[Ns-developers] New ns-3 Stable Release Posted (ns-3.2)

[Tom Henderson]

craigdo at ee.washington.edu wrote:
> 
> New user-visible features
> -------------------------
>   a) Learning bridge (IEEE 802.1D)
>     It is now possible to bridge together multiple layer 2 devices to
>     create larger layer 2 networks. The Wifi and Csma models support
>     this new mode of operation. (contributed by Gustavo Carneiro)
> 
>   b) Python bindings
>     It is now possible to write simulation scripts in python using our 
>     python bindings (contributed by Gustavo Carneiro).
> 
>   c) Real-time simulator
>     It is now possible to run simulations synchronized on the real-world
>     wall-clock time (contributed by Craig Dowell).
> 
>   d) Network Simulation Cradle
>     It is now possible to use the Network Simulation Cradle 
>     (http://www.wand.net.nz/~stj2/nsc/) in ns-3 and run simulations 
>     using various versions of kernel TCP network stacks. (contributed
>     by Florian Westphal as part of his Google Summer of Code work)
> 
>   e) A statistics framework
>     Joseph Kopena contributed a statistics framework which can be used
>     keep track of simulation data in persistent storage across multiple 
>     runs (database and ascii file backends are available).
>     More information on the wiki:
> 
>  
> http://www.nsnam.org/wiki/index.php/Statistical_Framework_for_Network_Simula
> tion
>  

Craig,
Thanks for serving again as release manager, which was difficult this 
time around, and thanks also to everyone who contributed to this 
release.  In particular, my hat is off to Gustavo for all of the Python 
bindings work contributed to ns-3.2.

There are a few features for which we are seeking feedback from users 
and developers.  First, how are our APIs (especially the container-based 
helper APIs) working out?  Second, there are some experimental features 
that would benefit from some user feedback.  Joe Kopena put together a 
nice wiki page (referenced above) describing how he envisions a 
framework for running large numbers of simulations might be put 
together-- what do people think?  We have put the code into the 
contributed code directory (src/contrib/stats) while we solicit 
feedback; it would be useful to have a few users try it out and let us 
know.  In addition, Mathieu has put together a GUI-based configuration 
tool for simulation parameters.  Many users may not even know of this 
feature (also implemented in src/contrib); I recently added a chapter in 
our manual to introduce it 
(http://www.nsnam.org/docs/release/manual.html#SEC42).  Should we move 
this feature to the main tree, and how might it be improved?

As for roadmap development and general program planning and discussion, 
I'd like to mention that we are going to try to get developers together 
each week for an hour from 9-10am PDT (1600 UTC) Tuesdays on our IRC 
channel, which is at irc.freenode.net#ns-3.

Furthermore, this coming Friday the 26th, from 1230-1430 PDT, I am 
planning to host a conference call for discussing ns-3.2 release 
feedback and for identifying priorities and schedule for the ns-3.3 
release.  Let me know if you would like to participate in this 
discussion; otherwise, I'll post some notes to the list later.  Feedback 
via email is always welcome.

Thanks,
Tom