Slashdot Mirror


Boeing 757 Testing Shows Airplanes Vulnerable To Hacking, DHS Says (aviationtoday.com)

schwit1 shares a report from Aviation Today: A team of government, industry and academic officials successfully demonstrated that a commercial aircraft could be remotely hacked in a non-laboratory setting last year, a DHS official said Wednesday at the 2017 CyberSat Summit in Tysons Corner, Virginia. "We got the airplane on Sept. 19, 2016. Two days later, I was successful in accomplishing a remote, non-cooperative, penetration. [Which] means I didn't have anybody touching the airplane, I didn't have an insider threat. I stood off using typical stuff that could get through security and we were able to establish a presence on the systems of the aircraft." Hickey said the details of the hack and the work his team are doing are classified, but said they accessed the aircraft's systems through radio frequency communications, adding that, based on the RF configuration of most aircraft, "you can come to grips pretty quickly where we went" on the aircraft. Patching avionics subsystem on every aircraft when a vulnerability is discovered is cost prohibitive, Hickey said. The cost to change one line of code on a piece of avionics equipment is $1 million, and it takes a year to implement. For Southwest Airlines, whose fleet is based on Boeing's 737, it would "bankrupt" them. Hickey said newer models of 737s and other aircraft, like Boeing's 787 and the Airbus Group A350, have been designed with security in mind, but that legacy aircraft, which make up more than 90% of the commercial planes in the sky, don't have these protections.

4 of 140 comments (clear)

  1. million dollars per line by Speare · · Score: 5, Informative
    I expect quite a few folks here are going to question the figure, "a million dollars per line changed."

    1. The airlines operate under a huge amount of regulatory oversight, and structure the development of avionics or engine control software accordingly. The terms ARP4754 and DO-178C are to aviation as ISO9002 is to business models. They provide guidelines on creating a rigorous development process, and regulators are keen to track how well companies develop logic and physical designs in line with best practices described by those guidelines.

    2. If you summarize DO-178C in one sentence, it might be "document the rationale for every change, and the means you employed to ensure it is the right change." Most companies follow a V-shaped change model where you trace from high level requirements to lower level requirements to implementation details, and then verify the code does what is expected and then validate that the requirements are being met (and the requirements are even proper in the first place). Once you have that framework in place, you have to document every step of the chain of review.

    3. For every change to a high level requirement, a low level requirement, an implementation, and sometimes even a change in a verification method, there typically has to be an independent review: you cannot trust the implementors to check that the change was appropriate and done correctly as it's easy to be blinded by your own thought process during development.

    So in a case like this, the customer needs to inject several new top-level requirement (which shockingly may not have been there in the first place), "the system shall be hardened against unauthorized changes in configuration/operation/state" and that has to flow down to subsystems "the component XYZ shall be hardened..." and that has to flow down a few more tiers before you even identify the protocols or chips or attack vectors to be changed. Then you have to verify the code change works in each component. Then a system-level review. Then a regulatory review to have the updated design certified as safe for test flight and finally safe for revenue service.

    Does this sound like a desktop software change control process? Sure, maybe you're really disciplined, but it's a matter of degree. It really can take fifty people or more, from regulators to systems engineers to coders to integration testers to work the process. And that all adds up in terms of time, opportunity costs, tools and tooling, lab test, systems test, hours and hours of live aircraft flight test, and so on.

    --
    [ .sig file not found ]
  2. Re:why should Southwest Airlines pay? and not boei by michelcolman · · Score: 5, Interesting

    What if a hacker takes down an airplane, people find out in the media, and nobody wants to fly on that aircraft type anymore? Or with that company because it didn't apply a fix that existed? Does the insurance cover that? Now that's something that could bankrupt an airline.

  3. Re:Useless metric spotted by leonbev · · Score: 5, Insightful

    With something like avionics software, it probably doesn't matter if one line or a thousand lines change... the entire application would need a full regression test for safety/certification purposes. That's where the million dollar estimate probably comes from.

  4. Re:Legacy aircraft by afidel · · Score: 5, Informative

    Actually the 737 is just as modern as any aircraft being produced. The current airplane with the designation 737 shares virtually nothing with the first plane to carry that designation. The fuselage is different, the wings are different, the engines are different, the avionics are different, and the interior packages are different. The currently produced aircraft with the designation (The Max series) is actually the 4th generation of 737. Basically saying it's a 737 is like saying it's a Ford Mustang, other than size and maybe some styling cues it's fairly meaningless as it tells you nothing about what's in the car/airplane.

    --
    There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.