No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

IoT Starts with 'Network'

The Internet of Things (IoT) will certainly be making the headlines this year, and networking pros should be upping their skills to play a meaningful role in this burgeoning revolution. Whether they realize it or not, IoT developers will desperately need the type of expertise only networking pros can provide if they are to have any hope of success.

 

IoT is such a broad category I try to describe it as a concept or an idea rather than any specific thing. The trouble is that the simple term "IoT" covers an enormously wide range of potential applications that hold the promise to deliver new capabilities for providing better, faster, and cheaper products and services and addressing bigger issues like energy management, environmental monitoring, and improved health care.

While civilians can swoon over visions of IoT bliss, we need to remember that "IoT" starts with "Internet" -- in other words, a network. So when all of the furor settles down and the time comes to deploy IoT applications, at least some of us who work in networking will be involved in building these things.

 

 

I became involved in IoT early on because the nature of the applications being described for IoT assume mobility, and mobility requires wireless (or at least a very long extension cord). While some applications will be served more cheaply and reliably over a wired connection, wireless will undoubtedly be key to IoT.

Two primary wireless technologies are lining up to provide the foundation for mobile-enabled IoT applications: cellular and Wi-Fi. While a variety of shorter range or specialized wireless technologies, like Bluetooth, might come into play for connecting IoT devices on a local basis, the backbone network that connects everything to the central intelligence, be it cloud- or premises-based, will be cellular or Wi-Fi.

The vast majority of wide-area IoT applications operating today make use of the cellular infrastructure. Of course, with mobile operators increasingly opting for Wi-Fi offloading, voice over Wi-Fi, and other strategies that incorporate Wi-Fi, the technology picture gets somewhat muddled. However, a mobile carrier is still operating the overall network.

Up until now, mobile operators have supported IoT devices over the same networks and the same cellular services they use to serve smartphones and cellular-equipped tablets and laptops. This has been a great deal for the operators, as they get millions of new endpoints and really don't have to invest much in terms of infrastructure to support them.

However, with the smartphone market nearing saturation while the IoT potential grows, mobile operators now consider meeting IoT requirements as a key factor in their service planning. The adoption of 5G is one indication of this.

5G as IoT Enabler
While many industry watchers have focused on the raw bandwidth improvements 5G promises to deliver, other key advances of the technology will be critical for IoT. In particular, 5G protocols are designed to support far greater device density (i.e., one million devices per square kilometer) and very low latency (e.g., one millisecond) services for applications like autonomous vehicles.

However, the operators aren't just waiting for 5G to start pushing into IoT. While our initial wave of wide-area wireless IoT apps has depended on existing 3G, 4G, and even earlier cellular data technologies, the carriers are now starting to roll out services based on low-power WAN (LPWAN) technologies designed specifically for IoT applications.

Based on my research with the carriers, most are now rolling out, or at least planning to offer, services based on a standard called LTE Cat M1 (also referred to as eMTC) that supports data rates up to around 200 to 350 Kbps. AT&T, Verizon, and T-Mobile have already rolled those services out nationwide, and Sprint is starting trials in the first quarter of this year with plans for deployment in mid-2018. Besides being economical to deploy, Cat M1 supports long battery life, a key requirement in many IoT applications.

Trailing not far behind is another IoT technology, called NarrowBand IoT (NB-IoT). NB-IoT runs on a much narrower frequency band than Cat M1 (180 KHz versus 1.08 MHz) and should support data rates roughly half of what Cat M1 can deliver. T-Mobile, Verizon, and Sprint are planning NB-IoT trials in the 2018 timeframe, and T-Mobile is talking about a nationwide launch in mid-2018. AT&T is being a bit more cautious in its approach to NB-IoT, but will likely take the plunge once it sees everyone else is in the pool.

A number of other LPWAN IoT transmission technologies have been announced, many of them from private companies rather than standards bodies. Those would include technologies like SigFox, LoRa, and Nwave, but the carriers don't appear to be interested in looking outside the familiar 3GPP-defined standards, which essentially ensure a high level of integration with their existing network technologies.

As the carriers roll out these new IoT services, we fully expect the equipment manufacturers to introduce compatible IoT hardware. However, a lot of that will depend on the chip manufacturers bringing out the components that will drive those devices.

Networking Needs a Say
What this all means for enterprise IT is that we will have a new selection of network services to evaluate as part of any IoT applications our companies intend to deploy, and we'll have to factor in the relative cost of the devices available for each network service option.

While cellular has the advantage of a massive infrastructure for macro deployment, Wi-Fi will still have a role. Today, virtually every residential IoT application from smart thermostats to video doorbells depends on home Wi-Fi, and there could be a place for Wi-Fi in the wide area as well. That configuration assumes a metro-area Wi-Fi deployment.

In the commercial space, Cisco has pile of IoT application case studies posted on its website. The vast majority of these depend on Wi-Fi, or in some cases a Wi-Fi router connected via a cellular modem. Among those are case studies from Kansas City and Copenhagen, both of which use metro-area Wi-Fi deployments to support a range of IoT applications along with traditional public Wi-Fi access.

And the Wi-Fi Alliance has introduced an IoT-oriented version of its radio link, dubbed HaLow (IEEE 802.11ah), that operates in the 900-MHz band and could support data rates from 650 Kbps to 234 Mbps. Thus far there appears to be little market interest in HaLow.

In short, it's time for network pros to start getting real about IoT. If IoT applications are going to be successful, network professionals will have to be part of the design team. The networking input would not just be limited to specifying, evaluating, and selecting the particular IoT transport to use, but also to provide guidance into the other requirements of networking like monitoring usage and performance, troubleshooting, maintaining records of device types, software/firmware releases, locations, providing software updates, and all of the other management functions needed to maintain and operate any network.

The foundation of any IoT application is going to be the network, and network professionals will be needed to provide that expertise. Any network pro with the least bit of ambition should want to get involved with what promises to be the next big revolution in networking.

Of course, if you don't have any ambition, I know a guy who installs payphones that might be looking for help...