- Introduction to the Scalable and Modular Control Plane on the ASR 1000
- NSF/SSO, NSR, Graceful Restart to Ensure Robust Routing
- Packet Capture Using Encapsulated Remote SPAN
- Achieving Segmentation Using MPLS over GRE and MPLS VPNs over GRE Solutions
- Scalable v4/VPNv4 Route Reflector
- Scalable and Flexible Internet Edge
- Scalable Data Center Interconnect
- Summary
- Chapter Review Questions
- Answers
- Further Reading
This chapter is from the book
This chapter is from the book
This chapter is from the book
Scalable v4/VPNv4 Route Reflector
With the growing adoption of MPLS in the enterprises to achieve large-scale virtualization and segmentation, there is also a need for enterprises to have their own route reflector (RR) for VPNv4 routes, deployed separately or combined in a PE router. An RR simplifies the iBGP full-mesh restriction where all PEs don't have to mesh with all other PEs, rather just with the RR.
Use Case: Route Reflection
Figure 12-6 shows the RR used by the enterprise in the self-managed MPLS clouds.
)
Figure 12-6 MAN using the same router for E-PE and VPNv4 RR roles.
To meet this requirement of avoiding the full mesh of iBGP, you need to configure the Cisco ASR 1000 as the RR for VPNv4 routes using the following steps:
- Step 1. Configure RRs to peer with PEs to reflect VPNv4 routing information learned from other PEs:
ASR1004-RR(config)# router bgp 100 ASR1004-RR(config-router)# neighbor A-PE peer-group ASR1004-RR(config)# neighbor A-PE remote-as 100 ASR1004-RR(config)# neighbor A-PE update-source Loopback100 ASR1004-RR(config)# neighbor PE loopback# peer-group A-PE - Step 2. Configure RRs for VPNv4 BGP peering between PEs and RRs:
ASR1004-RR(config-router)# address-family vpnv4 ASR1004-RR(config-router-af)# neighbor A-PE activate ASR1004-RR(config-router-af)# neighbor A-PE route-reflector-client ASR1004-RR(config-router-af)# neighbor A-PE send-community extended ASR1004-RR(config-router-af)# neighbor PE loopback# peer-group A-PE - Step 3. Configure the PE for VPNv4 BGP peering between PEs and RRs (thus enabling PEs to exchange VPNv4 routing information with the RRs):
ASR1004-PE(config)# router bgp 100 ASR1004-PE(config-router)# no synchronization ASR1004-PE(config-router)# bgp log-neighbor-changes ASR1004-PE(config-router)# neighbor ASR1004-RR loopback ip# remote-as 100 ASR1004-PE(config-router)# neighbor ASR1004-RR loopback ip# update- source Loopback0 ASR1004-PE(config-router)# address-family vpnv4 ASR1004-PE(config-router-af)# neighbor ASR1004-RR loopback ip# activate ASR1004-PE(config-router-af)# neighbor 172.16.1.1 send-community extended
Although this example uses the Cisco ASR 1004 as the VPNv4 RR, this is applicable to the IPv4 RR, too. The VPNv4 route scale is completely a function of the ASR1000-RP you have in the system. With the ASR1000-RP1 and ASR1000-RP2, the scale is up to 1M and 4M, respectively, for IPv4. For VPNv4 routes, ESP does not have to be in the data path, and therefore any ESP can be used. Currently for IPv4, FIB entries are still populated, hence limiting the RR scale. This will change in a future IOS XE version.
The Cisco ASR 1000, by virtue of the ASR1000-RP1 and ASR1000-RP2, provides the largest scale for Route Reflector deployments in the Cisco midrange routing portfolio. The ASR1000-RP2, with 16-GB DRAM, truly raises the bar, with 64-bit IOS XE that allows the routes to scale up to 20M, which essentially rivals even the largest core routers available today.
In general, the ASR1000-RP2 (16-GB DRAM) provides four times the route scale over RP1 (4-GB DRAM), three times the number of peers/sessions (with the given convergence time) and is at least twice as fast in terms of route convergence (for the given set of routes and peers).