Closed cyrill-k closed 3 years ago
I agree with your reasoning regarding the "endpoint transparency" - let's keep the definition as-is.
So I think this PR can be merged after changing the firewall example. :slightly_smiling_face:
I changed the firewall example. If you agree with the changes, you can merge the PR.
Looks good to me now, so I'm merging it.
I guess that is true, however even in such a case, the relevant piece of meta-information is how to identify the flow and associate it to host, with which the node communicated to perform a specific action. And the identification relies on the existence and content of protocol headers and payloads, which we cover with our definition. So, for example to model RFC8558 style interactions between hosts and on-path nodes using our path terminology, we could say that the explicit communication with a host changes the transparency property of the node.
I would suggest to keep the definition as it is right now and keep the definition of meta-information open-ended.