Maximum devices in OSPF area 0

Mailman List Mirror (Read Only)
1

Hello,

We are looking to deploy a greenfield MPLS network with OSPF as the IGP. I'm told OSPF areas don't play well with OSPF TED. For this reason, we are looking at using only area 0. Only Loopback interfaces and p-p core ethernet links will be in OSPF. What are the maximum number of Routers & Links that folks would be comfortable with putting in 1 area? If folks are already using this type of approach, could you share your current numbers?

Thanks,
Serge

2

If you aren't abusing your IGP with hundreds of thousands of routes, or
running some massive network with thousands of nodes on the edge, there
is no reason you can't (and shouldn't) run with a single area. IMHO
areas should be the anti-default, "if you has to ask, the answer is no".
The scenarios you see with lots of areas in books and labs are for
educational and testing purposes, not for sensible network design. :slight_smile:

3

Hello,

We are looking to deploy a greenfield MPLS network with OSPF as the IGP.
I'm told OSPF areas don't play well with OSPF TED.

Yep, OSPF-TE uses Type 10 LSAs, which only has an area flooding scope.

For this reason, we are looking at using only area 0. Only Loopback
interfaces and p-p core ethernet links will be in OSPF. What are the maximum
number of Routers & Links that folks would be comfortable with putting in 1
area? If folks are already using this type of approach, could you share your
current numbers?

I've seen single areas with as many as ~600 routers and as many as 6-7k LSAs
in the LSDB that functioned without any problems.

4

Hello,

We are looking to deploy a greenfield MPLS network with OSPF as the IGP. I'm told
OSPF areas don't play well with OSPF TED. For this reason, we are looking at using

you said .. greenfield.. why use OSPF?

5

> We are looking to deploy a greenfield MPLS network with OSPF as the IGP. I'm told
> OSPF areas don't play well with OSPF TED. For this reason, we are looking at using

you said .. greenfield.. why use OSPF?

I was thinking the same. If you run OSPF and want IPv6 some time in
the future you'll need to run OSPFv3 in addition. Much simpler to just
run IS-IS from day 1 and be done.

Steinar Haug, Nethelp consulting, sthaug@nethelp.no

6

Yet another vote in favour of doing IS-IS if you're going to do a greenfield
build;
it'll make your IPv6 native rollout go *so* much more smoothly (and you
*are*
planning for IPv6 now, since it's a greenfield rollout, right?

Matt