4 ms·

In a sense they are, but they send the same updates to everyone whereas a backdoor needs to allow malicious actions to be taken against specific targets only. I

by Canada 8y ago

In a sense they are, but they send the same updates to everyone whereas a backdoor needs to allow malicious actions to be taken against specific targets only. If vendors are forced to abuse updates in this way then users are going to demand a certificate transparency-like system to stop the abuse.

ENOTTY 8y ago
> they send the same updates to everyone Google's staggered updates to Chrome are updates sent to specific targeted users. The iOS App Store (and Google Play store, I believe) has an opt-in beta program that also sends betas to specific targeted users. > If vendors are forced to abuse updates in this way then users are going to demand a certificate transparency-like system to stop the abuse. But yes, leveraging users' trust in the update system is a risk that needs to be factored into the debate.
- Canada 8y ago
  Right, I'm not suggesting that it's even remotely possible for millions of systems to be updated at the same instant, just that outside of some special circumstances like beta channels (yes Google Play supports this) the huge numbers of devices are offered the same updates. If trust is lost in the update mechanisms then something will be done to restore it. That would likely be large developers like Apple, Microsoft and Google deploying a technical solution to prevent themselves from sending different updates to different devices without loudly warning the user and automatically reporting the attack to the world. That would please users and reduce compliance costs. Really, the major operating systems should just do it now to prevent this from becoming a problem in the first place.
AgentME 8y ago
>If vendors are forced to abuse updates in this way then users are going to demand a certificate transparency-like system to stop the abuse. I always thought this would be a neat use of Ethereum: a smart contract publicly stores the hash+url of the latest version of $softwareProduct (and maybe also the hash+url for the latest beta release, etc). The software would only update to a version named by the smart contract. The smart contract would only allow the application developers' addresses to be able to update the published hash+urls. The smart contract could have arbitrary logic here, like to allow the developers to add new addresses that can publish new versions, or to implement voting logic so a majority of the developers have to agree on publishing a new version, etc. In this system, it's not possible to secretly push a malicious update to one user.

by Canada 8y ago

ENOTTY 8y ago

> they send the same updates to everyone Google's staggered updates to Chrome are updates sent to specific targeted users. The iOS App Store (and Google Play store, I believe) has an opt-in beta program that also sends betas to specific targeted users. > If vendors are forced to abuse updates in this way then users are going to demand a certificate transparency-like system to stop the abuse. But yes, leveraging users' trust in the update system is a risk that needs to be factored into the debate.

Canada 8y ago
Right, I'm not suggesting that it's even remotely possible for millions of systems to be updated at the same instant, just that outside of some special circumstances like beta channels (yes Google Play supports this) the huge numbers of devices are offered the same updates. If trust is lost in the update mechanisms then something will be done to restore it. That would likely be large developers like Apple, Microsoft and Google deploying a technical solution to prevent themselves from sending different updates to different devices without loudly warning the user and automatically reporting the attack to the world. That would please users and reduce compliance costs. Really, the major operating systems should just do it now to prevent this from becoming a problem in the first place.

AgentME 8y ago

>If vendors are forced to abuse updates in this way then users are going to demand a certificate transparency-like system to stop the abuse. I always thought this would be a neat use of Ethereum: a smart contract publicly stores the hash+url of the latest version of $softwareProduct (and maybe also the hash+url for the latest beta release, etc). The software would only update to a version named by the smart contract. The smart contract would only allow the application developers' addresses to be able to update the published hash+urls. The smart contract could have arbitrary logic here, like to allow the developers to add new addresses that can publish new versions, or to implement voting logic so a majority of the developers have to agree on publishing a new version, etc. In this system, it's not possible to secretly push a malicious update to one user.