Reliable Radio Access for Massive Machine-to-Machine (M2M) Communication

Machine-to-Machine (M2M) communication is a term that identifies the emerging
paradigm of interconnected systems, machines, and things that communicate
and collaborate without human intervention. The characteristics of M2M Communications are small payloads and sporadic transmissions, while the service
requirements can range from massive number of devices to ultra-reliable. This
PhD thesis focuses on novel mechanisms to meet these requirements in a variety
of wireless systems, from well-established technologies such as cellular networks, to emerging technologies like IEEE 802.11ah.
Today an overwhelming 89% of the deployed M2M modules are GPRS-based.
This motivates us to investigate the potential of GPRS as a dedicated M2M network.
We show that by introducing minimal modifications to GPRS operation,
a large number of devices can be reliably supported. Surprisingly, even though
LTE is seen as the preferable solution for M2M, no mechanisms are in place to
guarantee reliable M2M access. Contrary to mainstream solutions that focus on
preventing overload, we introduce mechanisms to provide reliable M2M service.
We also investigate what cellular networks can do about upcoming smart metering traffic. Finally, we consider the 802.11ah protocol, which represents an
interesting alternative to cellular networks. We propose an allocation mechanism
that matches the dominant operation regime in the network (i.e., periodic
vs. alarm reporting). With the proposed improvements, 802.11ah exhibits outstanding results in terms of efficiency and reliability for massive M2M. The
research in this dissertation shows that reengineering of protocols has a great
potential to adapt the existing system to the new M2M requirements. However,
it also has limitations, that are visible when there is a requirement for a signi-
cantly higher data rate per device or requirement for latency that is lower than
the fundamental limit of the existing system.