We look at some of the important requirements for making Autonomous Underwater Vehicles (AUVs) more ubiquitous in their role in Oceanographic research in the deep ocean. We show how some of the constraints in working underwater lead to unique solutions for complex multidimensional sets of design possibilities for working with real vehicles in the deep ocean. We identify three components in elaborating on our system optimization concepts -the use of power, sensing strategies, and navigation. We analyse several possible deployments from a theoretical and practical standpoint in which the effects of each of these components interact. We look at the tasks associated with long distance transits, with conducting a sonar survey as well as the task associated with navigating multiple vehicles in the same acoustic network.