Thanks for the clear replies, Chris. One follow-up inline with the last point:
On Wednesday, July 15, 2015 at 1:44:47 AM UTC-7, Chris Vest wrote: > > Replies inline > > -- > Chris Vest > System Engineer, Neo Technology > [ skype: mr.chrisvest, twitter: chvest ] > > > On 14 Jul 2015, at 23:18, Zongheng Yang <zongh...@gmail.com <javascript:>> > wrote: > > Hi All, > > I've been reading up on neo4j's scaling features, and would love to > confirm with some devs about the following: > > 1. Is it correct that neo4j's scale-out feature (HA) is NOT available in > the community version (I'm using 2.2)? This is an easy one from the docs, > but just wondering if there's any latest change of plans. > > > Yes, HA is one of the main selling points of the enterprise edition. > > > 2. Focusing on reads -- is it correct that HA (the scale-out features) > would only improve read throughput, but not read latency? > > > Correct in the sense that going from 1 to 3 machines could take 3 times > the read workload for the same latency. On the other hand, a 66% reduction > in workload per machine could also improve latency, depending on the > workload. > > > 3. Does neo4j support scaling up (e.g. use more threads on a single > physical machine) to either improve throughput, or improve query latency? > (I'm using 2.2 community version in Java embedded mode.) > > > Yes. Having more cores means Neo4j will allocate more threads to process > queries and transactions. It also coalesces transactions that commit > concurrently, to reduce the overhead of the individual transaction at > commit time. > Is this something that neo4j does automatically, or is the behavior -- say, maximum number of "worker" threads to use -- configurable? Also, it sounds like, again, this improves throughput but not latency (since neo4j doesn't do partitioning). > > > Thanks in advance -- any pointers are greatly appreciated. > > Zongheng > > > -- > You received this message because you are subscribed to the Google Groups > "Neo4j" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to neo4j+un...@googlegroups.com <javascript:>. > For more options, visit https://groups.google.com/d/optout. > > > -- You received this message because you are subscribed to the Google Groups "Neo4j" group. To unsubscribe from this group and stop receiving emails from it, send an email to neo4j+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
