VMworld Sessions – Reference Design for SDDC with NSX and vSphere – Routing Protocol and Topology

This third section follows Reference Design for SDDC with NSX and vSphere – NSX Components, vCenter Topology, Connectivity Considerations and Edge Cluster Design

 

Routing Protocol and Topology

Enterprise vs Datacenter Routing
Don’t run multiple protocols
Multi-tenancy indicates use of BPG

Physical Core and Aggregation routing should be handled by the Enterprise team responsible for backbone Network.
Protocol choices may differ in each datacenter, with OSPF and BGP prevalent in large scale design.

Datacenter Routing Design and NSX

NSX routing domain is connected as Edge routing, and does not act as a transport network
Edge routing is sufficient in most cases
Use one protocol end-to-end between ESG – physical  network and DLR – ESG
Adopt established routing practices such as route summarisation.
Treat NSX as another autonomous system or as a stub network.

nsx-reference-architecture-b-copy-of-ecmp-with-dlr-control-vm-2

NSX Connectivity with BGP

BGP connectivity is preferred.
Multi-tenancy, compatible with any north bound protocol, better route control.
EBGP for Edge to physical network.
iBGP for DLR to ESG.

nsx-connectivity-with-bgp

North-South Traffic

Edge announces summary static network for NSX logical space to physical routers via EBGP
Necessary for ECMP control VM failure recovery
Not required in ESG HA model
DLR control VM announces routes via iBGP

South-North Traffic

Physical ToR send default routes, AND edge needs to allow redistribution of subnet of links connected to ToR to carry next hop into iBGP
Private AS can be used, additional configuration required to remove AS path being advertise into public BGP peering.

nsx-connectivity-with-bgp-2

 

 

This concludes part 1 of Reference Design for SDDC with NSX and vSphere
Session playbacks are available here search for NET7857

Leave a Reply

Your email address will not be published. Required fields are marked *