Right, but that's not my point.
My point is that it's possible to concieve of link specifric mechanisms that
make filtering work.
Furthermore, this is all hypothetical. The other SDOs are going to do it
however they want to anyway, and they, ultimately, are the ones who control
how paging works. If they think unsolicited RAs are going to screw up their
paging, they'll write it into their specs that unsolicited RAs aren't sent,
regardless of what IETF does. IETT really can't make a big case about how
that is going to present a serious problem, as was done for addressing or
security during 3G development
There's really nothing we can do at the IP layer except modify how and
whether unsolicited RAs are sent. And every time the topic comes up, people
in IETF push back on it for various reasons. So let's just not waste any
more bandwidth on discussing it.
jak
----- Original Message -----
From: "Manfredi, Albert E" <[EMAIL PROTECTED]>
To: "James Kempf" <[EMAIL PROTECTED]>
Cc: <[email protected]>
Sent: Wednesday, September 20, 2006 1:17 PM
Subject: RE: Proposal to change aspects of Neighbor Discovery
I think Pars' point is not restricted to L2 mechanisms, though. No
matter what layer has to wake up a dormant host, bandwidth will be
consumed at multiple base stations to achieve this?
Bert
-----Original Message-----
From: James Kempf [mailto:[EMAIL PROTECTED]
Sent: Wednesday, September 20, 2006 3:03 PM
To: Pars MUTAF
Cc: Erik Nordmark; [email protected]
Subject: Re: Proposal to change aspects of Neighbor Discovery
So here's a counter example.
Suppose there is an IP based but wireless link layer specific
protocol that
lets BSes communicate about dormant mode hosts. When a host goes into
dormant mode, all BSes in the paging area learn about it via
the protocol.
When paging happens, this protocol is used by the BS where
the mobile node
originally went into dormant mode to notify other BSes to
page. Etc. That
should take care of filtering.
Right now, paging is handled by L2 specific mechanisms. In
3GPP, I think it
even depends on the MSC, i.e. the circuit switched part of
the network.
Bottom line is, other SDOs get to say how it works, not the IETF.
jak
----- Original Message -----
From: "Pars MUTAF" <[EMAIL PROTECTED]>
To: "James Kempf" <[EMAIL PROTECTED]>
Cc: "Erik Nordmark" <[EMAIL PROTECTED]>; <[email protected]>
Sent: Wednesday, September 20, 2006 12:49 PM
Subject: Re: Proposal to change aspects of Neighbor Discovery
> Selon James Kempf <[EMAIL PROTECTED]>:
>
>> There is an assumption here that the "L2 paging system" is
run off the
>> AR.
>> This is not necesarily the case. It won't be for Wimax,
for example.
>
> Hello,
> OK. The discussion about Wimax now ;-)
>
> But this should be orthogonal to our problem. In paging, in general,
> the system (AR in this case) doesn't know where the dormant host is
> located. The mobile host is "asked" to report its exact location,
> i.e. cell. (This is what is meant by "paging").
>
> *The BSs don't know anything about the dormant host*. The
host is paged
> in all cells of the paging area. That's why wireless
bandwidth needs to be
> consumed for paging in all cells of the paging area. (This
is a well-known
> problem of paging. 100s of research papers attempted to reduce this
> bandwidth cost.)
>
> Then, the host hears the paging message while sleeping in one of
> the cells, wakes up, and a location update is sent.
>
> The AR has now learned the current BS of the host. The packet that
> triggered paging is forwarded to the current BS of the host.
>
> Filtering of the RA by the BS is too late. Because the host was
> already paged in all cells and woken up.
>
> pars
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
