Why the band OK Go is involved in systems engineering
In recent months, the band OK Go has managed to produce several videos that have gone viral on the internet. This included a Video recorded in parabolic flightwhich takes place in apparent weightlessness, or a 4-minute video, recorded in 4.2 seconds was.
What does this have to do with systems engineering? I will discuss this below.
This article was inspired by the TED talk "How to find a wonderful idea", in which OK Go discusses where all the ideas come from. Because many fans of the group wonder how all these crazy ideas come about.
Generating ideas is certainly also an important aspect of systems engineering. But I found another aspect of the lecture almost more interesting: The lecture begins with a Rube Goldberg machine. This is an unnecessarily complex machine without a precise purpose, in which a chain reaction unfolds. And such a machine is complicated and has significant complexity. These are characteristics that we also find in systems engineering. Furthermore, such a machine consists of many parts. The machine shown here consists of 130 elements.
OK Go has now asked itself how reliable such a machine is. Or more precisely, how reliable it has to be so that it doesn't drive the band members crazy.
Statistics
To determine reliability, OK Go initially assumed a reliability of 90% per part. The math is simple, 0.9130. The result is staggering: statistically speaking, none of the parts would fail in just one in a million attempts. Even at 99% per part, four attempts would be necessary.
Even at 99% reliability per part, a machine with 130 parts fails 75% of the time [tweetthis]Even at 99% reliability per part, a machine with 130 parts fails 75% of the time[/tweetthis]
OK Go did what many systems engineers would do: They encapsulated and focused on the parts first to make them highly reliable. In this phase, the band then changed some of the parts considerably or discarded them. Only then did they consider how the highly reliable parts could be put together.
Of course, someone who builds a Rubert-Goldberg machine has an advantage: after all, no predefined function has to be implemented. There is only the non-functional requirement: "It has to be fun and surprising".
Brainstorming
This process also explains why the band rather modestly says that they discover ideas, not invent them. And this discovery process is also something that is often used in systems engineering: Using prototypes. When OK Go has an idea, it simply tries it out: On the one hand, to find out whether the idea really has the desired surprise effect. But also to ensure that the idea is 100% reliable.
We use a Sandbox to find ideas that are not just surprising, but surprisingly reliable (OK Go) [tweetthis]We use a Sandbox to find ideas that are not just surprising, but surprisingly reliable (OK Go)[/tweetthis]
Conclusion
The band doesn't talk about one last aspect, but in my opinion it's at least as important as everything else: Enthusiasm. But the enthusiasm is clearly recognizable. In Marco Schmid's keynote speech at the ESE Congress 2017, the enthusiasm that led to the success of the project described above was also evident.
You can certainly argue about taste in music, but the videos by OK Go should inspire every systems engineer.
Image: TED / OK Go
The statistics section describes a major problem with micro-services. A problem that most so-called architects are not aware of. The company can then see how it can operate the oh-so-elegant architecture with the required reliability.
I have already been involved in two cases in which architectures based on micro-services had availabilities of approx. 40 to 50% in production operation and had the existing micro-services reprogrammed so that the required >97% availability was achieved.
As a little mental exercise:
Where was the error in the original realizations?
/pica