ISIS on SDA

[JUANNN]

Hello,

Why is IS-IS level-1 not used in SDA, instead of level-1-2? Why on an SDA site we need to mantain a level-2 adjacency and database too, if we can just have everything on the same area? In what production scenario do we benefit from level1-2 on SDA?

Thanks,

Juan

[muhammadawaisaz17]

I agree with your point. In a self-contained SDA fabric, keeping L2 only because it is automated does not necessarily provide much additional value. If L1 already provides the required topology and reachability, using a simpler hierarchy can reduce unnecessary routing state while still meeting the design requirements. L1/L2 makes more sense when the fabric needs to connect with a larger ISIS domain or another routing area.

[swept]

I was wondering about this as well when looking at the IS-IS design used with SDA. If all the SDA devices are intended to operate within the same area, I initially thought level-1 would be enough.

What I’m trying to understand is the practical reason for keeping the level-2 adjacency and database. Is there a specific SDA topology or production scenario where level-1-2 provides a real advantage, such as connecting different sites, areas, or providing better path selection?

It would be helpful to see a simple production example showing where level-1-2 makes a difference compared with using only level-1.

[Andrii Oliinyk]

"practical reason for keeping the level-2 adjacency and database." - the same as keeping L1 concurrently with L2.
"Is there a specific SDA topology or production scenario where level-1-2 provides a real advantage, such as connecting different sites, areas, or providing better path selection?" -u can imagine such a scenario w/ huge IS-IS domain as underlay for fabric site where u would decide (again, just hypothetically) to implement partitioning in L1& L1/L2. CatC wont automate it for u. u'll need to develop both target IS-IS configuration & (likely) Day0-templates. But so far i can only interpret this topic as purely academical interest, coz to decide something like this one need to face scalability issue w/ L1/L2 as default. 

[Andrii Oliinyk]

Just borrowed it from linkedin chat:
https://bst.cisco.com/quickview/bug/CSCvy10542 (thanks to Freerk Terpstra)

[julian.bendix]

Hey!

Check this out - https://bst.cisco.com/quickview/bug/CSCvy10542

Taken from a recently deployed lab:

sda01#show run | i is-type
 is-type level-2-only
sda01#

BR Jules

[JUANNN]

I guess in scenario where there is an IS-ISI hierarchy like M02@rt37 said, level-2 only ensures that all nodes get for example the /32 route required for certain services, like the WLC. With level-1, again supposing there is an actual backbone, some nodes may receive just the default route created from the attached bit of an area border router (level1-2)...

[Andrii Oliinyk]

imagine chain of routers R1-R2-R3 in area X. R1&R3 run L1, & R2 runs L1/L2. imagine there are /32 addressed loopbacks running isis process as well & let them be L1-only. will R1&R3 have /32 routes of each other in their ribs by default? yes they will.
with regard to default route, each L1-router will have it as soon as L1/L2-router in area generates LSP w/ ATT-bit set.