Neither you have to configure this transit logical switch nor configure IP addressing on the link (unless, user wants to change the subnet range) nor configure any routing between DR and SR. This link defaults to an IP address in 169.254.0.0/28 subnet. An internal link called Intra-Tier Transit link is auto created using a Transit Logical switch between DR and SR. When we enable a centralized service on that logical router, a SR is instantiated. As shown in the diagram below, traffic from a VM hosted on a compute hypervisor goes through the Edge node on a overlay network, to connect to a device in physical infrastructure.Р”Р‡igure 1: Physical Topology showing Compute hypervisors and Edge NodeР”В«efore we get into a detailed packet walk for this N-S traffic, it is imperative to understand the architecture details as to how the two routing components (DR and SR) are connected.Р”Р„s soon as a logical router is configured via NSX-T manager, a DR is instantiated in all transport nodes. These Edge nodes are also configured as Transport nodes and are assigned a TEP (Tunnel End Point) IP just like compute hypervisors to send/receive overlay traffic to/from compute hypervisors. To provide connectivity to the physical infrastructure, SR has been instantiated on the Edge nodes. As shown in the diagram, Distributed Router (DR) component of a Logical router is instantiated on all the transport nodes (Compute hypervisors and Edge nodes). The following diagram shows a typical leaf and spine topology with Edge nodes providing connectivity to the physical infrastructure. Edge nodes (all VM or all Baremetal) can be logically grouped into an Edge cluster to provide scale out, redundant, and high-throughput gateway functionality for logical networks. Similarly, when a centralized service like NAT is configured on logical router, a SR or service instance for that particular logical router is instantiated on the Edge node.

So, when a logical router is connected to physical infrastructure, a SR is instantiated on the edge node. They are available in two form factors: Bare Metal or VM(both leveraging Linux FoundationРІР‚в„ўs DPDK Technology). These appliances where the centralized services or РІР‚СљSR InstancesРІР‚Сњ are hosted are called Edge nodes. We need a centralized pool of capacity to run these services in a highly-available and scale-up fashion.

NSX-T achieves best of both worlds with DR/SR approach, distributed routing (in kernel) for E-W traffic, centralized routing for N-S traffic and other centralized services.

To avoid exponential growth of BGP peerings from each hypervisor and reduce complexity of control plane, a dedicated routing component (Service Router) is designed to serve the need. In a datacenter leaf and spine topology or any other datacenter topology, there are designated devices that peer with WAN routers to exchange routes in BGP and provide N-S connectivity. Connectivity to physical topology is intended to exchange routing information from NSX domain to external networks (DC, Campus or WAN). LetРІР‚в„ўs take a look at one of these services (connectivity to physical devices) and see why a centralized routing component makes sense for running this service.

There are some services today on NSX-T which are not distributed such as:Р”РЋ) Connectivity to physical infrastructure Whenever a service which cannot be distributed is enabled on a Logical Router, a Service Router (SR) is instantiated. Before we get into the N-S routing or packet walk, letРІР‚в„ўs define Service Router. This N-S routing is provided by the centralized component of logical router, also known as Service Router. In this post, I will explain how North-South (N-S) routing is done in NSX-T and we will also look at the ECMP topologies. In the first part of this blog series, NSX-T: Routing where you need it (Part 1), I discussed how East-West (E-W) routing is completely distributed on NSX-T and how routing is done by the Distributed Router (DR) running as a kernel module in each hypervisor.

Vmware mac address showing on router