[RRG] RRG shouldn't try to directly address mobility, including qualifying mobility attributes

["Fleischman, Eric" <eric.fleischman@boeing.com>]

Team-mates:

I've been grappling with mobility affects upon routing systems for much
of this decade. I am not optimistic that this group will be able to
directly address mobility in a manner that will result in an RRG
consensus. I also do not believe that this group is likely to be able to
qualify mobility attributes such as "granularity and churn" in a
satisfying manner. Rather, the following is the view from my knothole:

1) The semantics of mobility differs in different contexts. In my own
research, I have found it helpful to distinguish between geographical
mobility (i.e., the affects of one node moving relative to another,
which can also include the signal intermittence affects caused by
attenuation (distance) or signal blockages due to landforms, foliage,
buildings, particulate matter (dust storms), the weather, or the pitch,
roll, or yaw of the moving node impacting its own antenna's signal
propagation or reception) and organizational mobility (i.e., moving a
computer from one vehicle to another, moving part of an organization
into another part of an organization).

2) Different systems interpret geographical movement as being "mobility"
very differently. For example, for certain satellite systems, an entity
can move hundreds of miles and not leave the satellite's beam. From the
perspective of that satellite, that entity has not moved (i.e., no
mobility has occurred). By contrast, stationary nodes may exhibit
"mobility" given certain weather or particulate matter environments
(e.g., as a function of frequency ranges).

3) Routing protocols confuse underlying signal intermittence with
mobility. The impact of signal intermittence varies in terms of which
part of the protocol stack is being focused on. The higher one goes up
the protocol stack, the longer the duration the signal intermittence
event can be before the protocol must address it. One can take steps to
dampen these affects, but ultimately signal intermittence events will be
noted if their duration is long enough.

4) Mobility affects are a partial function of the network architecture.
Architectures similar to 3GPP, for example, naturally view mobility from
a mobile IP (MIP) perspective. By contrast, I have primarily been
working in MANET environments and those environments view similar
mobility events quite differently. It is my personal belief that
adherents of those two perspectives naturally interpret the same facts
differently -- and certainly various mobility attributes have very
different routing implications to each orientation.  

5) Ran Atkinson's posting on February 21st focused on a technical report
that for entities moving every few minutes or less, the protocol system
works better if mobility is handled at L2 than L3. Because I largely
agree with this conclusion, I mention Ran's posting here as another
datapoint suggesting that this group can't (fully) consider mobility
because L2 issues are out of RRG's scope. 

6) A great deal of published research has been addressing various
cross-layer protocol integration mechanisms for wireless systems over
the past many years. I believe that the technical literature is
concluding that cross-layer protocol integration techniques can
significantly dampen mobility affects upon IP routing protocols.
However, even if you agree with this conclusion, can RRG presume that
such techniques will be implemented in a consistent manner? I think not
(for the foreseeable future) ... thereby making the creation of a
consensus position for this group all the more difficult because we will
be talking apples and oranges -- and to what purpose?

7) I believe that the research community has already formed a consensus
that in wireless environments network performance and scaling is
enhanced if the network's topology is hierarchical. Of all of my
observations, I believe that this is the only useful observation that is
suitable for the RRG to use in order to create a consensus position.

Therefore, I believe that if RRG is to address mobility, it is because
the selected RRG approach is so clean that it incidentally also enhances
mobility (i.e., mobility is a secondary affect for RRG and not a gating
requirement).

--Eric
 

-----Original Message-----
From: William Herrin [mailto:bill@herrin.us] 

On Thu, Mar 13, 2008 at 6:18 PM, Tony Li <tony.li@tony.li> wrote:
>  Or, in other words, we should drop the mobility discussion for now, 
> deal  with the granularity and churn issues, and then see if we are 
> happy with the  bounds that that imposes on mobility.

Hi Tony,

Do we have a consensus on what target numbers for granularity and churn
could support mobility directly in the routing system? If not, I submit
that we should continue our tangent long enough to get them.

While we might or might not want to support mobility directly in the
routing architecture, it would be very informative if we could say:
yes, this proposal meets the target criteria for mobility in the routing
architecture and no, that proposal does not.

Regards,
Bill Herrin

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg