[Ns-developers] Underwater acoustic module for NS-3

[Leonard Tracy]

Hi Mathieu,

Sorry for the delay in response.

I'll take care of the formatting violations.  Not sure how the tabs slipped
in :)

As for the UanAddress block:  There is generally no standard address type in
an underwater newtwork.  The few devices I'm familiar with, such as the WHOI
Micromodem, only offer 4 or 8 bit address fields to distinguish between
senders.  Given the size of the ocean, it is highly unlikely that networks
would `accidentally' overlap, so uniquie ID's between networks are not
important.

As far as code reuse:  I think UanTxMode is definitely a candidate.  A
possibly significant difference between UanTxMode and the WifiMode code is
that I expect users to generate new modes.  I have been considering adding
the capability for user defined key/value pairs in UanTxMode in order to
account for features or specs of modulations or phys that are not already
accounted for.  I've created a modes list that allows for installing
`supported' modes in a PHY and then made the PER and SINR calculation
functions into attributes that can be replaced, so a PHY layer can respond
to these new modes in an appropriate way.

Basically I'm still not sure about which code may be reused.  Several
researchers have written me in recent days expressing interest in trying
this out, so I'm hoping for some feedback on how well I've covered their use
cases.

For NS3.  My feedback is all good.  So far I've found the learning curve
very reasonable.  I feel the API is very consistant and pretty intuitive.
So kudos back at all the maintainers.  I wish I had more constructive
criticism.  I'll do my best to find some soon :)

Leonard






On Mon, Oct 27, 2008 at 11:50 PM, Mathieu Lacage <
mathieu.lacage at sophia.inria.fr> wrote:

> hi leonard,
>
> I went through your code so, here are some (very minor) comments.
>
> 1) The code looks pretty good, kudos !
>
> 2) I have spotted a couple of [tab] characters: the ns-3 coding style
> disallows them (http://www.nsnam.org/codingstyle.html)
>
> 3) the following:
>  .AddAttribute ("Channel", "The channel attached to this device",
>      PointerValue (),
>      MakePointerAccessor(&UanNetDevice::DoGetChannel,
>                                           &UanNetDevice::SetChannel),
>      MakePointerChecker<UanChannel> ())
> would be better if re-indented as:
>  .AddAttribute ("Channel", "The channel attached to this device",
>                PointerValue (),
>                MakePointerAccessor(&UanNetDevice::DoGetChannel,
>                                    &UanNetDevice::SetChannel),
>                MakePointerChecker<UanChannel> ())
>
> The above is not currently present in the coding style document but
> there have been private discussions to try to indent through proper
> vertical alignment. We should try to resolve that issue officially and
> document it.
>
> 4) It would be cool to at least minimally document how your addressing
> scheme in uan-address.h. (I know _zero_ about underwater communications
> so, I have no idea what a single byte address is about).
>
> 5) the following:
> Address
> UanMacCw::GetAddress()
> {
>  return this->m_address;
> }
>
> could be written more simply as:
> Address
> UanMacCw::GetAddress()
> {
>  return m_address;
> }
>
> 6) coding style in UanMacCw::StartTimer
> if ()
>  {
>  }
> else
>  {
>  }
>
> 7) I see that you spent a lot of effort on the PHY models. It looks also
> that it might be possible to share at least the uan-tx-mode code with
> the wifi codebase. You probably know better which other parts of your
> code could potentially be reused/shared so, it would be nice if you
> could outline these.
>
> Finally, I can see that you spent a lot of time using ns-3 so, I would
> be curious to get feedback about what you liked best, what you disliked
> most, etc.
>
> regards,
> Mathieu
>
> On Mon, 2008-10-13 at 19:48 -0700, Leonard Tracy wrote:
> > Hi all,
> >
> > I have begun work on developing a module to enable simulation of
> underwater
> > acoustic networks in NS3.  This work is still early in the development
> > cycle, however, I am hoping to solicit code reviews and/or suggestions
> for
> > design improvement.  The freehg repository is avilable at
> >
> > http://freehg.org/u/ltracy/ns3-uan
> >
> > My ultimate goal is to develop a general framework that can be used to
> model
> > a variety of different underwater scenarios (e.g. research new MAC layers
> > with various underlying PHY/channel models).  To this end, I've followed
> the
> > design model of the wifi module included in current NS3 releases (with
> some
> > pointers from Mathieu) with some modifications to account for the
> difference
> > in intended purpose.  So far I've developed simple examples for that
> > exemplify the interplay between the channel, propagation and noise
> models,
> > and the MAC and PHY layers.  I think all of these work, although I have
> not
> > done much testing.
> >
> > I'm hoping for inclusion into the main NS3 dev branch at some point,
> > possibly as soon as ns-3.4.  Thanks in advance for any comments.
> >
> > Leonard Tracy
>
>