Network lifetime maximization in wireless mesh networks for machine-to-machine communication

Abstract

In this paper we present new optimization formulations for maximizing the network lifetime in wireless mesh networks performing data aggregation and dissemination for machine-tomachine communication in the Internet of Things. We focus on heterogeneous networks in which multiple applications co-exist and nodes may take on different roles for different applications. Moreover, we address network reconfiguration as a means to increase the network lifetime, in keeping with the current trend towards software defined networks and network function virtualization. To test our optimization formulations, we conducted a numerical study using randomly-generated mesh networks from 10 to 30 nodes, and showed that the network lifetime can be increased using network reconfiguration by up to 75% over a single, minimal-energy configuration. Further, our solutions are feasible to implement in practical scenarios: only few configurations are needed, thus requiring little storage for a standalone network, and the synchronization and signalling needed to switch configurations is low relative to each configuration's operating time.

Energy Versus Throughput Optimisation for Machine-to-Machine Communication

“…The main novelty of the presented work consists in introducing an original approach to optimisation of data aggregation and transmission in wireless mesh networks applying multicast packet transfer. The approach considerably extends our previous work described in [1][2][3].…”

“…There is a variety of papers in this field, see, for example, in [4][5][6], but most of them are not intended to consider traffic throughput efficiency and data aggregation in the precise way, as we do in this paper. Those papers (and others) are discussed in [2].…”

“…Note that energy required for signalling or other overheads to set up the routing configuration is not included. This signalling will be different depending on the particular network used: whether it has a connection to the Internet or is standalone, whether only one configuration is used throughout the lifetime of the network or the configuration can be changed during operation [2] and whether the configuration can be preprogrammed before deployment or the network is self-configuring. These are complex issues in themselves, and so here we focus on the operation of a single configuration, and exclude signalling and overhead so as to keep the results general to any specific protocol used.…”

“…In our previous work [1,2], we addressed the problem of determining optimal (data) routing configurations for data aggregation and dissemination in order to minimise the energy usage of the network spent on packet aggregation and transmission. These configurations specify routing of sensor measurements to actuators, as well as which intermediate nodes should aggregate and broadcast data.…”

“…This is because once all sensor measurements have been processed and transmitted, nodes may be able to sleep, thus saving energy. In [1,2], we considered only the energy required for aggregation and data transmission; however, the energy used while the node is idle but awake is also in many cases significant, and thus should be minimised for a complete, energy-efficient solution.…”