Tag Archives: computer security

Do you trust all your internets to strangers?

Question: Japanese government, Staat der Nederlanden and Starfield Technologies Inc. – what do they all have in common? Answer: You trust all of them to verify identities and security on company websites and internet services.

Are you Japanese or Dutch? Have you ever heard of Starfield Technologies? I guess not. Still you’re probably using a web browser which has these “certificate authorities” as trusted “roots”. In short this means they can – with no security errors whatsoever – impersonate your bank, eavesdrop on your logins and make your computer believe everything is just fine. But only if they were to man-in-the-middle your communication physically or somehow manage to poison your DNS lookups of course.

So the scenario isn’t really your average Joe security issue, but the heart of the issue is a very important one. It’s a question of trust – a deep conviction of truth and rightness – and “trust” – the simple term used for computing security (though without the quotes).

All computer interaction is based on trust. You trust the computer is doing what it tells you it’s doing (which Free software lets you verify). However you also trust the other person’s computer to do what you’re asking it to do – and nothing more! Given the global scale of networking and computing, this is a hard task to verify manually – thus trust is given to cryptographic algorithms mathematically ensuring your data is handled by the correct entities without unwanted manipulation.

No ordinary person can ever understand, verify or control all parameters required for 100% secure computing. This is probably the reason why browsers (Firefox, Chromium etc.) include packs of “globally trusted” certificate issuers, such as Verisign, GoDaddy etc. Private companies that are virtually the foundation of today’s DNS-based internet, controlling or supplying top domain names. This is however where one should start worrying about who controls what. Remember what Verisign did to Wikileaks.org? Did you know that they want to make it easier to happen again?

So when one can’t trust the big boys, how can one really have trust in small, unknown organisations or companies? What are their respective thoughts on free speech, free internet and policies on eavesdropping? Wouldn’t you actually prefer a system of peer-to-peer review? The “web of trust” model as it’s called when a user trusts its friends, verifies identities and thus algorithmically increases trustworthiness of each respective peer.

I’m not sure what to make of this post more than express my belief that the web’s SSL structure using pre-loaded certificate authorities in software designed to handle your private communication is flawed. For my part, I’m trying to push people into using semi-WoT CAs like CAcert.org and encourage the development of new DNS models. But my guess is I’m not influental enough on my own. Will you join the p2p revolution too?