Summarization of OSPF routes

Hi! first, sorry for my poor english…

Based on this post http://forum.mikrotik.com/t/summarization-of-ospf-routes-lsa-type-7/102384/1 and this http://forum.mikrotik.com/t/ospf-filter-external-e2/114690/1 I was able to make a summary of routes

Create a filter to not redistribute /32 that is part of the /24 ( so, no more /32 that is inside the /24 would be announced)
Create a /24 blackhole route (this will prevent static loop as well)
Inside OSPF instance, just mark redistribute static routes.

But what would happen if instead of redistributing static routes, I create a bridge, put the network that I want to summarize to that bride (i.e. /ip address add address=192.168.0.0/24 interface=bridge1), and put the option of “redistribute connected routes”.
That works, it works, but…
Is it correct this way?
Wouldn’t I have more control this way?

Thanks!

Area ranges are the best way to summarize OSPF routes, if they work for your situation. It is generally advised to only do redistribution if it is necessary.

Ok. Area ranges are configured in the ABRs. But in this case are internal routers, with a PPPoE server and hundreds of clients each. My goal is that it is not published every /32 even inside the area.

Use separate area for PPPoE server only.

What if i have this situation

In AREA 1 could make a virtual link and make a stubby area separately… but in AREA 2?

And this is another thing that I don’t quite understand… what if i grow geographically to one side so AREA0, AREA1 and AREA2 are big enough (more than 60 routers)… how i keep growing. I also couldn’t use small stub areas beyond AREA1… how would it be?

Well, looking at you picture: indeed, you can use virtual link for additional area on the pppoe server in area 1, but not the one on area 2.
So you might need to redesign your network somehow - to “stretch” your backbone. Probably have a partitioned backbone.
This might be even more true, having your second question in mind.

I could stretch my backbone… but keep in mind what I’m saying: the backbone is already big enough.
I’m not in that situation yet, but thinking ahead, it is not clear to me HOW to implement it. At some point, all areas will have too many routers to handle (i am an ISP)
From my point of view, it is a limitation that the protocol establishes that all areas must be contiguous to AREA 0. However, OSPF is presented as “highly scalable”… i don’t know… Could you point me in the right direction?

By “stretch” - I mean exactly stretch - make it thinner, but longer. :slight_smile:

Backbone should be exactly what it is named - some high-bandwidth links between high-capacity routers that serve only to interconnect other areas.

Okay. I will keep it in mind for the design. Maybe I need to think that, not only do I have to change the configuration but also upgrade some equipment and links… thanks for your time!