01-04-2015 10:28 PM - edited 03-05-2019 12:29 AM
Hi all
i have a doubt regarding area 0 which is known to be backbone area and all other areas attached to backbone area by an ABR . Why can we make any other areas as backbone say area 1 is made as backbone and all other areas connected to this area 1 via an ABR . Why its not possible , why its made as a not possible topology ? Please let me know whats the reason for this . or is it possible
01-05-2015 01:11 AM
Basically is the way the protocol was made. Just as ISIS has level-2 and level-1 areas, OSPF has an area 0 (backbone) and non-backbone areas.
This can not be changed, you can make areas work non directly connected to backbone with virtual links and other methods, but you need a Backbone area.
Now, why would you like to have area 1 as backbone? Is just a simple number change from 0 to 1
01-05-2015 01:14 AM
Hi,
The need for a backbone area (ignoring its specific number at this point) in OSPF is given by the requirement to provide loop-free routing between areas. Inside an area, OSPF operates on a link-state basis and uses the Dijkstra algorithm to compute the shortest paths. The Dijkstra algorithm by itself never causes a routing loop, and so the loop-free routing inside an area is guaranteed. However, between areas, OSPF resorts to distance-vector routing which is prone to transient routing loops. Therefore, to prevent routing loops in inter-area routing, OSPF by its internal design mandates and imposes a logical star topology between areas: the center of the star is the backbone area, the spokes are the non-backbone areas. Even if you configure a physical link between non-backbone areas, OSPF will not use it. As a result, the logical topology of areas in OSPF is always a loop-free star topology with the backbone area in the center and all other areas attaching to the backbone.
Notice that there cannot be multiple backbone areas as a star cannot have multiple centers.
Now because only a single area can ever act as a backbone area, OSPF designers decided that this area will be assigned the number 0 from beginning and that this number will not be configurable. In theory, they could adopt your approach: decide which single area is going to be a backbone, and then configure the routers to know that it really is the backbone. However, notice that this approach only adds to the complexity of the configuration and does not bring you any added flexibility: only a single area can ever be a backbone, and if the backbone area number was configurable, you would have an additional burden of configuring all routers with the information what area number should be interpreted as the backbone area ID. Simply put, there would be no added value in this: The backbone area would still be only a single area, and you would in addition need to configure extra commands to tell routers who the backbone area is. With the established rule in OSPF that the backbone area is always the area ID 0, the life is simpler :)
Best regards,
Peter
01-05-2015 01:30 AM
Hi Peter
Thanks a lot for the detailed explanation .
Regards
Binoj
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide