Chasis_GND question, LCDK, Hawkboard, Beaglebone, SOM-M1

Unfortunately the answer you seek is a lot more complicated than just one right answer.  The solution affects not only ESD, but EMI.  Both ESD and EMI are dependent on other variables like if the case is metal, or what the power supply will be (battery or plug in?), what other external cables connect to the unit, etc.  It's even more complicated than what type of connection you use (resistor, ferrite, etc.), since it actually matters where and how many times it's connected to ground.

The good news is that if you're just trying to make a board that works, it's hard to make it not work with the wrong connections.  The bad news is that describing how to best connect it is beyond the scope of what we can recommend, especially since I'm not an EMI expert, but I have some experience with it, and have heard several experts talk on the subject.

I did find this paper http://www.ti.com/lit/an/snla107/snla107.pdf describing one method that seems to be accepted by several experts as a good way.  It shows a picture of a separate chassis ground plane around the connector and ferrite beads on each side of the ethernet signals (sounds like the hawkboard).  I'm not sure if this will be the best solution for you, but it is one solution that I've seen used more than once.

If you do use this method, know that all the plane shapes (if any exist) underneath the chassis ground also need to be chassis ground and disconnected from the main planes or they will be capacitively coupled to the chassis ground, which is bad.  Also know that the gap from the chassis ground plane to the main planes should be at least 0.050".

I hope that helps.  Like I said, incorrect chassis grounding will most likely not break a working system, but it will make it more susceptible to EMI and ESD problems.