CoopStor: a cooperative reliable and efficient data collection protocol in fault and delay tolerant wireless networks
- PDF / 1,577,009 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 22 Downloads / 255 Views
(0123456789().,-volV)(0123456789(). ,- volV)
CoopStor: a cooperative reliable and efficient data collection protocol in fault and delay tolerant wireless networks Georgios Z. Papadopoulos1 • Alex Mavromatis2 • Antoine Gallais3 • Fabrice Theoleyre4
Ó Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Internet of things consist in the deployment of constrained and battery-powered devices with a radio interface. Most industrial applications require to respect strict reliability and delay constraints. Lossy links imply mechanisms to retransmit the packets to improve the reliability. However, transient faults (e.g. obstacles, interference) may also impact the reliability, and require to change the routes. In this article, we present cooperative storing mode (CoopStor), that detects the faults and triggers action before a packet is dropped. In particular, the packets are cooperatively stored in the network, until the routing protocol re-converges. The congested zone gradually increases to serve as a local storage facility. CoopStor relies on a selection of relay and leaf nodes to reduce the energy consumption while providing a low end-to-end delay in steady state. Our performance evaluation campaign highlights a reduction of packet drops and an improved energy-efficient data collection procedure. Keywords Fault-tolerance Congestion Distributed data storage Reliability Queuemanagement Storage nodes Synchronized MAC
1 Introduction More and more smart objects are now connected to the Internet, to create the so-called Internet of Things (IoT). Typically, smart buildings rely heavily on collecting a large volume of data in real-time [30]. The sensors send
& Fabrice Theoleyre [email protected] Georgios Z. Papadopoulos [email protected] Alex Mavromatis [email protected] Antoine Gallais [email protected] 1
IMT Atlantique, IRISA, Rennes, France
2
Faculty of Engineering, University of Bristol, Bristol, UK
3
LAMIH Laboratory, CNRS, Polytechnic University Hautsde-France, Valenciennes, France
4
ICube Laboratory, CNRS, University of Strasbourg, Pole API, Boulevard Sebastien Brant, 67412 Illkirch Cedex, France
their measurements to a cloud infrastructure, through a border router. Most IoT networks exhibit a convergecast traffic pattern, where all the packets are forwarded to a border router. If the routing protocol balances the load around the border router, many nodes have to forward packets, resulting in a high contention [3]. It is rather more efficient for the link layer to select only a small set of nodes to forward most of the traffic [33]. This way, we reduce the number of collisions and, thus, increase of network reliability and energy efficiency. Hence, energy consumption is balanced by changing the role (relay / passive) of the nodes regularly through the routing protocol. However, the routing protocol has also to construct efficient routes on top of a lossy topology. In particular, external interferences and collisions require to retran
Data Loading...
