"Kenneth D. Litwak" wrote:
>
> Since tunneling was raised, does nayone know of any statistics for the
> performance of tunnelingIIOP or RMI vs.directly calling thorugh a specific port?
> Thanks.
I just made some measurements. In each case, I ran my test case four times and
got an average result. (I'm not trying to publish a research paper here, so no
apologies for the poor standard of measurements).
On a machine with Java/EJB client, Sybase EAServer (locally, no network),
simple ping requests, HTTP tunneled IIOP is 25 times slower than native IIOP
in my environment (800 requests per second untunneled, 32 requests per second
tunneled).
Over the Internet, (New Zealand to California with 128Kbps cable modem Internet
connection) HTTP tunneled IIOP is 25% slower than native IIOP in my environment.
In this case, due to network latency, everything is much slower (1.70 requests
per second untunneled, 1.36 requests per second tunneled).
The IDL interface that I was using for the test is:
interface Manager
{
void ping();
};
A 25% performance degradation using HTTP-tunneled IIOP over the Internet is not
bad at all. Large data sets will introduce the speed of your Internet connection
as the limiting factor (e.g. methods returning 80Kb result sets).
<vendor note="Please ignore if you don't like product pitches">
This is all using Sybase EAServer 3.5, due for release next month with full
EJB 1.0+ support (we don't claim to support 1.1, since it is not final),
including entity beans with bean managed persistence, stateful failover and
2PC support.
</vendor>
________________________________________________________________________________
Evan Ireland Sybase EA Server Engineering [EMAIL PROTECTED]
Wellington - New Zealand +64 4 934-5856
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
