The setup is relatively simple but to do this as one shot you will need space. An awful lot of space. Doing this indoors in a regular sized house is not going to work, the walls/ceiling are going to reflect light back filling in the shadows and you wont get the fading-into-black effect.

The easiest way to shoot this (short of renting a studio) would be to use a large outdoor space at night (gardens work well for this as you can get power from indoors with an extension lead).

Apart from that you just need a single large softbox positioned in front and to the side. It mustn't be too close (you wont be able to use the trick of having a small softbox close to the subject in place of a large one) otherwise the closest and furthest family members will receive different amounts of light and one will be over/under exposed.

Shooting from farther away with a telephoto lens is also required, not for the lighting but to ensure the faces remain "in scale" with each other by compressing the perspective. The other alternative is to shoot each person individually and composite the shots as Dan mentions. It's possible to do this in camera, and it's impossible to tell just by looking how the original was achieved.