Zoom will not add end to end encryption anytime soon, that would defeat their business model as a data stealing mechanism for the Chinese government. You thought corona was bad, wait until all US business fold in the next decade. Fact.
Zoom used to let me use my Canon as the cam now since this update it doesn't allow it. The second problem is the app crashes out of nowhere PLEASE FIX IT. I desperately need this for work and my kids for school. I don't want to have to switch to something else when it work fine before. just add it back and fit the crashing.PLEASE!!!
Upping the security isn't that much of an improvement if it isn't end-to-end. The whole architecture of Zoom is basically a man-in-the-middle vulnerability.
I keep seeing people decry its lack of end-to-end encryption. Their initial instance that they provide it was stupid, as was how long they held on to that claim before eventually dropping it, but beyond that I do not understand the shortcoming. There is no video conference service that offers end-to-end encryption at scale for large, multipoint sessions. How could that possibly work? That's a genuine question, not rhetorical. I cannot fathom how multiple video sessions could be combined into a single session without a central server that decrypts the individual sessions, combines them, and then sends the combined stream to each percipient. The alternative would be fully meshed connections of each endpoint to all the others but that can't scale out.
Well, iMessage does it with group chats; PGP can do it, too. It's quite possible to encrypt data to multiple recipients. Every endpoint shares its public key with a central server, which then distributes them to all other participants. No one (including the central server) has access to any others' private keys.
Upping the security isn't that much of an improvement if it isn't end-to-end. The whole architecture of Zoom is basically a man-in-the-middle vulnerability.
I keep seeing people decry its lack of end-to-end encryption. Their initial instance that they provide it was stupid, as was how long they held on to that claim before eventually dropping it, but beyond that I do not understand the shortcoming. There is no video conference service that offers end-to-end encryption at scale for large, multipoint sessions. How could that possibly work? That's a genuine question, not rhetorical. I cannot fathom how multiple video sessions could be combined into a single session without a central server that decrypts the individual sessions, combines them, and then sends the combined stream to each percipient. The alternative would be fully meshed connections of each endpoint to all the others but that can't scale out.
Well, iMessage does it with group chats; PGP can do it, too. It's quite possible to encrypt data to multiple recipients. Every endpoint shares its public key with a central server, which then distributes them to all other participants. No one (including the central server) has access to any others' private keys.
Key distribution among a large number of users isn’t what I’m talking about; that’s a solved problem. Messaging apps can use this to do end-to-end encryption because the nature of those applications lets a central server simply redistribute messages to each group member without having to decrypt them.
Large multipoint videoconference sessions are different: Every participant’s outbound network stream has to be combined with every other and then sent to each participant as a single, combined stream. I don’t know of any way that can happen without a central server being able to decrypt everyone’s video. If you’re aware of some way of doing this without decryption in the middle, I’m genuinely interested. The only alternative that I can think of is for each user to sent their video to every other user, and in turn receive every other user’s video in separate streams, and there is no way that can work at scale.
Upping the security isn't that much of an improvement if it isn't end-to-end. The whole architecture of Zoom is basically a man-in-the-middle vulnerability.
I keep seeing people decry its lack of end-to-end encryption. Their initial instance that they provide it was stupid, as was how long they held on to that claim before eventually dropping it, but beyond that I do not understand the shortcoming. There is no video conference service that offers end-to-end encryption at scale for large, multipoint sessions. How could that possibly work? That's a genuine question, not rhetorical. I cannot fathom how multiple video sessions could be combined into a single session without a central server that decrypts the individual sessions, combines them, and then sends the combined stream to each percipient. The alternative would be fully meshed connections of each endpoint to all the others but that can't scale out.
Well, iMessage does it with group chats; PGP can do it, too. It's quite possible to encrypt data to multiple recipients. Every endpoint shares its public key with a central server, which then distributes them to all other participants. No one (including the central server) has access to any others' private keys.
Key distribution among a group of users is a solved problem and not the issue that I'm talking about. With low-bandwidth applications like messaging it's entirely feasible to send messages directly to everyone in the group, even for large numbers of users. Even if you have a central server acting as a broker, it only has to relay the messages to each member of the group and can do so without having to process the encrypted contents of the messages.
Shared video conferences are different. Each participant's transmitted video stream has to be combined with all of the others' and then sent as a combined stream to all of the participants. Without being able to decrypt the streams, a central server has no way that I'm aware of to combine the streams. If you know of some way, I'm genuinely curious as to how that could possibly work without decryption in the middle. The only alternative that I can see is for every participant to send their video stream to every other participant, and in turn, to receive one from every other participant. That cannot scale.
To be clear, I'm talking about multipoint video conference sessions at scale, not small groups or person to person. There is no good reason that person to person can't be end-to-end encrypted.
Key distribution among a group of users is a solved problem and not the issue that I'm talking about. With low-bandwidth applications like messaging it's entirely feasible to send messages directly to everyone in the group, even for large numbers of users. Even if you have a central server acting as a broker, it only has to relay the messages to each member of the group and can do so without having to process the encrypted contents of the messages.
Shared video conferences are different. Each participant's transmitted video stream has to be combined with all of the others' and then sent as a combined stream to all of the participants. Without being able to decrypt the streams, a central server has no way that I'm aware of to combine the streams. If you know of some way, I'm genuinely curious as to how that could possibly work without decryption in the middle. The only alternative that I can see is for every participant to send their video stream to every other participant, and in turn, to receive one from every other participant. That cannot scale.
To be clear, I'm talking about multipoint video conference sessions at scale, not small groups or person to person. There is no good reason that person to person can't be end-to-end encrypted.
Zoom doesn't need to combine the video feeds, and I highly doubt that they do. All it needs to do is relay each incoming video feed to other participants, then the client side plays each stream separately. Otherwise, how would the client software so easily re-arrange the video windows as different people speak? There's no decryption required for a relay.
Now that said, there are good reasons you might want to modify video feeds for each participant, say, for participants with low downstream bandwidth vs higher bandwidth. And they may do that for all I know, but such a thing isn't inherently necessary. You could have participants send high- and low-bandwidth versions of their streams (upstream bandwidth allowing) so that the relay could just choose one or the other (without decryption) to send to high- or low- downstream bandwidth participants, without doing any decryption or stream modification, at the expense of participants' upstream bandwidth who had it to spare. Low-upstream participants could send the low-bandwidth version only.
Anyway, the main issue here isn't so much that Zoom didn't have end-to-end encryption - they're the only one that made such a claim, afaik. It's that they lied about having it - and on top of a host of other fairly egregious security faux pas.
Comments
Well, iMessage does it with group chats; PGP can do it, too. It's quite possible to encrypt data to multiple recipients. Every endpoint shares its public key with a central server, which then distributes them to all other participants. No one (including the central server) has access to any others' private keys.
Key distribution among a large number of users isn’t what I’m talking about; that’s a solved problem. Messaging apps can use this to do end-to-end encryption because the nature of those applications lets a central server simply redistribute messages to each group member without having to decrypt them.
Large multipoint videoconference sessions are different: Every participant’s outbound network stream has to be combined with every other and then sent to each participant as a single, combined stream. I don’t know of any way that can happen without a central server being able to decrypt everyone’s video. If you’re aware of some way of doing this without decryption in the middle, I’m genuinely interested. The only alternative that I can think of is for each user to sent their video to every other user, and in turn receive every other user’s video in separate streams, and there is no way that can work at scale.
To be clear, I'm talking about multipoint video conference sessions at scale, not small groups or person to person. There is no good reason that person to person can't be end-to-end encrypted.