Impala: A Middleware System for Managing Autonomic, Mobile, Wireless Sensor Networks

Опубликовано 6 сентября 2016, 5:16

Impala: A Middleware System for Managing Autonomic, Mobile, Wireless Sensor Networks Sensor networks are computer systems with many sensing/compute nodes working to gather information about their environment, process and fuse that information, and in some cases, actuate control mechanisms in response. Sensor networks operate under severe energy constraints that call for very lightweight software layers. On the other hand, sensor networks are long-running systems, intended to operate in situ, with minimal direct human intervention, for months or years. This requires support for adaptive operations and ease of upgrades over the life of the system. In response to these opposing needs, this work has investigated middleware layers for sensor network systems. In particular, in this talk I will discuss Impala, a middleware layer developed for ZebraNet. ZebraNet is a mobile, wireless sensor network aimed at improving tracking technology via energy-efficient tracking nodes and store-and-forward communication techniques. Impala acts as a lightweight operating system, but also has been designed to encourage application simplicity, modularity, adaptivity, and reparability. Impala provides an interface for on-the-fly application adaptation in order to improve application performance, energy-efficiency, and robustness. Impala allows application updates to be received via the node's wireless transceiver and to be applied to the running system dynamically. Impala adopts an operation scheduling mechanism and an event handling model to capture the regular and irregular aspects of system activities. Impala unifies media access control and transport control into an efficient network protocol that supports a broad taxonomy of sensor network communication styles. In the context of Impala implementation on ZebraNet hardware nodes, I discuss how system constraints and goals led to the layering and interface designs in the system architecture. Overall, the contributions of this work include: 1. Established a system architecture with minimal layering and clean interfaces. The entire system consumes 16KB code memory and 320B data memory. 2. Implemented efficient regular operation scheduling and irregular event handling. The typical overhead is at 100's cycles. 3. Developed a unified network protocol to provide a broad range of message models. Packet header overhead is about 10 and update strategy saves software transmissions by 85. 5. Provided quantitative basis for justifying the needs for a special middleware layer in building efficient sensor network systems.