This past weekend my family watched the new Netflix movie Enola Holmes, which was a thoroughly entertaining story that cared little that its characters bore minimal resemblance to the “real” Holmes family.

Mycroft diminished
The biggest departure from the original stories was Mycroft. For those less nerdy than me, Mycroft Holmes is Sherlock’s older brother, who is smarter even than Sherlock, but who operates purely in the theoretical realm.
In the movie he’s shown to be a less-talented and rigid member of society, but in the source material he rarely leaves his arm-chair and has an even more impressive brain than Sherlock.
If the art of the detective began and ended in reasoning from an armchair, my brother would be the greatest criminal agent that ever lived. But he has no ambition and no energy. He will not even go out of his way to verify his own solutions, and would rather be considered wrong than take the trouble to prove himself right.
“The Greek Interpreter” from The Memoirs of Sherlock Holmes, by Sir Arthur Conan Doyle
By contrast Sherlock is full of energy, and goes to great lengths to physically investigate and prove his theories.
Network redundancy in theory
We’ve been doing some work recently on network redundancy and resilience, and I’ve been struck by how beneficial both the theoretical and the practical approach can be.
As we attempt to replicate Mycroft’s powers, we can imagine a list of all the possible things that could fail in a network (each piece of hardware, each IP connection, each third-party service, and so on) and ask ourselves two theoretical questions:
- What would happen if X failed?
- How could we design the voice network to be resilient to that failure?
This is a powerful approach, and it’s often possible to spot design flaws, or identify areas that need more research purely by thinking systematically about the network architecture.
Network redundancy in practice
But Mycroft was only able to get so far by smoking a pipe in his armchair – the master detective was Sherlock. Sherlock not only spent time thinking, but also went out into the world to gather more data and to test his theories.
In the same way, we can only feel truly confident that a network is resilient to a particular failure if we test it. Most people are understandably reluctant to do much failure testing – it’s a lot of effort, and there’s risk involved even in the testing – but if you want a truly redundant and resilient voice network, you need to prove it really works.
If you’d like us to conduct either a Mycroft audit or a Sherlock audit of your voice network, drop us a line and we can talk.