasis > > On 28 Αυγ 2013, at 11:23 π.μ., Tommaso Teofili <[email protected]> > wrote: > > > IMHO the only thing that makes sense is the latter. > > Regards, > > Tommaso > > > > 2013/8/28 Edward J. Yoon <[email protected]> > > > >> http://people.apache.org/~edwardyoon/screenshot1.png > >> > >> During drawing the Hadoop ecosystem diagram, one question has arisen. Do > >> you think we can create the another ecosystem around us in the future? > or > >> should we try to live in harmony with already existing projects in > Hadoop > >> ecosystem? > >> > >> -- > >> Best Regards, Edward J. Yoon > >> @eddieyoon > >> > > -- ------ Yexi Jiang, ECS 251, [email protected] School of Computer and Information Science, Florida International University Homepage: http://users.cis.fiu.edu/~yjian004/
