Whatever happened, something was way off the way they used HBase. Adding machines adds latency whaaaaat?
Also this is not the YCSB like everyone wrote, the post was updated with this : Update: A few users that had more luck reading the details on the slides have pointed out that this is not the YCBS benchmark, but rather a new one developed by the presenter. Another detail that’s important is that data used was rather small and could easily fit in memory. It seems they released the code 2 days ago: http://www.nosqlbenchmarking.com/ so I guess someone will have to look at how they used HBase. Ryan found that the YCSB was adding a lot of latency just the way they used the Result API (it's now fixed), maybe we'll find the same here. Finally, in research the most important step is validation of the results. They wrote that in the HBase case they "checked the logs" and "ran the balancer". Basically their numbers were not validated, the most basic thing to do would have been to contact the devs of each nosql database like the YCSB guys did. J-D On Fri, Feb 11, 2011 at 7:01 AM, David Engfer <david.eng...@gmail.com> wrote: > http://nosql.mypopescu.com/post/3163240962/ycbs-benchmark-results-for-cassandra-hbase-mongodb > > Not sure what the issue was there with HBase, but would this be much > improved if they tried it again with 0.90? > > Thanks, > _______________________________ > David Engfer > [817.360.4923] > david.engfer@gmail > http://www.engfers.com >
