ARP requests affecting ECOM100's

[jsatkanpak]

We've noticed that a high number of ARP requests (> 40 per second) on the network can cause a lot of issues with the ECOM100 modules, such as Transport Protocol Errors and other issues, like "breaking" a DirectSoft link so that it has to be rebuilt.  We haven't seen the issue on the regular ECOMs and are wondering what the difference is between the modules that would cause this.   

Obviously, high ARP traffic is generally a "Bad Thing" and we've always been able to trace it down and eliminate it -- and it's usually due to software that "doesn't belong". 

[BobO]

ECOM100s use a full TCP/IP stack, as opposed to the limited UDP implementation that we used in original ECOMs. While we have made every effort to test the stack...and fixed everything we've found wrong...there always seems to be yet one more stack related issue. At some level it is simply impossible to anticipate every possible network config, and our users seem to catch the stuff we've missed. Thanks for the help!

Anyone else here at Host have some specific thoughts about what might be going? Too much ARP traffic definitely isn't a good thing, but we shouldn't be killing stuff with it...