Cube Wall

Sarah and Phil were system performance engineers that sat in the same corporate cubicle island with a shared adjoining back wall. They talked constantly about work and it was creating problems. The other local inhabitants of their area wined when they talked too loudly through their adjoining wall. So they stopped.







Building Services then protested that it was a safety hazard for them to stand on chairs and have conversations over their cube wall.






So they stopped.









And they were fresh out of ideas, when finally their manager complained that they were wearing out the carpet walking around from one cube to another.





So they stopped.



Although intended to promote productivity, modularized office cubicles systems sometimes get in the way of how people do their work. Desperate for a solution, they negotiated with Building Services to remove one of the cubicle wall panels between their cubes, making it possible for them to communicate in an effective manner.



Sarah: “Marketing sent out a confusing customer mailing, causing thousands of unnecessary calls to customer service. On the back end database we observed five times more activity than on any normal day, until database redo logs fell behind resulting in an outage. The emergency fix was a change to a database configuration parameter.”

Phil: “No performance simulation can predict the impact of a confusing customer mailing.”

It took production performance problems to provide the necessary feedback to stabilize the environment.


Phil: “In performance testing of the new web based document retrieval system, the CPUs were idle, but now in production the CPUs are pegged at 100%. What went wrong?”

Sarah: “The older system did not support search, so users were forced to navigate to documents using hyperlinks. When we simulated the users in performance testing, we falsely assumed the users would continue to use hyperlinks. However, those dang users changed their behaviors and started using the new search engine far more than we anticipated.”

In performance engineering, real users will frequently change their behaviors in creative and unexpected ways when they interact with new systems, creating workloads that affect change. No one was able predict a change in the way users interacted on the new system, just as no one was able to predict the way Sarah and Phil interacted.

By removing the wall panel between their two cubes, Sarah and Phil achieved the minor celebrity status with their co-workers. They demonstrated that they were able to apply their system performance engineering skills to solve a cubicle performance engineering problem. As the news of their solution spread, it drew attention to their new cubicle arrangement and people came by to admire their innovation.


But now a new problem emerged. People started using Phil and Sarah’s cubical opening quicker thoroughfare to the company cafeteria.


Copyright Dave Nocera 2007





Office Terms