How does OSPF know when a link is down vs a neighbor is down?

[RachelG]

Hi!

I’m a beginner in this area and I have a question. How does OSPF know when a link is down vs a neighbour is down? Which information give us that information? When I neighbour is down, LSA and Hello messages are used... is it the same when a link is down?

Thanks for your help!

[Flavio Miranda]

Hi @RachelG

 OSPF is a link-state protocol, so, identify link down must be a priority. OSPF use link-state  advertisement (LSA) . If there´s something OSPF knows well, is when a link fail.

-If I helped you somehow, please, rate it as useful.-

[RachelG]

Hi @Flavio Miranda!

Thank you for your answer. So OSPF works the same way even if it's about knowing if a neighbor is down or a link? Should it not be a different? Or can't we see the different if it is the link who is down or the neighbor?

[Flavio Miranda]

No. It has different mechanisms to identify both events. Neighbors down will first identified by Hello packets absent.

 Link down is known via SLA.

OSPF uses hello packets and two timers to check if a neighbor is still alive or not:

• Hello interval: this defines how often we send the hello packet.
• Dead interval:  this defines how long we should wait for hello packets before we declare the neighbor dead.

The hello and dead interval values can be different depending on the OSPF network type. On Ethernet interfaces you will see a 10 second hello interval and a 40 second dead interval.

[Richard Burts]

Perhaps I am not understanding something in this discussion. OSPF does use Hello messages to build neighbor relationships and also to verify that the neighbor is still functioning. If OSPF stops receiving Hello messages from a neighbor then it will mark that neighbor as down. OSPF will also mark a neighbor as down if it determines that the link connecting to the neighbor is down (this is quicker than waiting for the Hello and Dead timers). In my experience a link going down typically means that the interface carrying that link has changed line protocol state to down. OSPF will know that a link is down if it observes the IOS message that an interface changed link state to down. Or are we talking about something different here? If something different then please clarify what it is.

HTH

Rick

[RachelG]

Thank you for your answer! My question is; how a router knows if a connected link is down? You say it is by IOS messages, not by LSA? I haven't heard about IOS messages how does it work?

[Richard Burts]

IOS uses inter process messaging to communicate between various processes. So the process that manages the interface will communicate with the process for OSPF about the event when an interface changes to the down state.

HTH

Rick

[Joseph W. Doherty]

When a link goes physically down, the hosting device's hardware/OS is designed to notify routing protocols. This is to "speed up" link down detection. For example, hardware detection of link down might take 50 milliseconds or so, while detecting link lost via lost of neighbor adjacency might take 40 seconds. Regardless of how link down is detected, the OSPF notification and reconvergence process is begun.