Developing applications for ad-hoc and sensor networks poses significant challenges. Many interesting applications in these domains entail collaboration between components distributed throughout an ad-hoc network. Defining these components, optimally placing them on nodes in the ad-hoc network and relocating them in response to changes is a fundamental problem faced by such applications. Manual approaches to code and data migration are not only platform-dependent and error-prone, but also needlessly complicate application development. Further, locally optimal decisions made by applications that share the same network can lead to globally unstable and energy inefficient behavior. In this paper we describe the design and implementation of a distributed operating system for ad-hoc and sensor networks whose goal is to enable power-aware, adaptive, and easy-to-develop ad-hoc networking applications. Our system achieves this goal by providing a single system image of a unified Java virtual machine to applications over an ad-hoc collection of heterogeneous nodes. It automatically and transparently partitions applications into components and dynamically finds a placement of these components on nodes within the ad-hoc network to reduce energy consumption and increase system longevity. This paper outlines the design of our system and evaluates two practical, power-aware, online algorithms for object placement that form the core of our system. We demonstrate that our algorithms can increase system longevity by a factor of four to five by effectively distributing energy consumption, and are suitable for use in an energy efficient operating system in which applications are distributed automatically and transparently.