You may want to try creating an Outage record for Incidents related to a service interruption (or CI outage). The Outage record tracks the start time, end time and duration of an outage for a CI.
Using Wikipedia's definition (Mean time between failures - Wikipedia, the free encyclopedia) MTBF is calculated by:
Sum of (start of downtime - start of uptime)
---------------------------------------------------------
number of failures
Using Outage records, you should be able to calculate the time difference between Outages (difference of end time of one Outage and start time of following Outage), and divide their sum by the number of Outages for a given service/CI. A business rule triggered by an end time being inserted into an Outage record could calculate the MTBF value.
You could add a custom field to the CI record to store the calculated MTBF value and use that for reporting.
Hope this helps.
