The Internet of Things is at the stage of development in which everyone is talking about the wonderful things it will do for the enterprise and the digital economy as a whole – what Gartner calls the Peak of Inflated Expectations.
We all know what comes next, of course: the Trough of Disillusionment where both the realities of implementation and more candid views of ultimate capabilities lead people to overinflate the risks and underappreciate the rewards.
In recent weeks, it seems we have begun to see the first trickle of disillusionment creep into the IoT. It has a lot to do with the substantial networking challenges that such a bold endeavor requires.
According to Andrew Hay, director of research for web security firm OpenDNS, IT departments across the board are still in the dark as to what is about to hit the enterprise. With smartcams, sensors and a host of other connected devices already pinging data to who knows where, few organizations fully appreciate the scope at which potentially critical data is being offloaded to non-secure assets on a regular basis. In many cases, employees are taking it upon themselves to deploy tools like Dropcam and Nest into the enterprise data environment without fully comprehending how they treat data flows and with little or no documentation from the manufacturers.
Security is always a serious concern in advanced network architectures, but there are a number of operational issues that have yet to be fully worked out as well. One is the development of real-time functionality, according to TTech’s Dr. Markus Plankensteiner. This is a particular concern in industrial settings, where split-second timing is crucial for machine-to-machine (M2M) performance. The IEEE currently has a Time-Sensitive Networking (TSN) task force working on the issue, focusing on an open standards-based, high-bandwidth solution that provides reliable and deterministic M2M and machine-to-cloud performance within a reasonable cost envelope.
Addressability will be another important feature of IoT networking, says Enterprise Networking Planet’s Julie Knudson. With virtually everything we touch acting as a tiny data point, keeping their identities straight will be a huge undertaking, and there has been little action on this front from the tech community. A key problem is a lack of standards when it comes to addressability, leaving manufacturers to devise their own rules. Without broad standards, even automated management systems will have trouble keeping track of the constantly shifting relationships between devices.
The good news is that leading networking vendors are at least trying to sort out a workable IoT stack, although their efficacy cannot be determined until they are deployed at scale. Huawei recently released the Agile Network 3.0 architecture, featuring the Agile Internet of Things solution, which purports to deliver full digitization of production, manufacturing and logistics. The framework is built around the LiteOS operating system, which incorporates functions like auto-discovery and auto-networking within a mere 10 kB footprint. Its small size and the fact that it requires no configuration allows it to oversee everything from automobiles to wearables. Huawei has even opened it up to developers to foster an IoT application ecosystem.
The Internet of Things will not pop up overnight, of course, but industries across the spectrum are ramping up plans for a world of widespread connectivity. It would be a shame if such a promising development were to be derailed simply because initial deployments failed to address the impact that dynamic scale will have on basic networking functions.