AWS eliminates human interactions with its customers’ data

Amazon Web Services (AWS) is "paranoid" about protecting customer data, according to CISO Stephen Schmidt, who is on a mission to cut out human error by increasing automation.

Speaking at AWS Enterprise Summit in London today, Schmidt highlighted some of the protective measures its data centres feature, which its forthcoming UK facilities will presumably share.

"Access to customer data, it's something that we are incredibly paranoid about," he told delegates. "Rather than saying the rule is that customer data on disk can't leave our data centre, so you have to check and see if there was ever customer data on this disk before it can go out, we decided it was more straightforward and enforceable if we said no hard disk will leave our data centres intact, period."

Staff access to data centre facilities is also restricted, with some teams able to access the physical infrastructure but not access to the software stored on it, while the opposite is true for other teams.

AWS has long expected to launch its UK data centres at the end of this year, or early 2017, a timeline that the company didn't narrow down today, though UK and Ireland MD, Gavin Jackson, said: "We're bang on track and we can't wait to open the doors to that in the near term."

CISO Schmidt also warned against the threat of human error, "the number one cause of mistakes". In a bid to widen AWS's appeal to enterprises, he described how over the course of 2016 he has automated many procedures that used to be carried out by humans.

"Even though the business is doubling I wanted 80% reduction of staff access to data," he said. "Because people make mistakes. People's credentials get compromised, people make errors."

Automating human interactions with data cuts out these errors, leading to better uptime and protecting against hacking threats from the likes of nation-state actors, according to Schmidt.

"You remove the opportunity for people like the Chinese government to steal credentials of your staff and reuse them for access to your data," he explained.