Hi rufi,
Your understanding of traffic flow is correct. The traffic from on-premises does indeed pass through an edge router and is sent to the ExpressRoute circuit. It then enters Azure via Tenant A’s ExpressRoute circuit and reaches Tenant A’s ExpressRoute gateway.
On-Premises
|
| (Edge Router)
|
ExpressRoute POP
| |
| |
Tenant A Tenant B
(ERCircuit) (ERCircuit)
| |
| |
Azure Azure
(Tenant A) (Tenant B)
Therefore, the traffic coming from on-premises, passing through an edge router, and being sent to the ExpressRoute circuit in Tenant A. The traffic then enters Azure via Tenant A’s ExpressRoute circuit and reaches Tenant A’s ExpressRoute gateway. After that, the traffic is routed back to the ExpressRoute Point of Presence (POP), before returning back to Azure and reaching Tenant B.
However, it’s important to note that traffic going between your two tenants does not go direct. It will route all the way back to your ExpressRoute Point of Presence (POP), before returning back to Azure. This might not be ideal if latency is important, depending on where your POP is located
References:
- https://video2.skills-academy.com/en-us/azure/expressroute/expressroute-faqs
- https://video2.skills-academy.com/en-us/answers/questions/1394753/sharing-azure-expressroute-connection-across-multi
Let me know if this solve your doubts,
Luis