Progressive network re-programming model for wireless sensor networks

Abstract

A sensor network consists of from several sensor nodes to tens of thousands of sensor nodes. Basically each sensor node contains simple software to sense surroundings and communicate with other sensor nodes. However, for more complex applications, more advanced technologies, such as operating system, layered network stack, localization, and time-synchronization, considering a sensor node as a system have been developed. It is nature to modify and update the software to fix bugs, and adapt the changes of surroundings and software requirements. Therefore, network re-programming in wireless sensor networks is an essential facility, especially in practice, for deploying and maintaining highly dynamic sensor nodes. ANTS-PNP, or Progressive Network re-Programming, proposes an efficient way to update whole binary image based on the layered architecture with multi-hop dissemination protocol. The layered architecture consisting of four layers including Application, Buffering, Storage Management and Bootloader layers takes a role of storing a new binary image downloaded from other nodes and guaranteeing its integrity. It utilizes 3 different memories including flash, RAM, and EEPROM to store a new binary image in faster and more efficient manner, but it requires storage management scheme for efficient, safe use of the memory. Dissemination protocol of PNP for multi-hop environment is based on hierarchical routing protocol providing information of parent/child nodes. It forms a so-called ‘Reprogramming Network`` to update only sensor nodes interested in the new binary image. This leads to low-power consumption for the network re-programming. Moreover, it makes it possible for all sensor nodes to perform periodic jobs even during the update. We implement ANTS-PNP targeted for ANTS-H20 sensor node hardware. In addition the implementation is done with full functional ANTS platform including operating system, network stack with MAC and routing protocol, and real appli...