OSPF Network Types

• 
  

• 
  
  
  
  Here we will discuss, what are the OSPF network type’s which supports on interface level to form adjacency. Some time we have to configure different network types in OSPF configuration as per network requirement. Let’s see what they are,
  
  
  *But before you will read more… To run OSPF network, we preferably use same network type in network, but if in such conditions when we have to configure OSPF between different OSPF network types, what we have to do? It is possible but till some extent by tuning hello and dead timers.
  You can use different OSPF types as mentioned below:
  Point-to-Point with Point to Multipoint (Tune hello and dead timers)
  Broadcast with NON-broadcast (Tune hello and dead timers)
  Let me know what your results…are. It will work.
  
  
  Below are OSPF network types:
  
  
  Point to Point networks:
  This network type works on point to point networks. If you are using this network type, no DR/BDR selection in this network, 10 sec hello timer and 40 sec is dead timer, Routers will communicate to only one multicast IP address 224.0.0.5.
  This network type is majorly used between two routers.
  
  
  Broadcast:
  This network type works on Ethernet networks/Shared media. In this network DR/BDR elected for further communication between Routers. In this every router can reach each other, so it is Multi access. 10 sec hello timers and 40 sec dead timers, in this network type router will communicate to two multicast IP address 224.0.0.5 and 224.0.0.6.
  
  
  NBMA, Multi-access networks:  This is having five modes which we can use, two are industry standard and three are Cisco proprietary.
  
  
  Non Broadcast mode –
        It’s a default mode for frame relay networks and ATM networks.
  In this mode no broadcast and multicast are allowed, so here is issue. How the OSPF neighborship is going to form. We have to statically configure neighbors with neighbor statements in Hub location/DR, the router will communicate to neighbors by unicast.
  
  
  router ospf 1
  neighbor x.x.x.x priority 0
  
  
  This mode acts like a LAN environment, So DR and BDR selection will happen. (Full state with DR/BDR)
  The whole network should be in one single subnet.
  30 sec hello timers and 120 sec dead timers.
  In frame relay network make sure about frame relay map statements to all neighbors with same DLCI. They should able to ping each other.
  
  
  Point to Multipoint mode –
       No DR/BDR in this mode. It overcomes the limitations Non- broadcast mode, No need of extra frame map statements. The whole network should be in same subnet. 30 sec hello timers and 120 sec dead timers. No need to configure neighbor statement manually. Each router will be in full state with each other. It allows broadcast and multicast in network, so neighborship is auto discovered.
  
  
  ip ospf network point-to-multipoint
  
  
  Point to Multipoint, NON- broadcast mode – (Cisco)
  
  
  • 
    
    It is Cisco proprietary mode, This is same like Point to multipoint mode, But in this broadcast are not allowed, So we have to configure the neighbors manually with neighbor statements in OSPF Process and same time in interface specify the network type –                                                                                                                                                
  
  
  
  
  
  Router(config-if)#ip ospf network point-to-multipoint ?
   non-broadcast  Specify non-broadcast point-to-mpoint network 
  
  
  In this also NO DR/BDR elections and network should be in one single subnet.
  
  
  Broadcast mode – (Cisco)
  
  
  • 
    
    It works like LAN network in NBMA cloud. But Full mesh is required in this network, one subnet is required in network. DR/BDR is elected. Neighbors are auto discovered. 
  
  
  
  
  
  Syntax – ip ospf network broadcast.
  
  
  Point-to-Point Mode – (Cisco)
  Use separate sub interfaces, No DR/BDR is elected, requires different subnets, Neighbors auto discovered.
  
  
  This networks are depends on network design and your preference.