Re: [Neo4j] REST Performance

2011-10-02 Thread jadell 
Christian,

Here are some benchmarks from a few weeks ago that I did comparing batch
operations to single operations using a PHP client over REST.  Don't know if
they help at all, but maybe get you started?

http://neo4j.org/forums/#nabble-td3282984

There is another set at the bottom of that thread after some server side
improvements were made.

Good luck!

-- Josh

--
View this message in context: 
http://neo4j-community-discussions.438527.n3.nabble.com/Neo4j-REST-Performance-tp3386667p3387804.html
Sent from the Neo4j Community Discussions mailing list archive at Nabble.com.
___
Neo4j mailing list
User@lists.neo4j.org
https://lists.neo4j.org/mailman/listinfo/user

Re: [Neo4j] REST Performance

2011-10-02 Thread Jacob Hansson 
On Sun, Oct 2, 2011 at 10:36 PM, Mattias Persson
matt...@neotechnology.comwrote:

 2011/10/2 Jacob Hansson jacob.hans...@neotechnology.com

  Hey Christian,
 
  we don't have any proper benchmarks for the server currently, but it is,
  unfortunately, still quite a bit slower than the core embedded database.
 
  There are a few things you can do to get a faster experience:
 
- Use persistent HTTP connections
- Use HTTP pipelining (this has, in my late-night-hacky tests, given
significant performance boosts, close to twice the speed)
 
  Any code/config/hints as to how to achieve this?


Well, most modern HTTP implementations will use persistent connections if
they can, but it's implementation dependent. In Java, the built in
UrlConnection (or whatever the name of the class is, I don't remember. The
thing you get when you do URL#getConnection) reuses TCP connections per
default, on the precondition that you read the full response before sending
the next request.

As for pipelining, this is (supposedly) supported by Jetty, and so should
work on the neo4j server (my experiments did not use Jetty to serve our API,
but the principle for why it is faster still applies, and the same
performance gains should be seen). So really, it's just a matter of using a
HTTP client that supports pipelining, unfortunately, there are very few to
choose from.

Here is one for Java:
http://hotpotato.biasedbit.com/




  /Jake
 
 
  On Sun, Oct 2, 2011 at 6:37 AM, Christian Straight 
  protosap...@mightyelephant.com wrote:
 
  
   I'm evaluating using REST-ful calls to neo4j stand-alone
 servermainly
  I
   would like to understand the performance issues better.  Can anyone on
  this
   list direct me to some neo4j benchmarks?
  
   many thanks,
   Christian Straight @
   WildGigs
  
   .:. Typed with thumbs and sent from my HTC Evo
  
   ___
   Neo4j mailing list
   User@lists.neo4j.org
   https://lists.neo4j.org/mailman/listinfo/user
  
 
 
 
  --
  Jacob Hansson
  Phone: +46 (0) 763503395
  Twitter: @jakewins
  ___
  Neo4j mailing list
  User@lists.neo4j.org
  https://lists.neo4j.org/mailman/listinfo/user
 



 --
 Mattias Persson, [matt...@neotechnology.com]
 Hacker, Neo Technology
 www.neotechnology.com
 ___
 Neo4j mailing list
 User@lists.neo4j.org
 https://lists.neo4j.org/mailman/listinfo/user




-- 
Jacob Hansson
Phone: +46 (0) 763503395
Twitter: @jakewins
___
Neo4j mailing list
User@lists.neo4j.org
https://lists.neo4j.org/mailman/listinfo/user

[Neo4j] REST Performance

2011-10-01 Thread Christian Straight 

I'm evaluating using REST-ful calls to neo4j stand-alone servermainly I 
would like to understand the performance issues better.  Can anyone on this 
list direct me to some neo4j benchmarks?

many thanks,
Christian Straight @
WildGigs

.:. Typed with thumbs and sent from my HTC Evo

___
Neo4j mailing list
User@lists.neo4j.org
https://lists.neo4j.org/mailman/listinfo/user