We have analyzed the design windows for laser fusion power plants based on direct-drive fast ignition concepts and have examined the issues of chamber technologies and the feasibility of a small laser fusion experimental reactor suitable for developing their power plants. Target gain curves are assessed for power plants having 90- to 200-MJ fusion yields with 600-kJ to 1-MJ lasers, and for an experimental reactor [the laser fusion experimental reactor (LFER)], having a 10-MJ fusion yield with a 200-kJ laser, i.e., 100 kJ for implosion and 100 kJ for heating. The fast ignition LFER can produce its fusion output approximately one order of magnitude smaller than that of the central ignition design, so that we can use a rather small solid-wall chamber for the first stage of the LFER operation. We can also expect to decrease laser cost drastically, although for the heating laser we must develop a long-life final optics system. Using fast ignition direct-drive targets, we could design a smaller ~300-MW(electric) reactor, with 200-MJ fusion pulse energy and 4-Hz repetition rates. The smaller pulse energies mitigate pulse loads on the chamber walls and the final optics; then, we can flexibly design large 1200-MW(electric) modular plants by using multiple reactor modules. We identified the issues of liquid-wall and solid-wall chambers and proposed basic reactor concepts for a power plant (KOYO-Fast) and an experimental reactor using fast ignition direct-drive cone targets.