Is there any binding (server or client) that will expose the client IP address to the server? If so, what are those bindings, and what bindings are purely safe from this perspective.
For this question I'm trying to help a client use the service bus as a firewall and only allow connections from "known IPs".
I'm also trying to help another client that needs to help reporters where freedom of speech is repressed, maintain an anonymous connection.
We don't give listeners the sender's IP address at this point. If we were to introduce block-lists, we would likely do so at the service edge, meaning that we would enforce them inside Service Bus without getting the listeners involved. We don't support user-settable per endpoint block-lists today, but we might add support in a future release.
For the second scenario I can't give any assurance that we will never disclose the sender's IP address to the listener via the Relay or even Messaging because we will introduce statistics and logging capabilities down the road. If you need complete insulation where the client must never be disclosed to the listener under any circumstances, I would suggest that you take the in ingress traffic into a Windows Azure Web Role and forward messages from there to the Relay target.
- Yanıt Olarak Öneren clemensv 06 Mart 2012 Salı 17:39
Thank you; even though the ServiceBus doesn't actually forward or relay the message, is there anything within the WCF client library that will transmit/embed the client IP in the XML (or otherwise) payload.
Pardon me if this is what you meant; but I want to ensure 100% clarity.