4 ms·

I'm really glad to see this on the front page of HN. My company is building API-based security tools to help medical device vendors prevent these kinds of vulns

I'm really glad to see this on the front page of HN. My company is building API-based security tools to help medical device vendors prevent these kinds of vulns. There is huge risk in this class of IoT device, but not a ton of engineering effort going into addressing it.

falcrist 8y ago
It isn't just the internet connectivity. There are devices that are controlled by wireless handsets. Even it those handsets aren't internet connected, they constitute a potential vulnerability. I recently worked on an insulin pump that was both wirelessly controlled and potentially internet connected. Unit tests and communications security seemed to be an more of an afterthought than anything.
- mikekij 8y ago
  not surprising. The FDA (and other regulators) didn't start talking about cybersecurity (in earnest) until 2015, so most device vendors didn't do much "securing". this has definitely started to change though. Unfortunately there is a huge backlog of old devices out in the field that had little to no security designed into them. It will take ~15 years for them to be flushed out of the market.
- pdkl95 8y ago
  Using radio at all is probably always going to be a terrible design choice for any medical device that relies on a non-rechargeable internal battery. Last year's pacemaker security alerts included the usual braindead protocols and broken authentication, but the really interesting vulnerability was about power usage. The pacemaker didn't rate-limit it's "RF wake-up" actions; anybody with a radio could spam those commands just to get the pacemaker to turn on it's transmitter to quickly ("hours") drain the entire internal battery. The older designs that used magnetic coupling to communicated required holding a bulky device right next to where the pacemaker was implanted may have been annoying to use, but that kind of roadblock is a feature for pacemakers and other devices where changing the battery requires major surgery. Also, for anybody that missed it: Karen Sandler gave a very good followup talk[2] earlier this year where she discusses her attempts to get the source code for her own pacemaker. [1] https://ics-cert.us-cert.gov/advisories/ICSMA-17-241-01 https://ics-cert.us-cert.gov/advisories/ICSMA-17-241-01 [2] https://www.youtube.com/watch?v=8wPAHu_zYDw https://www.youtube.com/watch?v=8wPAHu_zYDw
  - jaclaz 8y ago
    >The older designs that used magnetic coupling to communicated required holding a bulky device right next to where the pacemaker was implanted may have been annoying to use, but that kind of roadblock is a feature for pacemakers and other devices where changing the battery requires major surgery. Right, not so unexpectedly, physically limiting access to devices (air gapped) remains the safer practice.

falcrist 8y ago

It isn't just the internet connectivity. There are devices that are controlled by wireless handsets. Even it those handsets aren't internet connected, they constitute a potential vulnerability. I recently worked on an insulin pump that was both wirelessly controlled and potentially internet connected. Unit tests and communications security seemed to be an more of an afterthought than anything.

mikekij 8y ago
not surprising. The FDA (and other regulators) didn't start talking about cybersecurity (in earnest) until 2015, so most device vendors didn't do much "securing". this has definitely started to change though. Unfortunately there is a huge backlog of old devices out in the field that had little to no security designed into them. It will take ~15 years for them to be flushed out of the market.
pdkl95 8y ago
Using radio at all is probably always going to be a terrible design choice for any medical device that relies on a non-rechargeable internal battery. Last year's pacemaker security alerts included the usual braindead protocols and broken authentication, but the really interesting vulnerability was about power usage. The pacemaker didn't rate-limit it's "RF wake-up" actions; anybody with a radio could spam those commands just to get the pacemaker to turn on it's transmitter to quickly ("hours") drain the entire internal battery. The older designs that used magnetic coupling to communicated required holding a bulky device right next to where the pacemaker was implanted may have been annoying to use, but that kind of roadblock is a feature for pacemakers and other devices where changing the battery requires major surgery. Also, for anybody that missed it: Karen Sandler gave a very good followup talk[2] earlier this year where she discusses her attempts to get the source code for her own pacemaker. [1] https://ics-cert.us-cert.gov/advisories/ICSMA-17-241-01 https://ics-cert.us-cert.gov/advisories/ICSMA-17-241-01 [2] https://www.youtube.com/watch?v=8wPAHu_zYDw https://www.youtube.com/watch?v=8wPAHu_zYDw
- jaclaz 8y ago
  >The older designs that used magnetic coupling to communicated required holding a bulky device right next to where the pacemaker was implanted may have been annoying to use, but that kind of roadblock is a feature for pacemakers and other devices where changing the battery requires major surgery. Right, not so unexpectedly, physically limiting access to devices (air gapped) remains the safer practice.