I don't know where the 127 limit comes from. The numbers I recall are much higher (e.g. 1024 IP stacks) but my milage may vary. You're not running into some arbitrary limit that someone created by defining just 127 devices in the IOCP are you? But there will be a limit somewhere (if only the locked buffers under the bar).
When you get close to such a limitation, I would not suggest the approach that Adam suggests. Unlike what was hinted at before, we don't have a HiperSockets version of VSWITCH, but an OSA Express shared between the two LPARs does get close. The OSA microcode will send back the packets into the other LPAR without touching the wire. I expect the cost of a virtual router and the dynamic routing stuff that Adam suggests will take so many resources that your effective throughput will be lower than when you let the OSA Express card do it for you. If you have some Linux servers where you really care about high traffic etc, then you could consider to give those a dedicated HiperSockets device. Rob -- Rob van der Heij rvdheij @ gmail.com ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
