Ultra-Low-Latency Wireless Networking for Mission-Critical Applications

Abstract

In a Wireless-sensor network (WSN), low-power sensor devices are connected wirelessly to perform distributed tasks. Such type of networks require low-power wireless networking (LPWN). LPWN provides efficient all-to-all communication and the nodes transmit and receive periodically. Cyber-physical systems (CPSs) are another example of the systems that need LPWN. CPSs also consist of small devices that are connected together to sense the environmental changes, such as heat sensors, motion sensors and light sensors. Both WSN and CPS can also be examples of mission-critical systems. Mission-critical systems require low latency and high reliability. Usually in a WSN and CPS, the devices have a limited source of energy and they rely on small batteries. As the devices are wirelessly connected, they consume most of their energy in radio-transmissions. Traditional LPWN provides centralized processing, and enables all-to-one and one-to-all communications. In such communications, data is first collected at the sink, the sink processes the data and sends the results to all nodes. Such type of communication schemes consume a lot of energy in sending data towards the sink and receiving back from the sink. Chaos is a new LPWN scheme that allows distributed processing of the data. Chaos provides in-network processing, efficient all-to-all communications and a network-wide agreement using Two-Phase Commit (2PC) and Three-Phase Commit (3PC). In this thesis, we design and implement a quorum protocol on Chaos. The protocol enables the Single-Writer/Multiple-Reader (SWMR) and Multiple-Writer/Multiple- Reader (MWMR) protocols using a majority quorum configuration. We implement the protocols on the Contiki operating system using Chaos communication primitives. The SWMR/MWMR protocols act as emulators of the shared-memory systems, and provide highly reliable wireless communications with low-latency. The protocols also significantly improve the energy-consumption in LPWN.