Zoran Perkov is one of the heroes portrayed in Flash Boys. Someone liked my review of the book, which made me re-read the review and setup a Google alert on him. This video was one of the first hits on a search for him. His thing is:
Failure is feature of complex systems.
Basically when you build a complex system, you risk failure. Complexity allows us to more easily adapt to customer need. Complexity also increases the risk of novel failure. Things we did not even know could go wrong (unknown unknowns) probably will. The trick is, responding to them.
Something he mentioned I want to study more of:
Basically, we should have mechanisms in place which act to reduce the effect of performance spikes. The more options we have to deal with these events, the better we will be able to handle them. The smaller the toolkit, the less capable we are to respond.
The delivery is not that great, but the ideas behind it is something I’ve agreed with for a long time.