OSPFv2 Designated Router

The Designated Router (DR) is any OSPF router connected to a multiaccess network segment and that has the highest priority or Router-ID at the time of the DR election. The role of the DR is to simplify SPF calculations and relieve the need for all neighbors to be fully adjacent on a shared segment and thereby reducing the flooding of LSAs on shared segment with a large number of neighboring routers.

The DR election process

The OSPF routers will wait for the same amount of time as the Dead timer before electing a DR on the shared segment. This is known as the OSPF Wait Time and allows for a grace period for routers to boot up on the shared segment before electing a Designated Router.

On a shared segment there will be an election of a Designated Router (DR) and a Backup Designated Router (BDR). All other routers on the segment, that are neither DR nor BDR, will be a DROther. 

The specifics of the election process is describe below.

• Routers on a shared segment will listen to neighbors Hellos and collect the priorities and RIDs during the wait interval (wait interval is equal to the dead timer for the interface)
• If a Hello packet is received during the wait interval from a neighbor, which states that, that neighbor is in fact a BDR (which would mean that a DR is also present) the wait interval will expire immediately and the router will proceed to the DR/BDR election process.
• Likewise, if a Hello packet is received during the wait interval from a neighbor, which states it is the DR, but no BDR address is set, the wait interval will expire and the router will proceed to the DR/BDR election process.
• A router examines the RIDs and priorities collected during the wait interval and chooses the highest priority as the Designated Router and the second-highest priority as Backup Designated Router, though only if roles are not advertised by another RID already. If the priority is tied (it is set to 1 by default) the highest Router ID breaks the tie.

There is no preemption of the DR role in an OSPF network; if a DR is listed in the Hello packet it means that no election will be held - the router will join the network using the DR specified in the received Hello packet. This avoids unstable routers continually prompting a DR election, when they come online on the shared segment.

If for some reason the same segment has two DRs elected, each of the DRs will receive the others Hello packet stating itself as the DR for the segment. This will then prompt a new DR/BDR election on the shared segment to mend the network back into a single segment with only one DR and one BDR as the outcome.

OSPF DR/BDR Example

I will detail the workings of the DR with an example configuration using 5 routers connected over a shared ethernet segment. The topology is depicted below.

OSPFv2 DR/BDR Topology

The router R1 is the DR, which was ensured using the command ip ospf priority 255 under the interface connected to the shared ethernet segment. The router R2 was configured with the command ip ospf priority 254 to ensure it got elected as BDR. The routers R3, R4 and R5 are all DROther and configured with the command ip ospf priority 0 to prevent them from ever take part in the election process. Without the priority command the routers would use the default priority of 1 and the DR/BDR would be elected based on the OSPF RID. In the example here the RID is configured as X.X.X.X where the X represents the router number (R1 = 1.1.1.1, R2 = 2.2.2.2, etc). Assuming the priority was not set at all the router R5 would become the DR and the router R4 would become the BDR.

Below is the same topology shown, but this time with arrows indicating the neighbor adjacency states between the routers.

OSPFv2 DR/BDR/DROther Adj. States

The DR and BDR both establish a full neighbor adjacency between each other as well as with all the DROther routers. The DROther routers establish full adjacencies with the DR and BDR only. Between the DROthers they will be stuck in the 2-way state - meaning they receive each others Hello packets, but they do not exchange Database Description Packets with each other.

The process of flooding LSAs is shown and described below. In this case the flooding occurs in response to a simulated link failure on R3 (by shutting down the interfce). When the link goes down, router R3 sends an update to the multicast address 224.0.0.6, which only the DR and BDR listens to.

OSPFv2 Update Step 1

The DR receives the update and proceeds to flood it out its ethernet interface to the multicast address 224.0.0.5, which all OSPF routers listen to.

OSPFv2 Update Step 2

The routers receiving the update will then send an acknowledge to the DR and update their routing information according to the new information - in this case they will remove the network 140.100.3.0/24 from the RIB.

The DR is also the only router allowed to originate a type 2 LSA. The type 2 LSA describes the shared network segment as well as all the attached routers (meaning the routers that have established a full adjacency with the DR on the segment). The LSA type 2 for the example network is shown in the output below.

R1#show ip ospf database network
            OSPF Router with ID (1.1.1.1) (Process ID 1)
                Net Link States (Area 0)
  LS age: 1098
  Options: (No TOS-capability, DC)
  LS Type: Network Links
  Link State ID: 130.1.255.1 (address of Designated Router)
  Advertising Router: 1.1.1.1
  LS Seq Number: 80000007
  Checksum: 0x590A
  Length: 44
  Network Mask: /24
        Attached Router: 1.1.1.1
        Attached Router: 2.2.2.2
        Attached Router: 3.3.3.3
        Attached Router: 4.4.4.4
        Attached Router: 5.5.5.5

The type 2 LSA turns the shared segment into a star topology with the DR in the middle, connected to all routers (including the router that is the DR) on the shared segment.

So, the DR will prevent OSPF routers from excessively flooding the network with control packets and it simplifies the SPF algorithm by providing a single point from which to base calculations - the DR sees all and knows all for the shared segment.