The CERN linear collider (CLIC) [W. Schnell, Proc. Workshop on Physics of Linear Colliders, Capri, 1988, eds. l. Palumbo, S. Tazzari and V.G. Vaccaro (1989) p. 345] is a two-beam-accelerator (TBA) scheme in which the driving beam consists of an intense 3 to 5 GeV electron beam bunched at 30 GHz. One possible way to produce this drive beam is to start with a low-energy (or order 10 MeV), high-current (about 5 kA) beam from an induction linac. In passing through a wiggler, this beam is bunched at 30 GHz into micropulses, each with about 10 12 electrons. To construct the format required for the CLIC drive beam, the bunched beam is subsequently chopped at 350 MHz. It is then accelerated to 3 to 5 GeV in an rf linac driven by conventional, low-frequency klystrons. Rf power is extracted at the bunch frequency of 30 GHz and fed into high-gradient structures to accelerate electron or positron beams to TeV energies. The drive beam is repeatedly reaccelerated in 350 MHz superconducting cavities. This study examines the design trade-offs of the proposed “front end” of the CLIC TBA, the linear induction accelerator and FEL. We examine the relevant figure of merit, the efficiency of bunching, as a function of beam energy, current, and emittance, and we consider the effects of wiggler errors, energy spread and slew, and beam misalignment.