This article is part two of the previous article on M2M to IoT, and the adaptation of LTE where we introduced the concept of Machine-to-Machine (M2M) and its expansion to the Internet of Things (IoT).
The IoT has the potential to have transformational impacts on our society and the way in which we provide health, utilities, transportation, and the various other services required of a modern society. Whilst it is easy to appreciate the benefits of the IoT – estimates range from $7.1 to $19 trillion in global GDP by 2020 – the concept presents a number of architectural and technological challenges for the underlying networking, computing and storage platforms, which are not nearly as well understood.
The focus of this article will be on the challenges IoT poses to the networking platforms, particularly the wireless platforms which will underpin it. These challenges broadly include:
- capacity and spectrum
- latency and virtualisation
- connectivity and mobility
- convergence and security
- energy and environment
The issue of capacity and spectrum are closely related in that the aggregate data rate of the radio or “air” interface is constrained by the modulation (better signal quality equals more capacity) and the bandwidth (more bandwidth equals more capacity). The modulation methods used today are complex and nearing the theoretical limits (the Shannon capacity) consequently any advancement in the near future is likely to be relatively minor through increased spatial multiplexing ie MIMO. The most significant improvement in capacity is likely to come from the introduction of higher frequency bands – eg 45 GHz, 60 GHz – and the aggregation of disparate frequency bands ie carrier aggregation.
The requirement for reduced latency is not only driven by the end-user experience (real-time applications), but by the drive to adopt cloud-based radio access network architectures. A cloud-based architecture enables the dynamic dispatch of radio resources where and when they are required in the network ie front-haul, radio access virtualisation and CoMP. This means network hardware can be shared or pooled, and the provisioning and optimisation of network resources can be automated.
In the context of M2M and IoT, the question of connectivity and mobility is probably the most poignant. Whilst the total amount of data likely to be generated by IoT presents somewhat of a challenge for our wireless platforms, the amount of data required of each discrete intelligent device is far less troubling – it may only be in the order of bits per second. The concern here, rather, is the sheer number of end-devices requiring seemingly simultaneous connectivity while on the move – as many as one mobile end-device per square metre! Additionally, these end-devices will need to be able to ‘roam’ seamlessly across various private and public networks while retaining a connection to their ‘home’ networks, services and applications.
The concept of IoT encompasses a range of intelligent end-devices, some being critical eg intelligent transport and others being non-critical eg connected home. Whilst these services could be delivered on independent platforms, it is not resource (ie spectrum, power, space, hardware) efficient nor cost effective to do so hence the drive to ‘converge’. To enable the convergence of critical and non-critical services onto a common platform however requires stringent quality of service (QoS) and security (ie tunnelling protocols, encryption) measures. These QoS and security measures need to be present to ensure reliable performance of applications, particularly across untrusted and congested network environments.
The final challenge is that of energy efficiency and environmental impact. To connect to the IoT, each intelligent end-device will require a radio transmitter of some form which consumes power proportional to its bandwidth and modulation order. Furthermore, a number of these end-devices being mobile, will not be connected to the electrical grid, and will therefore need to source power via a combination of micro fuel cells, batteries, photovoltaics and wireless power transmission.
These challenges are certainly not insurmountable, though will require further research and development into the areas of 5G, High Efficiency Wireless Local Area Networks (WLANs), Heterogeneous Networks (HetNet), and Software Defined Networking (SDN).
The next instalment on “M2M to IoT, and the adaptation of LTE” will expand on how these areas of research and development seek to address the challenges presented in this article.