It is nearly impossible to overestimate the potential benefits that the Internet of Things (IoT) offers. There is another thing that is just as difficult to overestimate: The challenges the IoT presents to network engineers, planners and the organization for which they work.
Two interrelated problems are high on the list of roadblocks for the IoT. The first is that each of the almost limitless number of endpoints that eventually will be deployed collectively poses significant security threats. The other is that many of these small and low-power endpoints require efficient ways to very quickly reach computing resources.
Both obstacles are daunting. Luckily for IoT proponents, however, an increasingly established platform offers some answers. Software-defined networks (SDNs) are, in many ways, a natural fit. "The popularity and proliferation of IoT devices simply adds to the scale of the SDN’s purpose, personality and workload," wrote Rumus Sakya, Edgewater Networks' senior vice president of Engineering, in response to emailed questions from IT Business Edge.
Networking executives see the close tie between the IoT and SDN. The reason is simply that the core of SDN is the separation of data payloads from the data controlling where those messages go. Disentangling these two levels of data goes a long way toward solving two of the IoT's most vexing problems. The two are "rolling out at the same time because IoT needs SDN," wrote Ken Hosac, Cradlepoint's vice president of IoT Strategy and Business Development, in response to questions from IT Business Edge.
Securing the IoT
On the security front, a universe of IoT devices, such as those controlling an organization's autonomous vehicles, for instance, can be segregated onto an SDN network that has an especially high level of security. Thus, the network – not each device – is the point at which the major security steps are taken.
This is a far more efficient approach that is made practical using SDN. "For security reasons, IoT devices are best deployed on their own networks," Hosac wrote. "However, it’s not always possible or practical to have a separate physical network for many IoT deployments. SDN provides the ability to create private virtual networks that overlay physical networks without all of the complexity associated with legacy VPN technologies."
The other key capability SDNs provide to the IoT is access to computing power. Many IoT applications are time-sensitive. Using computing assets in the data center in many cases won't work. In other scenarios, the results may be suitable but the process inefficient and needlessly expensive. In both cases, it pays to have the computing near where it is being used. "Some IoT devices, such as autonomous vehicles, require a real-time stream of data that needs real-time decision making," Hosac wrote. "SDN components dealing with such IoT devices cannot tolerate latency or jitter. Therefore, IoT supporting SDN infrastructure may have to be located very closely to the IoT farm, hence what is beginning to be known as fog computing to support such IoT deployment."
Migrating to the Edge
The IoT, therefore, is one of the catalysts for the migration of computing power to networks' edge. A key challenge, according to NS1 CEO Kris Beevers, is enabling edge computing assets to communicate fluidly with IoT endpoints. "Alongside these new architectures come significant internetworking challenges: How can highly distributed and dynamic application deployments interconnect effectively while remaining secure, optimizing application delivery performance, and responding to shifting workloads and noisy network conditions?"
IoT devices use a well understood core internet standard, the domain name system (DNS), to find the computing resources they need. This now will be done in conjunction with SDN. SDNs are an inherently local-area network (LAN)-based technology, he wrote. Combining its capabilities with DNS gives it the capabilities of operating in what he calls" distributed and dynamic topologies" that characterize ever-evolving and shifting wide-area networks (WANs).
It's clear that SDNs and the IoT are joined at the hip. It seems, however, that SDN, which already is being deployed deep within carrier networks, will thrive. The IoT, however, may not find as great success if it doesn't team with SDN.
The combination of SDN and the IoT is gaining traction. "Most communication service providers are in the process of deploying SDN in 2018 for specific use cases," wrote Kevin Wade, the senior director of solutions for Ciena Blue Planet. "These early deployments are helping them modernize their operational processes, which is a necessary step to prepare for widespread IoT implementations, so in our view, SDN lays the foundation for IoT networks to be implemented."
Sakya wrote that the IoT and SDN "absolutely" should be rolled out in tandem. "SDN has to be part of the management infrastructure for the IoT deployments because the nature of vast scale will require efficient and automated on-boarding, provisioning, data collection, analytics and interfaces to third-party applications APIs needing such data," he wrote. "SDN will be needed also for the most important aspect of health and security monitoring of the IoT farm."
Carl Weinschenk covers telecom for IT Business Edge. He writes about wireless technology, disaster recovery/business continuity, cellular services, the Internet of Things, machine-to-machine communications and other emerging technologies and platforms. He also covers net neutrality and related regulatory issues. Weinschenk has written about the phone companies, cable operators and related companies for decades and is senior editor of Broadband Technology Report. He can be reached at firstname.lastname@example.org and via twitter at @DailyMusicBrk.