[Tinyos Core WG] Meeting: 3/11

[Prabal Dutta]

As a follow up to the call, the Quanto source code is available in contrib:

tinyos-2.x-contrib/berkeley/quanto

The changes are relatively lightweight and will be easy to digest.  A good
overview on the power state tracking and activity tracking is available in
Sections 2.4, 3.2, and 3.3 in the paper:

http://www.cs.berkeley.edu/~prabal/pubs/papers/fonseca08quanto.pdf

Let me (or Rodrigo) know if you have any questions.  Thanks.

- Prabal

On Tue, Mar 10, 2009 at 10:26 PM, Prabal Dutta <prabal at cs.berkeley.edu>wrote:

> Some background on Quanto for the call.  Quanto provides a network-wide
> time and energy profiling for embedded network devices.  It was built to
> help answer questions like:
>
> - What is the energy breakdown of my application by hardware subsystem
> (e.g. radio, MCU, flash, sensors)?
> - What is the network-wide cost of flood?
> - What fraction of time does the CPU/radio/flash do work on behalf of
> Deluge/routing/CTP?
>
> Quanto allows one to break down energy and time spent on behalf of logical
> activities (flows of execution) across hardware subsystems, nodes, and times
> -- and it allows one to connect the causal dots (it uses labels for this
> purpose -- think information flow control or taint tracking).  Quanto works
> by modifying a number of OS abstractions like timer, scheduler, queues,
> arbiters, interrupts, etc.  A total of about 40 core files are shadowed and
> modified for Quanto support (these exist in contrib/berkeley/quanto) on the
> Telos B and Epic motes.
>
> Maintaining a shadow copy of these files is a bit of a nightmare so the
> question is whether Quanto should be integrated into the core, and if so,
> how best to do this since Quanto touches so many pieces.  Although many of
> these can compiled out with dummy implementations, some things -- like
> interrupt handlers -- are very sensitive to timing concerns so that may be a
> worry.  The goal of the discussion is to (i) expose Core to the Quanto work;
> (ii) get feedback on whether it would be valuable to integrate it into the
> core tree; and (iii) discuss questions and concerns about how best this
> should be done.
>
> If you're curious, a copy of the Quanto paper is available online:
> http://www.cs.berkeley.edu/~prabal/pubs/papers/fonseca08quanto.pdf<http://www.cs.berkeley.edu/%7Eprabal/pubs/papers/fonseca08quanto.pdf>
>
> - Prabal
>
>
> On Tue, Mar 10, 2009 at 10:00 PM, Philip Levis <pal at cs.stanford.edu>wrote:
>
>> Wednesday, March 11, 2009, 09:30 AM US Pacific Time
>> Bridge: 5
>> Passcode: 4442994
>>
>> US: 1-916-356-2663 or 1-888-875-9370 (non-Intel)
>> UK: +44 1793 402663
>> Denmark: +45 4527 5090
>> Germany: +44 01793 402663 (Munich) +44 01793 402663 (Braunschweig)
>>          +44 01793 402663 (Ulm)
>>
>> Agenda:
>>   Status
>>   CC2420 security (John)
>>   AM/15-4/6lowpan (Jan, Miklos, Phil)
>>   Quanto (Prabal)
>>
>> Phil
>> _______________________________________________
>> Tinyos-2.0wg mailing list
>> Tinyos-2.0wg at millennium.berkeley.edu
>> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-2.0wg
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://www.millennium.berkeley.edu/pipermail/tinyos-2.0wg/attachments/20090311/484e8483/attachment.htm