As energy infrastructure is upgraded to meet modern digital society's need for secure and high-quality electric power, grid-tied energy storage systems are becoming a viable means for improving system reliability and realizing the benefits of Smart Grid. This paper investigates the optimal implementation of distributed storage resources in a power distribution system or islanded microgrid in conjunction with an intelligent load shedding scheme to minimize the societal costs of blackouts. First, a non-linear combinatorial optimization problem for prioritizing outage ride-through service is formulated and a solution methodology is proposed that combines a local search heuristic with simulated annealing. Next, a resource allocation optimization problem is developed to determine the optimal mix and placement of energy storage resources in a power distribution network for intentional islanding of sectionalizable sub-networks. Simulations are performed on the IEEE 123 node test feeder model subject to realistic power system constraints.