[SensorNetArch] draft slides for friday

[David Culler]

I think one can make to much of the "not a broadcast".  802.15.4 can 
broadcast.  Our implementation doesn't do it, but it should.




Philip Levis wrote:

> A couple of points on the functional architecture:
> 
> Although the proposal claimed broadcast would be the primitive, it isn't 
> in the SP that Joe will present. Certain link technologies prevent 
> broadcast from being the lowest common denominator. That is in and of 
> itself an interesting finding and worth bringing up.
> 
> SP is not exactly a narrow waist in the same way that IP is. Among other 
> things, it isn't a protocol, but rather a systems abstraction. Two 
> systems running SP on top of the same data link layer are not 
> necessarily interoperable, unless they also use the same SP 
> implementation. Due to the ways in which sensor networks are deployed 
> and developed (generally single administrative domain), this is OK. It 
> just needs to be made clear that while we're using the same term, it has 
> a different meaning.
> 
> Slide 23: The formation and management of the connectivity graph (sec 
> 4.3 in the proposal) is essentially the SP neighbor table. The fact that 
> we've moved it into SP is an interesting finding and worth bringing up.
> 
> Slides 26-7: We spent a whole bunch of weeks discussing cross layer 
> services and what they mean. Much of SP came out of taking a hard look 
> at power management with regards to communication (Joe and Jonathan's 
> first cut on Deluge). Putting some of our progress here would be nice: 
> right now it mostly restating the proposal.
> 
> A bit of an observation: defining a unifying abstraction is inherently 
> at odds with the notion of cross-layer services. For example, SP can 
> provide feedback to a routing layer on how it could improve energy 
> efficiency (phase, etc.): how does the routing layer then convey this 
> information to the application which is sending data? Do the notions of 
> SP -- an abstraction *between* two layers, defining the control and 
> feedback -- push up to successive layers? It would seem the inevitable 
> conclusion...
> 
> Phil
> 
> -------
> 
> "We shall not cease from exploration
> And the end of all our exploring
> Will be to arrive where we started
> And know the place for the first time."
> 
> - T. S. Eliot,  'Little Gidding'
>