Does anyone have a model that can be used to set some kind of availability service levels for Midrange Servers (UNIX, AS400 and also Intel/NT servers) that primarily run business applications?

My major areas of concern include: 1. being able to seperate time taken to restore the infrastructure and the application and attribute the responsibility accordingly 2. if an application has server redundancy (i.e. a number of servers that provide the same function and if one goes down, the application remains fully available), how does that impact the midrange server availablility ? 3. Following on from the above point, the SL needs to be able to measure true business impact instead of just server up/down.

-- Jimmy Bartolo (jimmyb@au1.ibm.com), April 06, 2003