BGP Unnumbered EVPN Fabric
Use Case |
Use the unnumbered BGP automatic peering feature to quickly deploy an IPv6 underlay for an EVPN-VXLAN data center (DC) fabric. |
Audience |
Network administrator, engineer, operators, and architects who want to understand how to deploy a BGP unnumbered Clos fabric underlay for a DC fabric. |
Knowledge Level |
General familiarity with EVPN-VXLAN data center network architectures and underlay and overlay routing. See the related topics section for background information about EVPN-VXLAN technology. |
Benefits |
|
Products Used |
For the full list of supported devices and OS versions, see Supported Juniper Devices. |
This guide demonstrates how to deploy and verify the BGP unnumbered peering (also referred to BGP auto discovery or BGP auto-peering). Juniper Networks supports BGP unnumbered peering starting in Junos OS Release 21.1R1. This feature allows BGP to auto-discover and to create peer neighbor sessions using the link-local IPv6 addresses of directly connected neighbors.
The BGP unnumbered peering solution uses Junos OS support for the following RFC's:
Overview
Today many enterprises and hyper-scale data centers use BGP as the underlay routing protocol. Unlike traditional IGPs such as OSPF and IS-IS, BGP typically requires that you explicitly configure peering, autonomous system (AS) numbers, and routing policies to control route exchanges.
Many operators are still relatively inexperienced with IPv6. Using BGP unnumbered peering, which dynamically discovers IPV6 neighbors, reduces the burden of manually configuring an IPv6 underlay in an EVPN-VXLAN DC fabric. Junos OS builds on the baseline IPv6 functionality by supporting BGP group configuration. BGP group configuration provides support for dynamic peering parameters (such as allowed remote AS numbers) used to support an unnumbered IPv6 fabric.
Use Case: Manually Configuring an EVPN-VXLAN Fabric
You can configure an EVPN-VXLAN DC fabric by manually configuring the fabric or by using the BGP unnumbered peering feature. This use case describes the complexity of manually configuration the fabric and why using the BGP unnumbered peering feature provides a much easier solution.
Consider a simple two-tier data center. This modest size fabric consists of four spine devices and 32 leaf devices. Each spine device has 32 links which attach to the leaf devices and each leaf has two fabric links, one for each spine device.
In this manual configuration, you first need to need to assign the IP addresses for the network. For this fabric you'll need to configure 4*32=128 IPv6 IP addresses. Each network requires two host address assignments.
Next, you configure the BGP peers and their associated AS numbers. For each end of every fabric link, you need one BGP peering session. In our example fabric, this calculation equates to a total of 4*32*2=256 BGP peer definitions, each of which requires a unique peering IP and remote AS number.
Manually defining 256 BGP peerings can be cumbersome is also prone to error. In a complex fabric, a simple misconfiguration can be difficult to isolate. Let's say that the fabric supports 128 leaf devices. You must now configure 4*128=512 IP IPv6 networks. It is clear from the math that the complexity of manually provisioning a large fabric quickly becomes a burden. Also, for IPv4 fabrics, an often overlooked factor is the large number of IPv4 addresses consumed by the underlay. In many networks, IPv4 addressing space is at a premium.
In contrast, BGP unnumbered peering requires no routable IP network assignments on underlay links. All BGP peering in the underlay uses only link-local IP's. Using link-local IP's means less configuration, less complexity, smaller routing tables, and IP address preservation.
Table 1 shows the configuration required for a simple two spine and two leaf EVPN-VXLAN fabric. Specifically, it compares the configuration of two underlay EBGP peers on a leaf device using manual peering versus unnumbered peering.
When comparing the configurations, consider not only the number of configuration statements required but also their relative complexity. As show in the table below, the manual configuration (IPv4) requires that you configuring the IP addresses, the remote peer IP address, and AS numbers for the remote peers. In contrast, with BGP unnumbered peering, you only need to define the interface names. No routable IP address assignments are required in the underlay. In addition, BGP unnumbered peering automatically configures the BGP neighbor IP and the remote AS number.
Manual Configuration (IPv4) | BGP Unnumbered Peering |
---|---|
set interfaces xe-0/0/0 unit 0 family inet address 10.0.1.2/30 set interfaces xe-0/0/1 unit 0 family inet address 10.0.1.6/30 set protocols bgp group manual_underlay family inet unicast set protocols bgp group manual_underlay local-as 65510 set protocols bgp group manual_underlay neighbor 10.0.1.1 peer-as 65001 set protocols bgp group manual_underlay neighbor 10.0.1.5 peer-as 65002 |
set interfaces xe-0/0/0 unit 0 family inet6 set interfaces xe-0/0/1 unit 0 family inet6 set policy-options as-list as-list members [65000-6510] set protocols bgp group bgp_unnum family inet6 unicast set protocols bgp group bgp_unnum local-as 65003 set protocols bgp group bgp_unnum dynamic-neighbor FABRIC peer-bgp_unnumovery family inet6 ipv6-nd set protocols bgp group bgp_unnum dynamic-neighbor FABRIC peer-bgp_unnumovery interface xe-0/0/0 set protocols bgp group bgp_unnum dynamic-neighbor FABRIC peer-bgp_unnumovery interface xe-0/0/1 set protocols bgp group bgp_unnum peer-as-list as-list |
Disadvantages
|
Advantages
|
You can easily use a configuration group to apply the inet6
family.
Using configuration groups greatly reduces the time it takes to configure a large fabric.
We show you a sample of a configuration group later in this NCE.
Next, we'll show you how to configure a BGP unnumbered peering on an VXLAN-EVPN fabric.