Blog

Forget “Mean Time to Innocence” and switch to collaborative problem solving

Well, let’s first define what we mean by Mean Time to Innocence?

This is when your main purpose is to prove that you are not to blame if something goes wrong somewhere. It can be seen as a more politically correct variation of the famous C.Y.A. — Cover Your Ass — approach.

The idea is to have a product — whatever it does officially — whose main use is to prove that you are not, in any way, responsible for the issue (e.g., failure, slowdown, user complaint) that occurred in your IT infrastructure.

Sometimes, even if no one would admit it — at least officially — it might have been the true reason for purchasing a solution.

The symptoms of “Mean Time To Innocence” syndrome

Especially when teams are more in a competition, or a fighting mode rather than a cooperative one, you will encounter these kinds of strategies:

  • The network team wants to prove that, as they always claim, it is not the fault of the network
  • The virtualization team wants to prove that, as they always claim, it is not the fault of virtualization
  • The system team wants to prove that, as they always claim, it is not the fault of the system
  • The application team wants to prove that, as they always claim, it is not the fault of the application
  • The database team wants to prove that, as they always claim, it is not the fault of the database
  • The storage team wants to prove that, as they always claim, it is not the fault of storage
Mean Time To Innocence involves a lot of fingerpointing
Mean Time To Innocence involves a lot of fingerpointing

Every team has a quick look at its own tools. So each team will search — not too long — for their usual suspects:

  • Connection, retransmission, TCP lifecycle issues …
  • Incorrect resource assignations, overloaded hosts, virtual machine moving issues …
  • Software updates incorrectly applied, outdated policies, driver or firmware issues …
  • Database workload, SQL errors, slow queries, …
  • Application errors, slow application response times, resource consumption issues …
  • I/O saturation, read/write policy balancing, caching mechanism issues …

Each of these efforts adds up, increasing the mean time to innocence.

What are the consequences?

Guess what! You will have quite uncomfortable meetings where everyone is trying to prove his or her innocence and eventually tries to reject the blame on one of the other teams, more or less randomly, depending more upon who you like or not, than on real technical and objective reasons. In the end, everyone may have succeeded in proving his or her innocence. And so what? The issue is still there. I’ll let you imagine the atmosphere of the next meeting on the same topic!

For your users and customers, time flies and your organization loses productivity and revenues! The mean time to innocence has no business value for your customers or your users.

Focus on solving the problem rather than fingerpointing
Focus on solving the problem rather than fingerpointing

This self-justifying approach sounds sad, but it is real! I have seen people claiming it directly on their t-shirts: “It’s not the network!” written in a big font size. In those cases, people tend to forget that the real priority is to have an optimal service running for their customers, not just proving that it was not their fault if something went wrong. Note that when this kind of thinking is in place inside teams, it is more likely the result of an inefficient punishment/rewards management policy against which people have developed such defensive strategies…

How to get to team collaboration

We believe IT organizations should get rid of those approaches to troubleshooting that are based on blaming so called guilty people and should adopt an approach to troubleshooting that is based on cooperation to make things work as quickly as possible in an efficient way.

Sharing performance analytics across teams (network, systems, applications, database) can be easy and quickly pointing people in the right direction to fix performance degradations faster doesn’t have to be complex. (In fact, with SkyLIGHT PVX, it isn’t.)


If you haven’t already given it a look, SkyLIGHT PVX is a network and application performance monitoring and management tool that can be used by network, virtualization, system, application, database, and storage teams to quickly locate the source of performance degradations and speed their resolution — collaboratively — with a shared view of IT infrastructure and objective measurements of performance.