The motor part of a crawling eukaryotic cell can be represented schematically as an active continuum layer. The main active processes in this layer are protrusion, originating from non-equilibrium polymerization of actin fibers, contraction, induced by myosin molecular motors, and attachment due to active bonding of trans-membrane proteins to a substrate. All three active mechanisms are regulated by complex signaling pathways involving chemical and mechanical feedback loops whose microscopic functioning is still poorly understood. In this situation, it is instructive to consider the problem of finding the spatial organization of standard active elements inside a crawling layer ensuring an optimal cost-performance trade-off. If we assume that (in the range of interest) the energetic cost of self-propulsion is velocity independent, we obtain, as an optimality criterion, the maximization of the overall velocity. We choose a prototypical setting, formulate the corresponding variational problem and obtain a set of bounds suggesting that radically different spatial distributions of adhesive complexes would be optimal depending on the domineering active mechanism of self-propulsion. Thus, for contraction-dominated motility, adhesion has to cooperate with ‘pullers’ which localize at the trailing edge of the cell, while for protrusion-dominated motility it must conspire with ‘pushers’ concentrating at the leading edge of the cell. Both types of crawling mechanisms have been observed experimentally.