Reviewing an exercise text for a local university, I ran into an issue that seems to be overshadowed by all attention for the new ITIL 4 terms like SVS, value chain, co-creation etc. What has become of the availability measures we used for some decades?
I think only two terms are really important to create meaningful measures of availability:
- the average uptime (or the complementary downtime)
- the distribution of this uptime (or downtime) over time
- The average uptime will be defined as the percentage of the time the service indeed delivered its agreed functionality. This is what ITIL v3 called MTBF - the Mean Time Between Failures.
- The average time between subsequent incident starting times tells you how downtime is distributed over time. This is what ITIL v3 called MTBSI - Mean Time Between Service Incidents.
Both factors are required to get a decent idea of what happens in practice. After all, 99.99% uptime for a 24*7 service translates to one annual incident of 53 minutes, but also to 100+ incidents of half a minute each. From a user perspective, this could make all the difference between an acceptable service and bankruptcy.
ITIL 4 presents a different set of definitions:
- MTRS - Mean Time to Restore Service - still is the same metric for service downtime: "a metric of how quickly a service is restored after a failure".
- MTBF - Mean Time Between Failures - however now is no longer a measure for uptime, but is now defined as "a metric of how frequently a service or other configuration item fails". This was previously called MTBSI in ITIL v3.
- The term MTBSI is not part of the ITIL 4 Foundation book, nor part of the ITIL 4 Glossary, so it seems to have been dismissed, just like the term MTTR.
This means that the ITIL v3 equation "MTBSI=MTBF+MTRS" is now replaced by the following ITIL 4 equation: "MTBF=MTRS+average uptime".
A few weeks ago, after a 'post implementation review', AXELOS published a number of errors in the ITIL 4 Foundation material of last January. These errors included some typos in text and graphics, and a few more significant changes, including the replacement of Change Control by Change Enablement, and some relationships between practices and value chain activities. The new definitions of availability terms were not part of the error list, and until now I haven't seen any of the ATOs or any of the regular ITIL experts writing about it. Maybe I've missed something?
The new definition of MTBF has been part of the Glossary and the Foundation book for almost a year now, and in the mean time it has been translated by many subject matter experts in the AXELOS translation teams. It seems we can only accept the fact that the availability equation has now changed permanently.
Have you changed your SLAs and service reports to reflect this?