4 ms·
HTTPS and SSL certificates are primarily used for encrypting the connection, thus protecting your data from snooping and modification in transit. They are not
by CydeWeys 9y ago
HTTPS and SSL certificates are primarily used for encrypting the connection, thus protecting your data from snooping and modification in transit. They are not actually that useful for authenticating identity. Users tend to simply ignore the padlock icon, which is why Chrome is moving away from an affirmative security display model (which users don't pay much attention to) to displaying prominent warnings for insecure or hijacked connections, which users do pay more attention to. SSL certificates' inability to properly authenticate identity is also why many of the largest tech companies in the world do not even bother with EV certificates (e.g. Google, Amazon, and Facebook for starters).
In order to fully establish identity and provide strong authentication, you would need a much stronger centralized system. Additionally, browsers would need to refuse to make connections to endpoints not authenticated through this system, otherwise you'd have the same kind of lackadaisical user response as we see now to the padlock icon. I don't think anyone present here wants anything like that, and I'm sure that if we were proposing such a walled garden vision for the Internet, the responses would rightfully be much more negative. Instead, we just want everyone's connections to be encrypted.
As for some of your other points, having a strong, short, memorable domain name is good for overall Web security and does help cut down on fraudulent apps and the other problems you've identified. To give you one example, I recently saw an ad for an app called "Curb" (it's a yellow tax ride-hailing app). Out of curiosity I wanted to see more information about it, but there was no domain name on the app; it simply said "Download the Curb app". So I searched the app store for "Curb" and there were dozens of apps with that name, many of them clearly imitators. The best "security" measure was trying to remember what the app's icon looked like.
Now imagine that that ad had instead said "Go to curb.app to get our app", and that said site had prominent links to mobile app stores. That's taking advantage of the global uniqueness property of domain names to guarantee that I won't be tricked away by a scammer. "curb.app" is short and memorable, gets users to the right place, and because .app is a fresh namespace, is still available. Good luck getting any domain name close to that good on .com. .com is fully mined out, and any good names that are available are being sold for minimums of tens of thousands of dollars by squatters and resellers.
I'll be talking about these issues and more in depth at I/O, if you'd like to watch/stream it: https://events.google.com/io/schedule/?section=may-8&sid=7ac8be59-1ecc-4d90-aea5-e62deab5ee7f https://events.google.com/io/schedule/?section=may-8&sid=7ac...
- spronkey 9y agoConsidering your "Go to curb.app" solution, to be brutally honest I think at least 50% of the world has no idea what the difference between "Go to curb.app" and "Download the Curb app" is. They are both practically the same. "Guarantee that I won't be tricked away by a scammer" - yeah, no. There are already too many TLDs to squat on that remove any usefulness of a "short and memorable name" on a new one. I feel like there's some sort of legitimate blockchain solution to be had here that doesn't involve a centralised authority. Who got there first, what domains they have owned continuously, and what products they control perhaps...
- danShumway 9y ago> Now imagine that that ad had instead said "Go to curb.app to get our app", and that said site had prominent links to mobile app stores. Now imagine it's a year or two from now and you have the exact same problems that you had with curb.com. What's the strategy for when .app gets mined out? Are we just going to keep creating new tlds over and over for eternity? Bear in mind that the only way .app won't get mined out will be if either very few people use it, or if users ignore it and don't treat it any differently than any other namespace. Ironically, the only way this will be useful for developers like me is if Google's advertising campaign utterly fails and it's ignored by most people. The solution to identity management can't be that users will stay one tld ahead of hackers. That's just never gonna happen. Users are slower than hackers.
- cornholio 9y ago> Are we just going to keep creating new tlds over and over for eternity? Yes. The "easily memorable global namespace" is a limited natural resource, there is no definitive fix and there can't be - unless you accept the totalitarian single gatekeeper scheme Cyde describes. In an open system, you always need to deal with the Sybil attack and the only solution is proof of work, resources - money. This means domain squatters will always capture a rent, by guessing domain names that will be valuable latter on. The way to minimize that rent on the long run is to periodically reset the game by bringing new TLDs into existence, thereby devaluing the investment in older TLDs, thereby making squatting a more risky activity. The GTLD liberalization greatly diminished the price pressure on .com .org and the like. Keep in mind the web has seen exponential increase in the last decades; the upside now is much more limited, maybe another order of magnitude until complete worldwide market saturation and web growth figures get more inline to global economic growth numbers. Once that is achieved the problem becomes much less serious.