5 ms·
Matrix and XMPP, for example, provide methods for end-to-end encryption. The operator of the node has no way to see the content of the messages. Deriving the co
by rglullis 7y ago
Matrix and XMPP, for example, provide methods for end-to-end encryption. The operator of the node has no way to see the content of the messages. Deriving the connection graph is not an issue exclusive to federated systems.
But most important, no one is ever bound to any federation system. This by itself should be strong deterrent of bad behavior by bad actors. If I don't trust one server I can always switch. In an extreme case of lack of trust I can set up a server just for myself and the federation devolves into a peer-to-peer system.
Either I am missing the point of the article or the author hasn't thought this through.
- fhennig 7y agoThe point is not the content but the connection graph as you mentioned > Deriving the connection graph is not an issue exclusive to federated systems. No it is not, but with a single large entity there is only a single entity that you have to trust. If facebook doesn't look at or leak the metadata then you're save. In a federated system there are so many people operating the system that it is much more likely that parts of your communication metadata gets exposed.
- rglullis 7y ago> If facebook doesn't look at or leak the metadata then you're save. Two words for you: Cambridge Analytica. The connection graph is something that is always going to be exploited at the big companies. It is not a matter of if or when, just how.
- fhennig 7y agoI was playing devils advocate here. Facebook has shown that they are not very trustworthy. But in theory it is easier to contain the data if it is only held in a single central location vs. distributed over many federated servers.
- rglullis 7y ago"Easier to contain" in this case does not mean "more likely to be safe". There is no incentive whatsoever for any large centralized body to not exploit the data. In any case, centralization is fragile. Even if there were some altruistic entity willing to run these services and not exploit it, it would still be a systemic risk.
- johnchristopher 7y agoMatrix has e2e but it's a pain in the ass to use (every time you log on a different browser of from a different client you have to import keys and validate clients). So most hub I have seen don't enforce users to use e2e and at least images are stored in the clear on the matrix server.
- rglullis 7y agoXMPP is still somewhat similar. But in any case it is an issue with the existing UX, not inherent to federated systems. I am not claiming it is easier. I am not even claiming it is easier than in the current centralized solutions. I am just saying that it is possible and that we don't need to throw the baby with the bathwater. It is much easier to improve on the existing federated solutions than trying to come up with some idealistic, absolutely-private decentralized technology.
- MayeulC 7y agoThat's true, but cross-signing is right around the corner for Matrix [1], which will enable e2e by default in direct chats [2]. Encryption is useless for most public rooms, and actually counter-productive most of the time. If the room is public, anybody can join it, idle there and save the logs on their computer, or publish them. Activating encryption is just annoying: from the top of my head, you can't read what was posted before you joined (unless someone sends you their keys), search is client-side only, it's hard to link to attachments or reuse mxc: (matrix content such as images) from one channel to another. [1]: https://github.com/vector-im/riot-web/issues/2714 https://github.com/vector-im/riot-web/issues/2714 and the two links below [1]: https://github.com/vector-im/riot-web/issues/10605 https://github.com/vector-im/riot-web/issues/10605 and associated labels [1]: https://github.com/vector-im/riot-web/issues?q=is%3Aissue+is%3Aopen+cross+sign+label%3Across-signing-sprint https://github.com/vector-im/riot-web/issues?q=is%3Aissue+is... [2]: https://matrix.org/blog/2018/11/02/user-experience-preview-end-to-end-encryption-by-default/ https://matrix.org/blog/2018/11/02/user-experience-preview-e... [2]: https://github.com/vector-im/riot-web/issues/6779 https://github.com/vector-im/riot-web/issues/6779
- johnchristopher 7y ago