On the need for system-level support for ad hoc and sensor networks
Authors:
Open Access
- 1 April 2002
- journal article
- research article
- Published by Association for Computing Machinery (ACM) in ACM SIGOPS Operating Systems Review
- Vol. 36 (2) , 1-5
Abstract
Ad hoc and sensor networks are an important, emerging niche that is poorly supported by existing operating systems. In this paper, we argue that network-wide energy management is a primary concern in ad hoc networks, and that this functionality is best provided by a systems layer. We are currently designing and implementing a distributed, power-aware, adaptive operating system, called MagnetOS, specifically targeting ad hoc and sensor networks. MagnetOS provides a single system image of a unified Java virtual machine across the nodes that comprise an ad hoc network. By automatically and transparently partitioning applications into components and dynamically placing these components on nodes within the ad hoc network, our system reduces energy consumption, avoids hotspots and increases system longevity. We show that a systems approach to automatic object placement in an ad hoc network can increase system longevity by a factor of four to five.Keywords
This publication has 15 references indexed in Scilit:
- Building efficient wireless sensor networks with low-level naming2001
- System architecture directions for networked sensors2000
- Quantifying the energy consumption of a pocket computer and a Java virtual machine2000
- The mobiware toolkit: programmable support for adaptive mobile networkingIEEE Wireless Communications, 1998
- A dynamic disk spin-down technique for mobile computing1996
- Safe and efficient sharing of persistent objects in Thor1996
- Fundamental challenges in mobile computing1996
- Rover1995
- Highly dynamic Destination-Sequenced Distance-Vector routing (DSDV) for mobile computers1994
- Fine-grained mobility in the Emerald systemACM Transactions on Computer Systems, 1988