3 ms·

What does "zero-trust" mean? I'd presume I want a network I can trust.

by toomim 4y ago

dovholuknf 4y ago
what a great, great question. it's a horrible term. really it means you have "no trust of the network, so you verify it before allowing connections"... The term is basically the OPPOSITE of what it is. You use a strong identity and you verify that cryptographically verifiable identity before allowing connections to your network, then on top of that you have a separate authorization tier for that identity to access services. it's a horrible term... :)
steveklabnik 4y ago
A perspective from someone who hasn't been super involved in network security: Historically, many folks would treat a VPN as something like a moat around the castle. If you're on the VPN, you can access stuff, and if you're not on the VPN, you cannot. However, this leads to a sort of ambient authority: break into the VPN, and you've broken into anything. "Zero Trust" is a model closer to capability-based security than to ambient authority: just because you're inside the VPN does not automatically mean that you have access to the servers inside of it, you must also authenticate with them, specifically. It's "zero trust" because no user is trusted by default, not because you as a user cannot trust the network.
- dovholuknf 4y ago
  If you're interested in learning more about it in video (still need to write the content up in 'blog/article' form) you can watch me blather about it at the beginning of the talk i gave to developer week... https://www.youtube.com/watch?v=kVW2SOETP90 https://www.youtube.com/watch?v=kVW2SOETP90 I actually use that castle/moat analogy and talk about how that's how vpn's work... it's a tried and true example. :)
- cyberge99 4y ago
  Thus, zk is akin to all doors within the castle being locked and only the appropriate persons can access their respective areas: chefs -> pantry; knights -> armory, etc.
  - PLG88 4y ago
    Aspects of it yes, that would be the principles of micro-segmentation and least privilege. There are other aspects of ZT principles, e.g., software-defined perimeters which OpenZiti delivers where this anology does not work. For example, SDP allows OpenZiti to do authentication/authorisation before any connectivity is established with outbound only connectivity as well as not even trusting the host OS if you use an SDK. I wrote a blog on this a few months back - https://netfoundry.io/demystifying-the-magic-of-zero-trust-with-my-daughter-and-opensource/ https://netfoundry.io/demystifying-the-magic-of-zero-trust-w...
0x457 4y ago
Zero Trust Network means a network without an implicit trust. Just because a device is present on a corporate office network doesn't mean it has access to all company resources. It's very common to have site-to-site connections and allow everything in-between or slightly limit things in between. Sometimes it even just allow-listing certain IPs like office public IP or employees home IPs. In a zero trust environment there are no such assumptions made, in fact, the only assumption that is made is that the network is already breached. No one is implicitly trusted, every access to a resource must authenticate, audit logs must be secured.

What does "zero-trust" mean? I'd presume I want a network I can trust.

by toomim 4y ago

What does "zero-trust" mean? I'd presume I want a network I can trust.

what a great, great question. it's a horrible term. really it means you have "no trust of the network, so you verify it before allowing connections"... The term is basically the OPPOSITE of what it is. You use a strong identity and you verify that cryptographically verifiable identity before allowing connections to your network, then on top of that you have a separate authorization tier for that identity to access services. it's a horrible term... :)

steveklabnik 4y ago

A perspective from someone who hasn't been super involved in network security: Historically, many folks would treat a VPN as something like a moat around the castle. If you're on the VPN, you can access stuff, and if you're not on the VPN, you cannot. However, this leads to a sort of ambient authority: break into the VPN, and you've broken into anything. "Zero Trust" is a model closer to capability-based security than to ambient authority: just because you're inside the VPN does not automatically mean that you have access to the servers inside of it, you must also authenticate with them, specifically. It's "zero trust" because no user is trusted by default, not because you as a user cannot trust the network.

dovholuknf 4y ago
If you're interested in learning more about it in video (still need to write the content up in 'blog/article' form) you can watch me blather about it at the beginning of the talk i gave to developer week... https://www.youtube.com/watch?v=kVW2SOETP90 https://www.youtube.com/watch?v=kVW2SOETP90 I actually use that castle/moat analogy and talk about how that's how vpn's work... it's a tried and true example. :)
cyberge99 4y ago
Thus, zk is akin to all doors within the castle being locked and only the appropriate persons can access their respective areas: chefs -> pantry; knights -> armory, etc.
- PLG88 4y ago
  Aspects of it yes, that would be the principles of micro-segmentation and least privilege. There are other aspects of ZT principles, e.g., software-defined perimeters which OpenZiti delivers where this anology does not work. For example, SDP allows OpenZiti to do authentication/authorisation before any connectivity is established with outbound only connectivity as well as not even trusting the host OS if you use an SDK. I wrote a blog on this a few months back - https://netfoundry.io/demystifying-the-magic-of-zero-trust-with-my-daughter-and-opensource/ https://netfoundry.io/demystifying-the-magic-of-zero-trust-w...

0x457 4y ago

Zero Trust Network means a network without an implicit trust. Just because a device is present on a corporate office network doesn't mean it has access to all company resources. It's very common to have site-to-site connections and allow everything in-between or slightly limit things in between. Sometimes it even just allow-listing certain IPs like office public IP or employees home IPs. In a zero trust environment there are no such assumptions made, in fact, the only assumption that is made is that the network is already breached. No one is implicitly trusted, every access to a resource must authenticate, audit logs must be secured.