Optical illusions intrigue me. What intrigues me even more is that illusions have little to do with intelligence but are fundamental to human nature – they reflect how our brains have evolved over several million years. Business decisions, like any other, are often affected by similar illusions. One such illusion became evident to me as part of exploring the true cost of application operations for an enterprise.
Over the last 12 months I have been working with customers to understand what it takes to operate an enterprise class application. I have engaged with more than 60 customers to understand their application operations and we have collectively learned a lot during this time.
One area of operations that consumes a lot of resources is change management. Any change in the operational environment can potentially lead to an adverse impact on SLAs to the business and hence has to be managed through a well-defined process. Changes come in many forms, but software patches are particularly interesting.
What we noticed was that the application owners often don’t understand the economic impact of patching the full application stack over the lifetime of the application. If you take a squinty eyed view of your operations, patching can look something different from what it really is – an illusion.
As an example we asked one of our customers that had been running the same version of the software for 5 years, how many patches they had applied across the application stack over the last 2 years. The initial answer was – “Very few… we think”. Most customers I have talked to think that if they haven’t upgraded in several years the system is relatively “change free”. But, the real answer surprised me as well. For this customer, the answer was 100+ patches over the last 2 years as shown in the table below:
Number of patches
Application Server /Middleware
The cost of these patches is quite significant when they are applied properly and even more significant when those patches are not applied or are applied haphazardly. The cost of sub-optimal patching often shows in the form of system instability and often results in creating security risks. The cost of applying patches correctly directly stems from lack of application specific full stack expertise, which is essential for patch selection, patch testing and patch rollout in a way that is right for the specific business application.
A squinty eyed view of IT operations can result in an optical illusion of low operating costs around change management and leave you blindsided. When the details emerge the picture is likely very different.