LDN - Luminis Developer Network

For quite a while, we have seen single sign-on from Luminis to Banner break when our systems are under heavy load (ie: registration, start of term) The errors in cp.log look like this:

[2007-03-09 14:43:13,063] [ERROR] WebServlet [cpip]: CPIP ERROR sct The system 'sct' is responding too slowly, please try again after a short wait
[2007-03-09 14:43:41,403] [ERROR] WebServlet [cpip]: CPIP ERROR sct The system 'sct' is responding too slowly, please try again after a short wait

After a (configurable) amount of time, CPIP finally gives up, and writes of that system as "permanently down". The only way to get single sign-on to Banner working after that is to restart the Luminis web server.

This looks like a combination of (a) deficiencies in the CPIP protocol itself, and (b) the way that Sungard chose to implement CPIP with respect to Banner.

My questions:
1) Are you experiencing the same "SSO breaks under heavy load" problems?

If so, then:
2) For how long? When was your first problem report to Sungard?
3) How often / what times of the year?
4) What have you done to monitor or resolve this?

Thanks,
Todd