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

stds-802-16: Re: Re: Published: MAC Modeling and Simulation Tools: Recommendations






Mat,

Thank you for your quick input.  Here are some of my thoughts in response.

I like the idea of using SDL for specifying the protocol state machines and
procedures.  At 3Com, we use some SDL tools (we use Verilog Object Geode,
Cadence Bones and Mil3 Opnet) for specifications and synthesizing protocol
implementations.  3Com also has an internally-developed Opnet model for cable
modem systems.  At the least, I think we should use tools like SDL editors, etc.
for specifying MAC protocol operations.  But even though 3com has an investment
in these tools, they are expensive, and I am not convinced that they are the
right thing for our committe.  The learning curve and the investment required to
host the tools (typically for UNIX workstations only) limits active
participation to larger companies that can afford it.
Additionally, if we chose a freely available tool that's in wide use in
acedemia, and make our models public, it could give 802.16.1 more attention, and
perhaps even attract researchers to work on 802.16.1 models.  Generally, I think
such availability would be good for the BWA industry.

When I undertook a project at 3Com that required discrete event simulation (for
evaluating MAC protocols for a wired network), I evaluated these tools and ended
up using ns, mainly for the reason that it modeled Internet protocols very well
(e.g., TCP).  I was pleased with the results, and think the task group should
consider ns.  I haven't looked in detail at the NIST simulator, but it obviously
could give us a leg-up because of it's 802.14 background.

I took a look at the Monarch stuff you mentioned.  That looks very interesting
and is a good example of how ns is used for 802.11 protocol evaluations and
mobile-IP.

Best Regards,

Brian Petry
3Com




Sent by:  Brian_Petry@3com.com


To:   stds-802-16@ieee.org
cc:    (Brian Petry/PA/3Com)
Subject:  stds-802-16: RE: Published: MAC Modeling and Simulation Tools:
      Recommendations







Forwarded message from Matthew Sherman regarding MAC simulation tools...

Roger / Brian,

Your option if you want to echo this to the 802.16 reflector.  Just a few
comments on MAC simulation.

I know that I'm looking into SDL for 802.11 simulations.  A couple of other
people are also looking at this approach and feel there is merit.  While SDL
is really a software development tool (in my mind) there appears to be an
ability to simulate MAC performance as well.  Both 802.11 and Bluetooth
(802.15) include SDL models for the their MAC's directly in the
specifications.  One of the neat things about this approach is that their
can be direct feed back between the specifications, the simulations, and I
think the actual design process (where relevant).  SDL is available from
Telelogic (www.telelogic.com).  They will generally make the editor
available for free for standards work, although the actual simulator is
quite expensive.  Still, I know of several people going this route.

I had recently discussed MAC simulations with Nada Golmie at NIST and got
the impression (perhaps falsely) that she more favored a tool such as Opnet
or custom coding in C for MAC simulations.  Also, note that while an 802.14
model may be available in NIST's ATM simulator, I have also seen Opnet code
for 802.14, and I think NIST coded DOCSIS 1.1 in Opnet.  I believe NIST
would make their DOCSIS model available for free.  While, Opnet is not free,
it is commonly available at most large companies.

If you are serious about ns, you might also want to check out the monarch
web site and see if there is anything useful for you there.

     http://www.monarch.cs.cmu.edu/

I was looking at it for 802.11, but making all the little pieces play
together looked a bit daunting to me.

Just some friendly thoughts and comments.

Mat

Matthew Sherman
PTSM - Communications Technology Research
AT&T Labs - Shannon Laboratory
Room 3K18, Building 104
180 Park Avenue
P.O. Box 971
Florham Park, NJ 07932-0971
Phone: +1 (973) 236-6925
Fax:     +1 (973) 360-5877
EMAIL: mjsherman@att.com




> -----Original Message-----
> From:   Roger B. Marks [SMTP:marks@boulder.nist.gov]
> Sent:   Wednesday, January 05, 2000 4:54 PM
> To:     stds-802-16@ieee.org
> Subject:     stds-802-16: Published: MAC Modeling and Simulation Tools:
> Recommendations
>
>
> Published Document: IEEE 802.16.1mc-00/06
> Title: MAC Modeling and Simulation Tools: Recommendations
> URL:
> <http://grouper.ieee.org/groups/802/16/mac/contrib/802161mc-00_06.pdf>
> Source: Brian Petry <mailto:brian_petry@3com.com>