3 ms·
With open systems and code, you can verify if code is doing the things you expect it to, and if you're paranoid enough, you can choose to build everything from
by heavyset_go 5y ago
With open systems and code, you can verify if code is doing the things you expect it to, and if you're paranoid enough, you can choose to build everything from the ground up. If there are enough security conscious people out there for there to be a market for security-focused systems, then companies can come in and offer verified systems and reproducible builds of things those people care about.
With closed systems, the best you can hope for is to treat it like a black box that can and will change out from under you.
- judge2020 5y agohaving a 3p come in and perform a build doesn't ensure there's not an attack targeted at you, though. You have no idea if yours is different unless you add PKI & CT logging to the mix (which wouldn't be horrible).
- heavyset_go 5y agoMy point is that with open systems, you have vastly more options depending on how paranoid you are and who your adversary is compared to closed systems.
- threeseed 5y agoWe can play this game all day. 99.999% of open source projects rely on Github Actions, Circle CI, Travis CI etc to build their code. Those are all proprietary so unless you are running your own CI/CD stack then you can't trust the code. And of course every open source project relies on libraries. So you need to make sure this applies equally to them as well.
- heavyset_go 5y agoYes, I'm pretty sure we're all familiar with "Reflections on Trusting Trust"[1]. Security comes in layers, and with open systems, if you want to, and if you have the resources, you can audit and verify those layers depending on your level of paranoia and your potential adversaries. You might throw your hands up in the air because security isn't assured and go all in on closed, black box systems, but there is value in open systems when it comes to security. [1] https://www.cs.cmu.edu/~rdriley/487/papers/Thompson_1984_ReflectionsonTrustingTrust.pdf https://www.cs.cmu.edu/~rdriley/487/papers/Thompson_1984_Ref...