I have some applications which require multiple queries to obtain the
all the data for any given interface. At this time I am executing all
queries and fetching the data consecutively. Which can be anywhere from
2-10 queries and resultsets back to back.

Now the queries aren't big, and at times don't return data. I have
tested the queries via some db gui's, and performance seems better even
over remote connections. Local connections (LAN) are fine. Remote ones
(WAN) tend to be pretty slow, 5-30 seconds or so for all queries and
resultsets. It's not a line speed issue, maybe latency, but db tools not
using Jaybird or java are faster.

I was curious if there were any tunable parameters I should start with
to see if I can expedite things at all. I am not dealing with any blob
data. I am not sure if the socketBufferSize would make a difference or
not? Speed is the same on Linux and Windows clients. Server is Linux,
FB SS 1.5.3, JayBird 2.1.0 beta1.

I was considering using Zebedee, but I don't really think it's a problem
compression could solve. Possibly, but seems the delay is mostly in
passing the query to the db, and waiting for a resultset. Emtpy ones
seem to take as long as ones with data. So it does not appear to be data
size or etc related.

Now I am still looking into how I am executing the queries. I am
considering using multiple db connections, and threading the app to make
simultaneous db calls. Just was hoping to avoid like 2-5 connections to
the db per remote user.

--
Sincerely,
William L. Thomson Jr.
Obsidian-Studios, Inc.
http://www.obsidian-studios.com