In many ways, distributed open source software gives more social attack surfaces, because the system itself is designed to be distributed where a lot of people each handle a different responsibility. Almost every open source license includes an explicit disclaimer of a warranty, with some language that says something like this:
THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
Well, bring together enough dependencies, and you’ll see that certain widely distributed software packages depend on the trust of dozens, if not hundreds, of independent maintainers.
This particular xz vulnerability seems to have affected systemd and sshd, using what was a socially engineered attack on a weak point in the entire dependency chain. And this particular type of social engineering (maintainer burnout, looking for a volunteer to take over) seems to fit more directly into open source culture than closed source/corporate development culture.
In the closed source world, there might be fewer places to probe for a weak link (socially or technically), which makes certain types of attacks more difficult. In other words, it might truly be the case that closed source software is less vulnerable to certain types of attacks, even if detection/audit/mitigation of those types of attacks is harder for closed source.
It’s a tradeoff, not a free lunch. I still generally trust open source stuff more, but let’s not pretend it’s literally better in every way.
It’s a tradeoff, not a free lunch. I still generally trust open source stuff more, but let’s not pretend it’s literally better in every way.
Totally agree.
All the push back I’m getting is from people who seem to be worried about open source somehow losing a positive talking point, when comparing it to close source systems, which is not my intention (the loss of the talking point). (I personally use Fedora/KDE.)
But sticking our heads in the sand doesn’t help things, when issues arise, we should acknowledge them and correct them.
using what was a socially engineered attack on a weak point in the entire dependency chain.
An example of what you may be speaking about, indirectly. We can only hope that maintainers do due diligence, but it is volunteer work.
In the broader context of that thread, I’m inclined to agree with you: The circumstances by which this particular vulnerability was discovered shows that it took a decent amount of luck to catch it, and one can easily imagine a set of circumstances where this vulnerability would’ve slipped by the formal review processes that are applied to updates in these types of packages. And while it would be nice if the billion-dollar-companies that rely on certain packages would provide financial support for the open source projects they use, the question remains on how we should handle it when those corporations don’t. Do we front it ourselves, or just live with the knowledge that our security posture isn’t optimized for safety, because nobody will pay for that improvement?
100%.
In many ways, distributed open source software gives more social attack surfaces, because the system itself is designed to be distributed where a lot of people each handle a different responsibility. Almost every open source license includes an explicit disclaimer of a warranty, with some language that says something like this:
Well, bring together enough dependencies, and you’ll see that certain widely distributed software packages depend on the trust of dozens, if not hundreds, of independent maintainers.
This particular xz vulnerability seems to have affected systemd and sshd, using what was a socially engineered attack on a weak point in the entire dependency chain. And this particular type of social engineering (maintainer burnout, looking for a volunteer to take over) seems to fit more directly into open source culture than closed source/corporate development culture.
In the closed source world, there might be fewer places to probe for a weak link (socially or technically), which makes certain types of attacks more difficult. In other words, it might truly be the case that closed source software is less vulnerable to certain types of attacks, even if detection/audit/mitigation of those types of attacks is harder for closed source.
It’s a tradeoff, not a free lunch. I still generally trust open source stuff more, but let’s not pretend it’s literally better in every way.
Totally agree.
All the push back I’m getting is from people who seem to be worried about open source somehow losing a positive talking point, when comparing it to close source systems, which is not my intention (the loss of the talking point). (I personally use Fedora/KDE.)
But sticking our heads in the sand doesn’t help things, when issues arise, we should acknowledge them and correct them.
An example of what you may be speaking about, indirectly. We can only hope that maintainers do due diligence, but it is volunteer work.
Forgot to ask, but I would love to hear your thoughts on what @5C5C5C@programming.dev has commented about this subject: https://lemmy.world/comment/9003210
In the broader context of that thread, I’m inclined to agree with you: The circumstances by which this particular vulnerability was discovered shows that it took a decent amount of luck to catch it, and one can easily imagine a set of circumstances where this vulnerability would’ve slipped by the formal review processes that are applied to updates in these types of packages. And while it would be nice if the billion-dollar-companies that rely on certain packages would provide financial support for the open source projects they use, the question remains on how we should handle it when those corporations don’t. Do we front it ourselves, or just live with the knowledge that our security posture isn’t optimized for safety, because nobody will pay for that improvement?