Re: SIMILE Store Report

Andrew Newman wrote:
> Having a quick look at the code that's using Kowari, it seems that there 
> might be some sub-optimal things being done, like creating a new 
> ItqlInterpreterBean every time which effectively sets up a new RMI session.
> 
> Also, by looking at the results it seems odd as to why "Load Page" 
> locally would be slower than over the network by two orders of magntitude.

The network interaction for all models was done in the configuration 
stage; after that, the application ran on a local, in-memory Jena model, 
so loading the page was fast.  You can probably attribute the slow local 
loading to my code; suggestions on improvements I can make are more than 
welcome.

> And the use of "In-Memory" should probably be "In JVM" or maybe "Local". 
>  By saying "In Memory" I think of an an in memory model, one that isn't 
> persistent.

Thanks, I'll make that change.

> Also, I wonder the usefulness of the "configure" test, it seems to vary 
> widely between implementations from 2ms to 50,000ms and the one that 
> does consistently better is the one that: "..naively queries a Model 
> using the RDQL query language instead of Jena's Model API. Very little 
> is cached, so all user interaction through the web application ends up 
> relying on accessing the store through non-optimal queries."

The 'configure' stage instantiates in-application caches and structures. 
  If it's blindingly fast, it isn't doing anything, which means the 
'browse' test is likely to be blindingly slow :)

-- 
Ryan Lee                 ryanlee@w3.org
W3C Research Engineer    +1.617.253.5327
http://simile.mit.edu/

Received on Thursday, 29 July 2004 12:56:20 UTC