TWiki > DNA > DNASolution1 > DNASoln1Issue008	TWiki webs: Main | TWiki | Know | Sandbox
	DNA . { History of Changes | DNA Documents | Search | Go }

DNAO

There was originally, and occasionally still is, talk about Learned
Prefix Information Options (LPIOs).  The draft actually just has a
DNA Option (DNAO) that includes all of the learned prefixes, and the
total number of prefixes thought to be on the link.

Is this acceptable? 

Discussion

Jim

Yes.

            jak

Erik

One thing I haven't sorted out is whether we need the valid lifetimes for the prefixes in the DNAO, or whether it is sufficient to say that the router MUST NOT include any prefixes with a zero valid lifetime in the DNAO.

   Erik 

Jim

Erik,

What use would valid lifetimes be? The purpose of the DNAO is to provide the
host with indication of prefixes that the router does not advertise, in
order to provide link identification. If the router doesn't put zero
lifetime prefixes in the DNAO, then shouldn't that be sufficient?

Otherwise, why not get rid of the DNAO and simply have all routers advertise
all prefixes?

            jak

Erik

James Kempf wrote:

> Erik,
>
> What use would valid lifetimes be? The purpose of the DNAO is to provide the
> host with indication of prefixes that the router does not advertise, in
> order to provide link identification. If the router doesn't put zero
> lifetime prefixes in the DNAO, then shouldn't that be sufficient?


I think so. I'm just exploring the semantics and how they are consistent.

I think the semantics I'd like for the DNAO list is:
    If you sent (well, had sent, since things might have changed by
    now) a landmark option for this prefix, I would have responded
    with a "yes".

And more generally,
    A router will respond "yes" to a landmark option iff the prefix
    is either one of the PIOs that it is advertising (and has a
    non-zero valid lifetime), or in the DNAO it is advertising.

With those invariants, the DNAO without a lifetime is the right thing to do.

  Erik 

Jim

These seem like the right semantics to me.

            jak

----- Original Message ----- 
From: "Erik Nordmark" <erik.nordmark@sun.com>
To: "James Kempf" <kempf@docomolabs-usa.com>
Cc: "Brett Pentland" <brett.pentland@eng.monash.edu.au>;
<dna-dt@eng.monash.edu.au>
Sent: Tuesday, April 19, 2005 6:14 PM
Subject: Re: [DNA-DT] Solution1, Issue 8: DNAO



>> James Kempf wrote:
>
>>> > Erik,
>>> >
>>> > What use would valid lifetimes be? The purpose of the DNAO is to provide

the

>>> > host with indication of prefixes that the router does not advertise, in
>>> > order to provide link identification. If the router doesn't put zero
>>> > lifetime prefixes in the DNAO, then shouldn't that be sufficient?
>
>>
>> I think so. I'm just exploring the semantics and how they are consistent.
>>
>> I think the semantics I'd like for the DNAO list is:
>> If you sent (well, had sent, since things might have changed by
>> now) a landmark option for this prefix, I would have responded
>> with a "yes".
>>
>> And more generally,
>> A router will respond "yes" to a landmark option iff the prefix
>> is either one of the PIOs that it is advertising (and has a
>> non-zero valid lifetime), or in the DNAO it is advertising.
>>
>> With those invariants, the DNAO without a lifetime is the right thing to

do.

>>
>>    Erik

Brett

This one seems resolved:

Brett Pentland wrote:

> There was originally, and occasionally still is, talk about Learned
> Prefix Information Options (LPIOs).  The draft actually just has a
> DNA Option (DNAO) that includes all of the learned prefixes, and the
> total number of prefixes thought to be on the link.
>
> Is this acceptable?


Yes, with the following semantics suggested by Erik:

"
   I think the semantics I'd like for the DNAO list is:
       If you sent (well, had sent, since things might have changed by
       now) a landmark option for this prefix, I would have responded
       with a "yes".

   And more generally,
       A router will respond "yes" to a landmark option iff the prefix
       is either one of the PIOs that it is advertising (and has a
       non-zero valid lifetime), or in the DNAO it is advertising.
" 

-- Main.BrettPentland - 18 Apr 2005

Back to DNA Issue Tracker Base Page