A "roaming network" is a wired network. You are simply connecting additional wireless access points to your wired LAN. This is a completely normal way of implementing what you seek.
It is preferable to have "stationary" devices wired, since wireless is subject to a variety of phenomena that can result in unreliable performance. This is especially true for streaming video.
And how should I configure from the modem into the network?
The usual setup is
Cable from the overhead line > ISP's modem > Extreme > Ethernet switch > wired devices (AppleTV, VoIP, desktop Macs, etc).
The other bandwidth - demanding use will be VoIP. Make it part of your wired LAN as well. Both Vonage and Ooma insist that you connect their supplied interface box directly to your modem. The reason is that they do not want customers complaining about poor voice quality; interruptions and dropouts that make conversation difficult. Installing it in the manner they recommend gives VoIP traffic precedence over the remainder of your network, by taking all the bandwidth it wants for phone calls and parceling out the rest. However, you don't have to install it this way and it will work if you choose to install their box after your router. Just remember that if you are streaming movies, downloading files, and performing other network-demanding tasks your phone call audio may begin to stutter.
For these reasons Vonage and Ooma will want you to place their device immediately after the modem, and use it for your router. This means your Extreme will need to be configured as a bridge, preventing you from using many of its features. In that case your installation will look like this:
Cable from the overhead line > ISP's modem > VoIP > Ethernet switch > wired devices (all of the above, plus your Extremes, now configured as bridges, which will serve as wireless access points).
Don't bother with an Ethernet hub. Switches are preferable and have supplanted hubs. They have become very inexpensive.
