|
|
|
|
|
by nickpsecurity
3997 days ago
|
|
Interesting article and interesting guy. I like the work he and his team does on these apps. Unfortunately, they typically run on the type of endpoints that everyone from script kiddies with money to High Strength Attackers can hit. Usually alongside apps not as strong as theirs on TCB's that can at best be described as insecure foundations. I recommend against such apps and platforms for anything other than stopping the riff raff. That's what I use them for. I pointed out the difference between secure code and secure systems in this [1] writeup. Shared much of my framework for analyzing or designing-in security in the process. The TCB of most solutions today is ridiculous: people are building on foundations of quicksand. There's only a few exceptions I've seen such as GenodeOS (architecturally) or Markus Ottela's Tinfoil Chat. Markus has been unusually alert to our concerns and updated his app appropriately even for covert, channel suppression. Quick question: which of the many crypto apps on the market can deliver a covert channel analysis to you at app and system level? Answer: few to none despite it's importance over decades with a rediscovery in past 5+ years in mainstream security. Strong security is hard. Moxie seems awesome as a coder and good to great in both crypto and OPSEC. Thing is, his offerings break the decades old rule of having a strong TCB. Just like most of the rest. It's why they're usually bypassed or broken by strong attackers. Gotta do the whole thing with concern for each aspect of the system. TFC is a clever cheat on that even more than my MILS scheme with a KVM and a highly-assured guard. If you don't cheat around it, you better do it right or your users will suffer the consequences. Those trying to contain vulnerabilities of mainstream OS's and components with any success are expending literally hundreds of thousands of dollars worth of labor per year. It's why I push for clean-slate, hardware and software platforms like DARPA and NSF have been funding recently (eg SAFE, CHERI processors). Alternatives using COTS tech are pretty complex and most users will probably fail to secure them to be honest. [1] https://www.schneier.com/blog/archives/2013/01/essay_on_fbi-... |
|