Multi-Application Deployment in Shared Sensor Networks Based on Quality of Monitoring

• Main Authors: Bhattacharya, Sangeeta, Saifullah, Abusayeed, Chenyang Lu, Roman, Gruia-Catalin
• Format: Conference Proceeding
• Language: English
• Subjects: Adaptive systems; Costs; Monitoring; Physics computing; Resource management; Sensor phenomena and characterization; Sensor systems and applications; Spread spectrum communication; Virtual machining; Wireless sensor networks
• ISSN: 1545-3421; 2642-7346
• DOI: 10.1109/RTAS.2010.20

Wireless sensor networks are evolving from dedicated application-specific platforms to integrated infrastructure shared by multiple applications. Shared sensor networks offer inherent advantages in terms of flexibility and cost since they allow dynamic resource sharing and allocation among multiple applications. Such shared systems face the critical need for allocation of nodes to contending applications to enhance the overall Quality of Monitoring (QoM) under resource constraints. To address this need, this paper presents Utility-based Multi-application Allocation and Deployment Environment (UMADE), an integrated application deployment system for shared sensor networks. In sharp contrast to traditional approaches that allocate applications based on cyber metrics (e.g., computing resource utilization), UMADE adopts a cyber-physical system approach that dynamically allocates nodes to applications based on their QoM of the physical phenomena. The key novelty of UMADE is that it is designed to deal with the inter-node QoM dependencies typical in cyber-physical applications. Furthermore, UMADE provides an integrated system solution that supports the end-to-end process of (1) QoM specification for applications, (2) QoM-aware application allocation, (3) application deployment over multi-hop wireless networks, and (4) adaptive reallocation of applications in response to network dynamics. UMADE has been implemented on TinyOS and Agilla virtual machine for Telos motes. The feasibility and efficacy of UMADE have been demonstrated on a 28-node wireless sensor network testbed in the context of building automation applications.