Technology innovation is about knowledge not gadgets

Data61's chief research scientist says the real innovation in cybersecurity will come from technologies that are thousands of years old, plus some professionalism.
Written by Stilgherrian , Contributor

Technology comes with a mental trap, says professor Bob Williamson: "It's easy to get trapped into thinking you're talking about an artefact."

Williamson is chief research scientist at Data61, Australia's data science research and engineering agency, and he prefers a definition used by economist professor Ola Olsson and others.

"Technology is useful knowledge, nothing more than that," Williamson told the SINET61 conference in Sydney on Tuesday.

"There's one technology that I think actually dominates all of them, which is mathematics," he said. "Mathematics is what has underpinned the scientific revolution. That's what allows you to build a computer screen like [the one on the stage], and buildings that don't fall down," he said.

The two bits of mathematics that really matter, Williamson said, are probability and proof.

Probability is the basis of insurance, a business that dates back to prior to Babylonian times, and was a result of the growth of shipping.

"What insurance does is it turns something that is either zero or one into a continuum," Williamson said.

"If you owned a ship and it went down, you were ruined. But fortunately, only one in five, or one in 10 ships went down, so if you spread that risk, life is good. Now it's not just about helping you not go bankrupt. What it does is it puts a price on the quality of the ship," he said.

"Piracy can ruin your cargo. Incompetence of the captain can ruin your cargo. Remember the Exxon Valdez? Don't be drunk when you're driving a supertanker. But there is a fundamental thing that you can fix, which is you can make the ships more reliable. Ships that don't break in big waves. Ships that don't get lost because their navigation systems aren't so good. Once you have maritime insurance, you can price the value of that."

Proof goes back Ancient Greek mathematician Euclid, according to Williamson. Certainly Euclid's Elements, published around 300 BC, was "instrumental in the development of logic and modern science", as Wikipedia puts it.

"Software is mathematics, and if you want reliable software, you have to be able to prove that it's true. Now the technology does exist to do this. It's reached the point where you can actually prove significantly complicated systems. But it costs, so people don't want to go and do it," Williamson said.

Data61's own formally proven seL4 microkernel is an example of such a system. It's now the basis of a Department of Defence project to allow classified and unclassified data to be processed through a single interface without compromising security or function.

"I think the innovation will be when there's a sufficiently liquid market caused by cyber risk insurance that will fund building proper software," Williamson said.

"Once you've built proper software, you've at least solved one of the problems. There's still pirates, there's still drunk captains. I can't fix that. But maybe we can make the technology work," Williamson said.

Checklists don't make you dumb, they make you professional

Aviation and medicine are industries that we trust with our lives, but they have very different attitudes to safety. According to Williamson, the number of iatrogenic deaths, which are deaths caused by medical mistakes, is equivalent to two Boeing 747 aircraft crashing per day.

"The aviation industry knows how to deal with error and failure. It talks about error and failure. The medical profession ... they don't want to talk about it, and their record in that is terrible. Putting a price on it, and having a feedback loop to fix it, is something the aviation figured out long ago," he said.

A key component of aviation's safety culture is the preflight checklist, which was developed after the fatal crash in 1935 of the prototype Boeing Model 299, the aircraft that became known as the Boeing B-17 bomber.

"Boeing built an aeroplane with four engines and all sorts of features. It was too complicated for a pilot to handle, and it crashed. They forgot to do all of the right things," Williamson said.

Smart people sometimes find checklists insulting, because they know they know what to do. But a true professional follows the checklist every time, because they know that like anyone, they can make mistakes.

Editorial standards