Category: RIP

Routing Protocol Key Concepts

These are the key concepts I need to understand for routing protocols:

  1. How neighbors are formed and maintained
  2. How the best path is calculated
  3. How aggregation is configured and deployed
  4. How external routing information is handled
  5. How protocols interact
Advertisements

External Routes Preference for IGP and BGP

  • In general, EGP prefers external route and IGP prefers internal routes.
  • BGP prefers eBGP routes (AD 20) over iBGP routes (AD 200)
  • EIGRP prefers internal routers (AD 90) over external routes (AD 170)
  • OSPF prefers intra-area route over external routes (E1 and E2). By default OSPF ASBR redistribute external routes as E2 routes (internal costs are not considered). OSPF uses same AD (110) for internal and external routes. OSPF prefers E1 over E2 routes regardless of the metrics
  • IS-IS prefers internal routes over external routes. IS-IS uses same AD (115) for internal and external routes

Classic Routing Network Meltdown

The classic scenarios of routing meltdown is links flapping that consumed high CPU utilization due to route convergence calculation.

The avoid routing meltdown, we actually want to slow down how routers’ reaction to changes on the network.

There are four methods to slow down routers’ reaction:

  1. do not report all physical links changes to routing protocol
  2. slow down is declaring a neighbor is dead (dead timer)
  3. slow down in telling other routers on the changes (LSA, LSP, routes updates)
  4. slow down the time that the routing protocol reacts to topology changes

Routing Protocols for IPv6

Overview

  1. The IGP and BGP will work pretty much the same for IPv6 as in IPv4, with some adjustments which are highlighted below.
  2. The IGP that supports IPv6 are: RIPng, EIGRP, OSPFv3 and IS-IS
  3. Link-Local Address is used in IGP to establish neighbor  relationship
  4. OSPF has been modified most heavily among all routing protocols to support IPv4. It is called OSPFv3. (OSPFv2 is for IPv4 only).

OSPF

  1. per-link, instead of per-subnet protocol processing because it is possible to have hosts that are on the same layer 2 link but with different prefixes.
  2. does not support authentication because IPv6 natively supports IPSEC AH and ESP
  3. it is possible to have multiple addresses per interface
  4. use of link-local addresses as neighbor addresses
  5. two new LSA:
    1. Link LSA (Type 8)
      1. Advertising the link-local address to all routers that are attached to the link
      2. Advertising the IPv6 prefixes on the link to the routers that are attached to the link
      3. Advertising options
    2. Intra-area LSA
      1. Associating a list of IPv6 prefixes with a transit network by referencing a network LSA
      2. Associating a list of IPv6 prefixes with a router by referencing a router LSA
  6. Flooding of LSAs has been changed ans was made easier. There are three flooding scopes defined:
    1. Link-Local Scope
    2. Area Scope
    3. AS Scope
  7. OSPFv3 can be used in conjunction with OSPFv2, because they run independently of each other. You can also run multiple instances of OSPFv3 on a single link

IS-IS

IS-IS was designed with extendibility in mind. It uses TLV (Type-Length-Value) to define information and can be easily extended to support new TLV.

Two new TLV were added to support IPv6:

  1. IPv6 Reachability
  2. IPv6 Interface Address

IS-IS can  only has one topology for both CLNS and IPv4, thus it is important to have CLNS and IPv4 configured on each interface. Having CLNS or IPv4 address only will create suboptimum and routing failures in the network.

Cisco introduces multi topologies IS-IS so that IS-IS can have difference topology for IPv4 and IPv6. Because many times IPv6 topology is different than IPv4 topology, for example some routers do not support IPv6.

Four new TLV are defined to support multiple topologies:

  1. Multi-Topology TLV
  2. MT IS TLV
  3. MT Reachable IPv4 Prefixes TLV
  4. MT Reachable IPv6 Prefixes TLV

By default, a single topology for CLNS, IPv4 and IPv6 is used in Cisco devices. You will need to specify “multi-topology” under IPv6 address family of router IS-IS if you want to run multiple topologies.

EIGRP for IPv6

EIGRP originally supports multi protocols already: IPv4, IPX and Appletalk. Each protocols use different topology tables. So it was easier to add IPv6 on EIGRP compares to OSPF.

EIGRP hello are multicast packets with destination of FF02::A.

Packets sent to a specific peer contain the unicast link-local address.

BGP

MP-BGP (RFC 2283) was developed to support multiple protocols on BGP. MP-BGP is used to route MPLS L3 IPv4VPN prefixes and labels. MP-BGP is also used to route IPv6 prefixes.

MP-BGP defines two new attributes:

  1. Multiprotocol Reachable NLRI
    1. Address Family Identifier (AFI)
      1. IPv4
      2. IPv6
    2. Subsequent Address Family Identifier (SAFI)
      1. Unicast
      2. Multicast
      3. Unicast and Multicast
      4. MPLS Label
      5. MPLS-labeled VPN
  2. Multiprotocol Unreachable NLRI
    1. Address Family Identifier (AFI)
      1. IPv4
      2. IPv6
    2. Subsequent Address Family Identifier (SAFI)
      1. Unicast
      2. Multicast
      3. Unicast and Multicast
      4. MPLS Label
      5. MPLS-labeled VPN