Very nice explanation as usual. Please let me know once you have uploaded VRF-lite, Inter VRF-Routing, EVN, NAT64, etc… on your web. I can’t wait to see those on your web coz I know you will make us understand clearly for sure.
That was a very detailed explanation. very straight forward. Thanks
Hi Rene, i am confused to understanding this concept. please help.
So what I could figure-out so far is :-
"In a transit AS scenario, Enabling BGP sync made sense. because otherwise our ASBR router will advertise EBGP routes (learned from EBGP neighbor) to other internal routers as well and ultimately may become a transit AS for both EBGP neighbors which we don’t want.
So the rule said if we learn a route from IGP then only we should advertise it to other EBGP neighbor. that means we are only advertising network that we own.
But now this rule is not needed in today’s scenario. …why ?(please help me in understanding)
BGP synchronization is an old rule in a time where not all routers within the AS were running IBGP. The problem here is not that we don’t want to become a transit AS, but that R5 is learning a route that it can’t reach (because R3 will drop the traffic).
By enabling synchronization, we will not advertise this network to R5 anymore which might be a better idea compared to advertising something that is unreachable. Only once the route is known in the IGP, we know that R3 will be able to forward the traffic and only then we will advertise the route to R5.
We don’t use this anymore since nowadays, we run IBGP on all our routers if we have a transit AS, or even better…we use MPLS VPN where we only require BGP between the PE routers.
Hope this helps!
Awesome work Rene!!!
on router 4 you type a “show ip bgp” and say that the “router will install the OSPF (AD 110) entry for 18.104.22.168/24
instead of IBGP (AD 200) route”
router 4’s BGP table says that LocPrf is 100. did you mean t6o say ad 100 and not ad110 ? if i have this wrong can you explain what the “LocPrf” as I think that it means the local preference of iBGP AS2
I didn’t watch the video again but if I was talking about the AD, it’s 110 for OSPF. If the router has to decide if it wants to use routing information from routing protocol X or Y then it will look for the AD. For example, 22.214.171.124/24 is learned through IBGP and OSPF…t will look at the administrative distance and prefers the lowest value. It will use OSPF since it has an AD of 110, compared to 200 for iBGP.
The local preference is used when the router has learned multiple paths for a network through BGP, and it has to make a decision which path to use. The default local preference value is 100. Here’s an example btw:
thank you Rene. I appreciate it
Hey Rene, I don’t quite understand how your BGP table can show RIB Failure in sub-section “BGP synchronization disabled” for R4’s route to 126.96.36.199, you have no IGP providing route to it, how can it be in RIB failure ?
In this output, it shouldn’t show the entry as a RIB failure. Probably something went wrong when I edited this post. Just fixed it so thanks for letting me know!
I have a very basic doubt. We configure the sync rule under the BGP configuration, so in this case, how would it know that there is a routing table entry for IGP route. Is the check for routing table entry performed by the router itself or the individual process such as BGP, OSPF or EIGRP.
When the sync rule is enabled, the route will not be installed in the routing table unless the IGP being used has learned about it.
It is not looking for the route in the routing table, since it will never be there because this is a condition for installing a route into the routing table. It will only be installed if this condition is met.
What the sync rule does is cause the router to first check if the route is “known” by an IGP. What this means is it looks in the topology table in the case of EIGRP, or the OSPF Link State Database to see if the IGP has a record of this route.
I hope this has been helpful!
Can you please update the IP address in the diagram, please
Can you be more specific? If there is an error in the lesson we’d like to have it fixed. Please let us know, thanks!
Can you add the subnet details on the diagram , I mean which interface is using whcih IP address so its easy to understand from the config?
Not sure what happened here but there should have been a diagram with subnets / IP addresses. Just fixed it, it’s there now.
Thanks for letting us know!
Very nice article. However redistribution BGP routes/prefixes to IGP should be avoided as possible all the time.
hi René and staff,
i am studying this lesson
What i understand is : with sync configured in our PE routers in AS transit, we prevent to advertise NLRI that are unreachable from others external AS. Is it right ? (reading the old posts seems to confirm this is right)
It was a configuration used in the old days, not now (so no sync is always the default behavior for routers now, i presume)
But i wonder something: what did prevent from using full mesh iBGP in the old days ??? so this sync concept had to be created ??
I am confused on this !
In “the old days” nothing technical prevented you from creating a full mesh iBGP. However, BGP was a protocol that was generally available on more expensive platforms, unlike today. So it wasn’t common to have all nodes within your AS participate in iBGP, but just the edge nodes that connect to other ASs. Now that it is common, the best practice is to either keep synchronization disabled, or redistribute the prefix into the AS via an IGP.
I hope this has been helpful!