It has been historical to measure availability as uptime % for most of the services. There has been known pitfalls in this measure of uptime and most of the time is does not make any sense to the customer when we talk of uptime % This paper deals with a different technique of base lining and measuring the availability of a service and NOT on service components.

The uniqueness of this technique lies in fact that we no longer measure availability using service component uptime % ( like server uptime, db uptime, network uptime , etc and them arriving at a service uptime by calculating the product of the uptimes) The second part deals with the calculation of throughput of the service based on volumes of transactions.

This technique is based on the moving average principle, where the volume of transactions is measured is compared against the same period of time intervals over a period of time ( say for the past 12 weeks). The average deviation % from the moving average gives the performance deviation of the service. This technique is very useful when end - to end measurement is not possible for transactions and where we ant to measure the throughput.