The term 'star-wiring' is a bit misleading as to the actual goal here. That goal is to insure that any block of track only has one path for power, and more importantly the DCS signal, to flow through. It doesn't matter if you use home runs to a central location, or a bus that delivers power through feeders, as long as any section of track only has power supplied from one place. Basically, you want to use center rail isolators every 10-12 track joints and run only one feeder to each of these blocks.
You can also access pretty much every feature the cab1L offers through the DCS remote if you connect the systems with a serial cable.
On to less technical information: (my opinion only)
If you have conventional engines, and want to run them through a command system, I think the DCS system offers the most bang for your buck when getting started, however... Over all, I find the TMCC/Legacy systems to be better made, more adaptable, and use superior (if still obsolete) technology. The command protocol for TMCC/Legacy is also available to the public allowing folks to build their own devices that can talk to the command system if they wish to. This is not a big deal for some folks, but for me, not being able to speak to the DCS system is a deal breaker on ever purchasing it... At least at retail cost, I'd pick one up if the deal was good enough.
The question to ask is what do you expect out of the system? If you are looking to run conventional engines, and you already have the cab1L/base, you can add a legacy powermaster or use a ZW-C to get good control of your conventional engines for less money. On the other hand, if you own, or plan to own DSC PS2/3 engines, you'll want the DCS system to have control of all their features. For me, I have no desire to purchase PS2/3 engines over TMCC/Legacy ones, so no need for a DCS system.
JGL