Which path does the router use to reach a redistributed?
In the topology we see an ASBR doing mutual redistribution between BGP and OSPF. It is connected to two areas, one normal area and one no-so-stubby-area. The question in the task is which path does the bottommost router use to reach a route redistributed into OSPF from BGP. That this is some kind of trap is obvoius. What happens to a redistributed route when traversing those areas in this topology?
Configuration of OSPF gives us a choice between 5 different areatypes:
- Totally Stub
- Not So Stubby Area
- Totally Not So Stubby Area (omg what a name?!)
To just list all the different types and behavior would not give the reader much help, it would just be confusing, so I decided to break it down the same way i did to learn it myself.
Area 0 is the only backbone area in ospf, all areas needs to be connected via some ABR to Area 0 in order to be routed to another area. For example:
Area 1 —Router1—Area 0—Router2—Area 2—Router3—Area3
Here will all routers act as Area Border Routers, R1 will be the ABR for areas 1 and 0. Router 2 will be the ABR for areas 0 and 2.
Router3 will be the ABR between areas 2 and 3… ?
Task: Make sure traffic to 10.100.123.0/24 uses the path via R2 by changing the bandwidth
What path does OSPF choose to a specific destination?
Recently when I have been labbing OSPF more than before I noticed I stumbled when verifying the path ospf chooses for a specific destination. Lets look for instance at the topology here. First lets see with ospf database commands which path the routers uses and why, then how we can manipulate this.
I find it both good to know and very likely that tasks like this will be tested in the ccie lab. So I did some extra training to become more comfortable with this kind of tasks with EIGRP. The task in my example is just taken out of nowhere, but Im going to solve it with a general approach.
Make sure that traffic to the network depicted as a switch is loadbalanced from Router E between both paths, and Path A should be used 6 times as much as the other! The routers are configured with metric weights 0 0 1 0 0, in other words, only k3=1, so delay would be the only parameter for metric calculations.
My friend, extremely skilled CCIE and collegue sent me this diagram showing a topoligy the other night. We were discussing metric weights and path selection with EIGRP.
The task was this “Calculate which routes different routers in this diagram will install to their routing tables. For instance if R2 is configured with a variance of 2, will it install the route it gets from R4?”
Its really no coincidence that im not very handy with RIP, yet. Its just not that commonly used nowadays, not at any customer I have worked with so far at least. Lack of real life experience sure makes it more to learn. So is route filtering with RIP so I made some notes on the topic.
ODR – On-Demand Routing was nothing that i had heard of before i hit it in the workbook. The idea is kind of clever, but limited. The idea is to use CDP to advertise a default route in a topology with hub and stub spoke routers. Where a default route and only a default route will ever be needed. The implementation is simple, one single command on the hub routers lets the router advertise a default route to connected routers via CDP.
To make ODR work you need to enable CDP, obviously. And also make sure no other routing protocol is running there. Also kind of makes sense. And ofcourse the one command to enable the ODR function – “router odr”
Here comes a brief guide on how to implement and use IP SLA.