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.
  
  
  
  

OSPF Neighborship

OSPF Neighbor Relationship Process

In this post , We will see about OSPF neighborship process.

•  First you have to give the IP address to any interface of router which is going to participate in OSPF, As you type router ospf 1,it will choose ospf router id. The criteria is highest IP address will became ospf router id but loopback interface will beat the election process. You can manually configure in router id in ospf process: router-id syntax. This will override every dynamic process.

• After that configure network command in ospf routing process for particular interface with proper subnet mask, this will add that interface in ospf routing process, After that particular interface will send hello packets to 224.0.0.5

• Have a look on OSPF Packet format:

• Hello packet format: (Wireshark view) This is OSPF hello packet captured from tool,You can see the required fields in OSPF Hello packet.

Different States of OSPF neighborship,
There are different states of OSPF neighborship process, Which can be seen by show ip ospf neighbor command.

• Down State: In this state no hello packet is received from neighbor within last dead interval.

•   INIT State: Router A sent Hello packet to router B. Router B received the hello packet but can’t see own router id in hello packet.(the valid hello packet is, receiving router should see its router id in hello in sender’s hello packet) 

• 2 WAY State: This state means in both router bidirectional communication is happened, both router can see each other router id in hello packets. At this state Router decides the how to form the adjacency with other (on broadcast or NBMA) same time it will form full state only with DR and BDR and for other neighbors it will in 2-way state. For point to point link it will form full neighborship. In the stage DR and BDR are elected.

• EXSTART State: In this state master and slave negotiation happens between Router and DR/BDR, to start the communication first and exchange the LSA’s. In Point to Point link it negotiation happens between neighbors. 

• EXCHANGE State: In this state both will exchange DBD (Database descriptors) packets with sequence numbers. Router will send LSR and LSU to each other.'

• Loading State: In this state, both routers will compare DBD sent by each other and if some information is missing, Then again router sends LSR and receives LSU. After all process all database is compared, it should be same in both routers. Both should be properly synchronized. 

• Full State: In this state, routers are fully adjacent with each other. All the router and network LSAs are exchanged and the routers databases are fully synchronized. SPF algorithm is applied and best path is calculated and Routing table of neighbors are completed.

Now we completed the neighborship process. Let's have a look on OSPF packet types:

OSPF Packet Types:

•  Hello Packet
•  DBD – Database description
•  LSR – Link state Request
•  LSU – Link state Update
•  LSACK – Link state Acknowledgement.