Hi Gavin! On Mon, Jun 29, 2015 at 11:27 AM, Gavin Li <lyo.ga...@gmail.com> wrote: > Hi Roman, > > I've taken a brief look at Geode. Pistachio and Geode both can support > similar in-memory consistent HA distributed storage, and the data aware > function routing idea is also similar. Here are what I think are different: > > 1. Pistachio actually uses a very different approach to resolve the > problem. Pistachio leverages Apache Helix to do partition management and > Apache Kafka as a WAL implementation to leverage the rich features and > great performances in those systems. > 2. Pistachio supports simple plugin of different storage engines like pure > memory or different disk based local storage engines like kyoto cabinet or > Rocks DB. On our production we mainly used kyoto cabinet based SSD storage > system to optimize for read to resolve the low latency high volume read > problem in RTB applications.
This is *really* useful to know. Thank you so much for taking time and providing this compare-n-contrast. Helps me a great deal to know where Pistachio is coming from. Thanks, Roman. --------------------------------------------------------------------- To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org For additional commands, e-mail: general-h...@incubator.apache.org
