The COVID-19 virus still dominates life as we know it, and this makes the news about vaccine effectiveness very welcome. Despite our familiarity and accommodation with computer viruses, we still find human viruses to be extremely disruptive. It's not just about money. When cyber malware hits healthcare providers, the results are distressing, as one of our news items demonstrates. All of this leads me to wonder if software and biology have any non-trivial similarities that could help us prevent future pandemics.
The critical point is that organisms can detect "foreign" proteins. Proteins have shapes, and some shapes are just innately non-human. We can detect unusual software by using learning to define "normal" or "self" patterns. System call patterns, memory usage patterns, and other easily monitored features can help distinguish normal from foreign. But biological proteins and immune systems have evolved over the course of a billion years, their "rules" are highly variable, yet combinatorialy limited and tuned to survival. Software is more flexible, less rule-driven, and tuned only to functionality.
Computers, vulnerable as they are, give us the tools to understand biology. With software, we can visualize and understand proteins, delineate and duplicate individual genes, design antibodies, and ultimately cure and prevent diseases. Perhaps biology will help us find the tools that will let us protect software and data. As we have been seeing, that kind of trust is essential to modern societies and their democratic processes. Christopher Krebs, until recently head of the Critical Infrastructure Security Agency of DHS, understands that well and is at the moment an unsung hero of this intersection of democracy and technology.
Now is the winter of our discontent
Lurching towards glorious summer, should
COVID-19 to mRNA relent.
And all the clouds that lour'd upon our world
In the warming bosom of the ocean buried.
(Apologies to The Bard),