I would add that in my experience, the biggest challenge for most people in thinking about these issues is in moving from what they know about Wi-Fi (which is a star topology) to mesh topologies like Zwave and zigbee home automation.
Mesh is designed for different goals and works very differently than what you will be used to from Wi-Fi. You don’t have constant monitoring of everything and you don’t have fixed routes. And you don’t have a massive computer brain (speaking relatively) figuring everything out.
Mesh was designed to be inexpensive, both in the devices and in operation. The first real practical uses were for massive sensornets with 1000 or more sensors. The goal was absolutely minimal energy use to prolong the battery life. The messaging model was very tiny messages sent infrequently.
For example, you would typically check battery status once a week, not even once a day, and certainly not in real time or once a minute.
This kind of model tends to map really well to both sensor operations and lighting operations as long as you keep human beings out of the mix.
As soon as you add smart phone applications, people tend to think they should have real time status of everything that’s happening, and it drives many of them crazy if they don’t. Consequently, they will keep trying to force a mesh model into a WiFi paradigm, and it’s really not a good fit.
So if you are interested in all these technical details just clear your mind of anything that you know about Wi-Fi or other kinds of star or tree networking and just start over. Little tiny, not very smart, devices sending little tiny messages at very low power over very short distances. And everything designed to reduce the total cost of the network.
It’s really cool technology and lots of fun to work with, once you get yourself out of the Wi-Fi mindset.