MP BGP Multiprotocol BGP
Formally referred to as MP BGP-4 (Multiprotocol BGP version 4), Multiprotocol BGP (MP-BGP) is a crucial addition to the conventional BGP-4 protocol. Its main purpose is to enable the simultaneous transmission of routing data for many network tiers and address families via a single BGP session. This multiprotocol capability was made possible by the original high extensibility of the underlying BGP version 4 protocol.
For contemporary networks that need to support several protocols, such as IPv6, multicast, and MPLS VPNs, MP-BGP is essential. Since a BGP-4 router is made to disregard communications it does not comprehend, MP-BGP is compatible with normal BGP-4.
You can also read What is Metro Ethernet Network? how does metro ethernet work
Core Mechanism and Components
The core innovation of MP-BGP is the use of identifiers and new path attributes to manage diverse routing information:
- Address Families (AFI and SAFI): To precisely define the type of routing information being transferred, MP-BGP employs two identifiers:
- Address Family Identifier (AFI): The network layer protocol is specified by the Address Family Identifier (AFI). For instance, IPv4 is denoted by AFI=1 and IPv6 by AFI=2.
- Subsequent Address Family Identifier (SAFI): The address family’s application or purpose is indicated by the Subsequent Address Family Identifier (SAFI). SAFI=1 for Unicast, SAFI=2 for Multicast, and SAFI=128 for VPN-IPv4 are a few examples.
- Only the IPv4 Unicast combination (AFI=1, SAFI=1) is supported by standard BGP. Multiple AFIs and SAFIs can be carried by a single BGP peering session MP-BGP.
- Separate Routing Tables: To keep routing data for various protocols segregated and manageable with varied policies, MP-BGP has separate routing tables for every address family. The AFI and SAFI are used by a BGP router to decide which routing table to put the route into when it receives an update.
- New Path Attributes: To transfer this varied routing data, MP-BGP adds two non-transitive optional path attributes:
- MP_REACH_NLRI (Multiprotocol Reachable Network Layer Reachability Information): Used to advertise reachable routes (prefixes) and the associated next-hop address for a specific address family.
- MP_UNREACH_NLRI (Multiprotocol Unreachable Network Layer Reachability Information): Used to explicitly withdraw unreachable routes for a specific address family.
You can also read What is Message Switching & advantages of message switching
Key Uses and Benefits
MP-BGP extends BGP beyond IPv4 unicast by supporting a large number of protocols, such as:
- IPv4 multicast
- IPv6 unicast
- IPv6 multicast
- Layer 2/Layer 3 VPNs
One of the main advantages is the capacity to handle and exchange different address families over a single BGP peer session.
IPv6 Routing
The particular improvement that gave BGP version 4 IPv6 capability is called MP BGP-4. It is the IPv6 Exterior Gateway Protocol (EGP) that is supported.
- It enables service providers to use their current infrastructure for IPv6 routing (using IPv6 Unicast AFI=2, SAFI=1).
- MP-BGP employs SAFI 2 messages for IPv6 multicast routing (SAFI 1 messages show routes useful solely for IP unicast).
MPLS VPN Services (Layer 3 VPNs)
For service providers in particular, this is the most important and prevalent use case for MPBGP.
- Provider Edge (PE) routers rely on MPBGP to perform the essential function of promoting client routes throughout the MPLS network.
- Route separation is the most important advantage in this setting. Many clients share a private IP network space (such as 10.0.0.0/8); MPBGP ensures that these routes are conceptually distinct, avoiding routing problems brought on by subnet overlap.
- By exchanging VPN-IPv4 routes (AFI=IPv4, SAFI=128), MPBGP accomplishes this separation. The VPN label, a special 96-bit address that consists of a 64-bit Route Distinguisher (RD) and a typical 32-bit IPv4 prefix, is used to carry the routes.
PE routers frequently employ route redistribution in Layer 3 MPLS VPN configurations to inject routes learnt from a customer’s Interior Gateway Protocol (IGP) (such as OSPF) into MPBGP, which then sends them over the service provider’s backbone.
Configuration Notes
When configuring BGP, neighbors are initially defined globally. By default, BGP peers only exchange IPv4 unicast address prefixes. To enable other address families, such as IPv6 unicast or multicast, neighbors must be explicitly activated within the respective address-family configuration block using the neighbor activate command.
To facilitate complex policies for different services, MP-BGP allows the application of route maps specifically within the context of an address family (e.g., address-family ipv6), simplifying the management of policies for each family.
You can also read Types of Broadband Networks, Definition, Uses, & Importance