re: Hybrid Relay

Question

I'm pondering this scenario. I've developed a service that exposes a hybrid endpoint to the Service Bus. This service is accessed by clients that exist on the same network as the service, as well as by clients on the internet.

When the connection between the same network clients and the service moves to a direct coupling, will that traffic still be routed out to the internet and then back in? Or will communication between the client and server only happen within the intranet?

I realize that the bus will no longer be part of communications once a direct connection has been made. But I'd like to know if the traffic is going to be consuming internet badwidth or not. Also helps me determine if I need to plan for service supporting multiple endpoints (both an internal one and a service bus one).

Answer 1

Hello, when using the direct mode, the initial communication between the client and the service still goes through the Service Bus. Once the connection between each endpoint and the Service Bus is established, Service Bus will try to see if it is possible to create a direct route between the client and the service. If it succeeds, the messages will no longer be relayed to the Service Bus. But until such a route is created, all messages are still relayed to the cloud. So initially, internet traffic is unavoidable. But if the direct connection can be established, internet traffic will be significantely reduced.

---

Lante, shanaolanxing This posting is provided "AS IS" with no warranties, and confers no rights.

Answer 2

Right, I understood all that. But my specific question was that once the direct connection is established, if both the client and server reside within the same intranet... will the direct connection be going using internet resources? Or will all traffic remain within the intranet.

The reason I ask is that some time ago I worked on a distributed project. After a firewall change, we started having issues with connectivity between one of the clients and our service. Both these processes resided within our network and therefore the external facing firewall should have not been impacting them. However, we found that the client, when addressing the service, but being resolved to an externally facing IP and thus was going out to the internet and then coming back in through the firewall. This resulted in additional resources (both bandwidth and router/firewall cpu) being consumed for each transaction.

Hence my query about the behavior of once the direct connection has been established. :)

I'm ok if connection continues to use behave like the errant client in my example above. That's then something I can code both the application and service to handle via multiple endpoints. :)