[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RRG] Benchmark scenario for modeling

Toni:

Thanks for your comments.

I tried to avoid using the word "mobility" in my message.
http://www.ops.ietf.org/lists/rrg/2008/msg00363.html
Mobility usually implies session continuity, tight delay constraints, etc.

Here I am trying to focus on what kind of extreme churn
a mapping distribution protocol should be able to deal with
and still maintain performance. You may perhaps like to think of
the scenario as migrating or "nomadic" subnets.
This scenario represents high dynamicity due to PI prefix splitting
into many subnets, and the subnets migrating to other regions and
being serviced by other ISPs'. The subnets (more specific prefixes but
still PI) are then injected into the map-encap space, and thereby give
rise to high dynamicity or churn for the mapping distribution protocol.
Not only that but also the subnets migrate back to home base
and are reposited into the parent prefix frequently.
The parent prefix is serviced via eBGP by the home base ISP.

This scenario can be viewed as another form of multi-homing with
high dynamicity. The multi-homing like behavior is due to the nomadic
behavior consisting of prefix splitting into subnets,
temporary re-homing of the subnets at other ISPs/regions,
and later the subnets returning to the home base.
This process repeats everyday perhaps multiple times.

BTW, I would include models for the common-case multi-homing as well
in the overall scenario, which should consist of multiple components.

Sriram

At 01:46 AM 2/19/2008, Tony Li wrote:


On Feb 18, 2008, at 9:07 PM, ksriram@nist.gov wrote:


(Note: I start with the airline-related scenario because I feel it
represents significant mapping dynamics / churn.  I mention other
plausible applications or components of the overall scenario later.)



Ok, but wouldn't it make some sense for the primary thrust of any
modeling effort to be the common cases?  While I want to see mobility
solved as much as anyone, that seems like the icing on the cake.
Shouldn't we first model operations in a basic multihoming
environment, with one of the various mapping solutions?

Tony







--
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