4 ms·
re firmware. It's true and orthogonal. Just as the DARPA project is doing, you use the best tech and method for each part of the job. The recommendations I gave
by nickpsecurity 11y ago
re firmware. It's true and orthogonal. Just as the DARPA project is doing, you use the best tech and method for each part of the job. The recommendations I gave her were for kernel up mainly for isolation. Drivers should be written robustly, synthesized, use trusted hardware, and/or mediate the hardware.
re x86 vs ARM. It's true. Remember, though, that I didn't tell her to use seL4: just re-use work in L4 family projects like Nizza or OKL4 to get their security + performance benefits. The idea is you get a good start, make sure you can swap in/out components, and someone puts a verified one in place later. This is exactly the approach GenodeOS team took along with using many things I recommended in other post. Only gripe with them is they're doing too many things at once rather than focusing on getting at least one in best, production mode.
That said, I agree with keeping off x86. I've recommended ARM, MIPS, SPARC, and RISC-V for future work. SPARC is nice due to open specs, no licensing past $99 trademark, and GPL hardware available. RISC-V people building all kinds of things. SPARC and MIPS have already been modified for enhanced security by academics several times over. So, SPARC (esp Gaisler), MIPS, and RISC-V are my recommendations in that order putting usability/extensibility over pricing.
re pragmatic approach. Sure, I agree with that part. The resulting approach combined things top on the CVE list onto a platform she published exploits for and criticized in other posts. Gotta wonder how far its security would go. ;) She was a good bug-hunter, though, when she understood how something worked. Didn't seem to understand strong, security engineering so I predicted QubesOS would be good for vanilla malware w/ some containment of advanced attacks. Basically, a nice improvement on and replacement for Compartmented Mode Workstations's. The TCB size and implementation are just too complex for high security.
re secure browsers. For full-featured experience, you'll need either virtualization or my old KVM-switch method. More research in the former has only made the latter look more trustworthy over time haha. Anyway, I referenced a number of more-secure browser architectures in the comment below. Enjoy seeing variations of The Right Thing in action. :) Now, if only mainstream would put effort into improving those instead of the monolithic ones. Chrome was at least an OP-inspired attempt. Due credit for them.
https://news.ycombinator.com/item?id=9969961 https://news.ycombinator.com/item?id=9969961
- hga 11y agoMy complaint about browsers is less about architecture (I mean, when using multiple processes is a vast improvement of the state of the art on the ground...), than what I gather is the vast amount of work required to get today's standard web pages to render well. There are, what, 4 state of the art rendering engines? (Two Microsoft, which has lots of money to spend, Webkit (now Apple and Google forked??), and Firefox's.) And there's much, much more to the security story for browsers. Then again, this is not a topic I've looked closely at, maybe there's more hope than I think, and continuing financial losses might prompt more rigor, although I don't recall the worse ones being purely browser exploits.
- nickpsecurity 11y agoOh I see what you mean. Yes, it's become an oligopoly with a huge barrier to entry. The typical solution is to separate all these components with safer implementations, isolation, and/or interfaces. What I referenced do those. If you want to solve it directly, you have to make the code trustworthy. That's... another matter entirely. There are a few approaches to this. One is to apply compiler transformations that make the code memory-safe along with interface checks at sys-call level. An example is Softbound+CETS albeit I'm not saying it's worked on a browser. Several have compiled C programs to run in a Java or other managed environment. Another example is a lighter security approach called control flow integrity where you just try to block code injections with data or pointer protection. Control Pointer Integrity is most interesting right now with several HW accelerators. The next approach is using software or hardware obfuscation to make it where the incoming attack (a) wouldn't hit its target or (b) would be mangled/detected by crypto work at HW level. This has been ported to Linux. CheriBSD is another, DARPA-funded effort that uses custom hardware (CHERI) and tooling to let you incrementally protect legacy stuff with capabilities. So, quite a few ways to deal with it. As usual, the problem is that our machines and languages make it easy to cause a vulnerability with common, fundamental operations. Making those operations fundamentally safe[r] knocks out a lot of that risk. Traditional techniques, like separation kernels, can try to isolate and detect the rest. That's my take on how to deal with being stuck with codebases like Webkit. Still work but feels less impossible.