Re: draft-pillay-esnault-ospf-flooding-05.txt

[Alex Zinin <zinin@psg.com>]

Bill, Randy-

Though Bill took the draft off the agenda, I'd like
to explain couple of things. Inline below, pls:

Wednesday, June 11, 2003, 5:23:43 PM, Bill Fenner wrote:

>> * DC circuits spec is from 1995; they are practically dead/useless today, 
>>who cares about the feature?  They're junk in the spec.

> Ok, you don't need 1793 for its original purpose, that's fine.

>> * Now, this spec creates a dependency on DC circuit spec; this is useless 
>>if DC circuit spec is not implemented in all OSPF routers

> Well, most OSPF extensions are useless if the extension isn't implemented
> in all OSPF routers.  If vendors want to offer this ability they can
> implement 1793 for this purpose, even though they don't need to implement
> it for its original purpose.

1793 defines 3 main changes in OSPF:

 1. Do-Not-Age LSAs, i.e., ability to mark LSAs to not expire as long
    as the originating router is available.

 2. Hello suppression to make sure DC links are not kept up/bouncing
    because of hello packets

 3. Change-only flooding over DC links to avoid periodic LSA refreshes
    from bringing links up

The draft in question only depends on (1) above, which is a generic
mechanism, which most of widely deployed implementations have already.

Also, assuming that the problem this drafts solves needs to be solved,
using DNA LSAs is the natural way of doing this.
    
>> * The bandwidth required for flooding a few LSA's every 30 minutes is 
>>minimal.  Why bother with something like this which could lead to a lot of 
>>issues?

> I'll bring this issue back; I know it's not just a thought
> experiment since there are implementations presumably stemming
> from customer request...

First-hand info (Padma and I were in the same company then): it was
asked by customers: "I see too many refreshes floating throughout the
network, and I can bump my refresh timer in ISIS, but there's nothing
like this in OSPF". As more stuff was put in the IGPs (e.g. TE
attributes) and networks grew bigger, customers asked more and more to
make the OSPF refresh interval configurable, rather than constant. So,
we have this hack now.

>>==> uhh, no.  what if you flood LSA's with noage bit, they stick around
>>forever and are never purged?

No. They would not be purged only as long as the originating router
is reachable. Once it's been unreachable for MaxAge seconds, it's
LSAs will be purged.

>>I'd guess that changes the OSPF protocol assumptions quite a bit.

The DNA mechanism does change the assumption, but this not something
new that this draft introduces. DNAs are defined in 1793.

>>  I didn't bother to check how well they were
>>documented in the OSPF DC circuits doc, but I wouldn't count on it.

In fact, 1793 does talk about it, and explains decrease in robustness
quite nicely. Please take a look at section 6.

Alex

> 1793 has the beloved 8-word security considerations section;
> 2328 talks about authentication but doesn't talk about problems
> or consequences.  I agree that this is a problem.

>   Bill