Re: RFC 2370 Update and a Proposed Change to Stub Area Behavior

[Mike Fox <mjfox at US.IBM.COM>]

Dave Katz wrote: 

> I guess the broader question is, does this really
matter at all?  Are  
> stub areas even useful any longer?

Yes and yes!!!

> Once upon a time, routers were memory-starved little boxes, but it's
 
> not clear to me at this point why anybody actually needs stub areas,
 
> unless they're still running AGS+'s someplace.

Not every box running OSPF is a router.  My platform,
z/OS, is an example.  It is far from a "memory-started little
box." However its main purpose is to be an application host but it
runs OSPF in support of value-added features like sysplex and dynamic VIPA,
and also to optimize failover recovery and interactions with its routers,
etc.  Our customers often need to run OSPF but they want to devote
as few resources to it as they can because they want to spend the cycles
running databases, CICS, IMS, and other business applications.  For
several years we have been recommending that our customers put our boxes
into totally stubby areas whenever possible, and have gotten very positive
results with that setup.  

So given the above you may not be surprised to learn
that my preference is to NOT allow unknown LSAs (or any additional LSA
type) in stub areas.  We don't support NSSA on our platform so I don't
really have an opinion on what you do with that. 

Mike
-----------------------------------------------------------------------
Enterprise Network Solutions
-----------------------------------------------------------------------
Research Triangle Park, NC  USA