What are DR and BDR?

DR and BDR

The DR and BDR (designed router and backup designated router) are common in Ethernet networks, AKA multiaccess broadcast networks. They work together in OSPF to reduce OSPF traffic and they dictate control over the topology – they are like a central point of management for the OSPF segment (or subnet) they are located on. The DR and BDR tell other routers on that subnet about other OSPF routers and networks in that segment and other OSPF segments. So because the DR/BDR is elected per subnet, we can have multiple DRs in an area such as Area 0.

The router itself is not actually a DR or BDR, but the interface is because OSPF is a link-state protocol, hence a router can have one interface as a DR, and another interface as a BDR. The BDR is a backup of the DR and simply maintains its accuracy consistent with that of the DR and it takes over as DR should the DR fail.

If the interface is neither a DR or BDR, then it is simply a Drother – the interface is participating in the OSPF topology and simply listens to what the DR and BDR tell it. Drothers send their OSPF information to the DR and BDR which is then propagated to other Drothers.

In short, the whole point of a DR/BDR combination is to reduce routing updates on multiaccess broadcast networks by preventing all the routers involved from spamming their updates to one another – this process is controlled by just the one router – the DR. They are the spokespeople for collecting and sending OSPF LSAs to everyone else.

DR and BDR Election

The DR and BDR are elected on a network segment when Hello packets are exchanged. There are various states of neighbor discovery with Hello packets – the DR/BDR election occurs at the ‘Two-Way’ state and is required before the ‘Exchange’ state starts at which point they are not establishing adjacencies with Drothers.

There must be a DR and BDR on a broadcast network (224.0.0.6 is used to reach the DR/BDR) unless the capability is manually disabled?Thus, the link is effectively a ‘point-to-point’ (224.0.0.5 is used for communication) and therefore no DR/BDR is required because it is assumed there are only two routers, and broadcasting is not supported.

If a DR fails and loses its position as DR, but rejoins the segment after a new election has occurred and it has been replaced, the original DR will not assume its position again until the neighbor adjacencies have been reset.

The rule for electing a DR/BDR:

It should be noted if a router joins the segment that should be DR, but is not elected DR, this is because the adjacency needs to be flapped or DR/BDR failure before this new router is recognized

Like most sports – the highest-value wins

1. The highest interface priority becomes DR. Second highest becomes BDR.

• A priority value set to 0 = Drother
• A priority value set to 1 = default
• Priority must be configured before the election occurs otherwise an adjacency reset is required
• Since interface priority may not be chosen, the second tiebreaker comes into play

1. The highest router ID becomes the DR. Second highest becomes BDR.

• An unconfigured router ID is ineligible
• A router configured with an ID of 255.255.255.255 is like an interface priority of 255
• If a router joins the segment with a router ID higher than the current DR/BDR, nothing will happen here until the adjacencies are reset

1. The highest physical interface on a router even if OSPF is not enabled on it

• Chose one method or the other previous methods at this point

1. The highest loopback interface regardless if OSPF is enabled on it

?

Characteristics of the DR, BDR, and Drothers (Point Form)

• DR is responsible for updating all OSPF routers on a subnet or segment with OSPF information provided by Drothers.

While there is a DR/BDR combination per each subnet segment, the routing updates are propagated throughout the OSPF area and domain

• The DR concept is at the link level therefore a router can be a DR on one segment and a Drother on another. The interfaces are DR, BDR, or Drother, not the physical router itself

• BDR monitors the DR and maintains consistency with its LSDB (link state database) – it takes over should the DR fail

• The DR passes its LSDB to all new routers on the LAN

• The DR passes updates received from one neighbor on the LAN to all other neighbors on the same LAN

• A DR and BDR must be elected before routers exchange link-state info

• The DR multicasts at 224.0.0.5 (to the Drothers). Packets from Drothers are multicast at 224.0.0.6 (to the DR & BDR)

• Drothers send all their LSUs to the DR and BDR and never to each other

• The BDR is fault tolerance only. It forms adjacencies but does not update other routers on the segment and therefore will not send LSAs on behalf of the DR, unless the DR dies. In this case the BDR takes over as DR

• The BDR observes the DR. If no activity triggers the timer, the BDR takes over. The BDR basically uses the heartbeat function you see with other technologies

• There is only one DR and BDR per segment. An interface becomes one or the other via the election process

• When the BDR is elected as DR, there is another election held to determine who is the replacement BDR

• The DR’s sole job is to ensure all neighbors on the LAN, the Drothers, have an identical LSDB with it. Any new routers that emerge on the LAN form an adjacency with the DR in this regard

• Hello packets are exchanged between neighbors and this starts the election process

• Link state info is not exchanged until the DR and BDR are elected

• A broadcast (Ethernet) network requires a DR/BDR unless the broadcast capability of the link is disabled

• Point-to-Point does not require the DR/BDR because they do not use broadcast capability. It is under the assumption there is only one router on the other side

• There are nonbroadcast multiaccess which interconnect multiple routers, like broadcast, but don’t use broadcasting technology. Like Ethernet, but without the broadcasting. An example is Frame Relay