[Ns-developers] wireless phy status report

[Ruben Merz]

Hi Mathieu, hi all

> Today, I think that this goal cannot be really achieved: it seems
> impossible to me to share the same interface to multiple different
> wireless PHY standards which have very different fundamental
> architectures: CDMA, TMDA and FDMA systems have such fundamentally
> different modes of operation that I see no real way from the point of
> view of the MAC to make them behave in a similar fashion.

Mmmm, I'm not completely convinced. But then, we don't have any 
experience to really tell whether or not it is possible. Then, I can 
understand that many PHYs are vastly different from each other and that 
the effort of trying to come up with a common interface is maybe not 
worth the effort. Especially if the PHYs are then used by some MAC 
layers that are completely orthogonal.

In fact, the only thing that I'm really worried about is that things end 
up as in  ns-2: today, the whole physical layer of ns-2 is more or less 
tied to 802.11. Hence, if you want to implement a new PHY in ns-2, it is 
much more complicated than simply adding the files that implement the 
new physical layer. In addition, you have to modify all sorts of code in 
several locations in the source code of ns-2. Now, if several PHYs are 
implemented in ns-3, this should probably not happen.

So, my point is that, if we go along the route of having different PHYs 
implemented independently, we should really make sure that it is easily 
possible.

The good thing with a common interface is that it somehow forces people 
to take a step back and think about what they do. The process we had 
about the PHY interface on this same mailing list was, I believe, quite 
interesting.

> The idea, today, is thus to scale down the ambition of the project and
> to try to focus on designing a good wireless PHY interface for each
> different PHY standard. Thus, what I would like to propose is to try 
> to find motivated developers for each PHY standard independently and 
> start development efforts on each PHY standard separately. 

My current interest and priority is 802.15.4a. That is, a UWB physical 
layer and a MAC that is close to the one of 802.15.4 (without the _a_). 
I really would like to see such a phy available in ns-3. So, I'm willing 
to try to contribute an 802.15.4a phy to ns-3.

What's the proper way if we'd like to start such a development effort?

Best,
Ruben

> For now, my personal priority is 802.11 so, I plan to start porting 
> the 802.11 PHY and MAC model from yans to ns-3 in early february. 
> Once this port has reached a point where it is usable, I hope we will
> be able to integrate new different 802.11-specific PHY-level models 
> in this framework and extend it when needed to match the needs of 
> different PHY-level models. Of course, this effort will
> take place in a public repository and will be announced here once the
> code has reached a usable state. 
> 
> My hope is that similar efforts will be undertaken for each wireless 
> standard (I am thinking about wimax and others here) and that the 
> current wireless phy wiki will serve as a central repository of 
> information on each of the models we want to implement for 
> each PHY standard.

> Are there comments or flames about this new plan ?
> 
> regards,
> Mathieu