Yes, just putting 1500 docs 3 times in a row into Xindice and meassure the
docs/sec.
The throughput differs only in less than 1% while storing messages into
Xindice. I'm using a testclass which connects indirect to the db via the
XindiceProxy of xmlBlaster.
I just added a test retrieving docs from Xindice by id and this test
didn't show much difference either, about 1% deviation with and without
Piccolo.
Retrieving 1500 docs by id gives me a result of more than 700 msg/sec
whereas a msg looks like:
<?xml version="1.0"?>
<message> <description>Message Text</description></message>