Release notes for Diffusion 3.0.3 Beta 1 for WINS NG 2.0
--------------------------------------------------------

In order to compile diffusion, please first look at the config.mk file
and follow steps 1 and 2 listed there. The default configuration will
compile diffusion for the Sensoria nodes using the sh4 cross-compiler
that is available at the BBN website (http://dstl.bbn.com).

In order to use diffusion on the Sensoria nodes, you will have to have
streamd running on your system.  We assume streamd will export the
device files /dev/r1c and /dev/rf2c.  Diffusion will not start if
these files are not present.

Since streamd has not yet shipped, we suggest two work-arounds:

Work-around 1:  (uses just the software originally shipped with the nodes)

If you have just a couple of nodes, you can try changing the BaseSlot
value at the base node using rfcfg in order to get large enough packet
size that will work for your application. After changing the BaseSlot
value, you can check the maximum packet size (MaxDataLength) at the
base node using rfcfg. Once this is done, you will have to create
symbolic links as follows:

bash# ln -s /dev/rf1 /dev/rf1c
bash# ln -s /dev/rf2 /dev/rf2c

Note that while this allows diffusion to run without streamd, it will
not operate properly if the packets sent are larger than the maximum
radio packet size.

Work-around 2:

We have also tried a preliminary version of streamd from Sensoria that
just does fragmentation (what we got is called "fragd").  To start
this, run:

bash# fragd /dev/rf1 /dev/rf1c &
bash# fragd /dev/rf2 /dev/rf2c &


In addition to streamd, diffusion also expects the radios to be
configured as base/remotes and cluster to be configured before it
starts. Use rfcfg to configure each radio with the desired
topology. Note that as of July 12th, the rf radio #1 is flaky and the
Sensoria driver to crashes after a few packets are sent/received. We
expect that this will be fixed in the near future. Meanwhile, we
recommend that radio #1 should be avoided and only radio #2 used.


Please note that this release is pre-configured for the Sensoria
radios.  If you want a version of diffusion to use over ethernet
(emulating the radios), just follow steps 1 and 2 and proceed to the
README file.
