Sent From: [log in to unmask]

At 12:12 PM 3/8/2001 -0600, you wrote: >Does anyone know anything about how well existing RTI implementations handle >non-trivial numbers of geographically distributed federates over the >Internet? During HLA compliance testing, I had trouble with federates >crashing while trying to join the federation if I tried to get more than >three federates on my end to join with the test tool at once. That leads me >to suspect that RTI 1.3NG can be very vulnerable to inadequate network >performance during the join process. I`m curious as to whether others have >encountered similar problems trying to use 1.3NG over the Internet, and also >whether other RTI implementations are designed to be more tolerant of >networks with relatively low performance. Nathan, Provided that you are not attempting to use ``best effort`` transport, the RTI-NG implementation should only have problems if ``inadequate network`` implies that the TCP connections are broken. If the connections are broken the RTI implementation would require some form of fault tolerance to automatically recover from connection failures (RTI-NG does not support this mechanism). But even in the case of network failure, a ``crash`` should not occur so there appears to be some problem with either the RTI or federate application. Was your problem submitted to the DMSO RTI helpdesk ? They may be able to provide assistance to understand the cause of the problem that occurred and if an issue exists within the RTI-NG implementation it can be fixed for future users. Steve

To unsubscribe from the Z-ARCHIVE-SIW-CFI list, click the following link: