Insights from the State of Application Delivery 2017.
An interesting thing happened to the Internet of Things on the way to pwning every other trend as the technology darling du jour: it split itself up.
There are actually two very distinct markets under the “Internet of Things” umbrella. The first is the one we all know and love (to hate) – that of the consumer gadget market. We all have something that qualifies, and if studies are correct we probably have 7.8 things that qualify. These are the home and personal gizmos that delight, amaze, and confound us. Whether it’s related to home automation or personal fitness, the consumer consumption rate for things shows no signs of stopping.
The other, less sexy and less often mentioned, market is the Industrial Internet of Things (IIoT). The distinction is not only important but necessary, as the uses and challenges associated with the industrial side of things are very different from those on the consumer side. While availability is always a concern for both, it becomes critical in the industrial market as business operations and manufacturing become almost wholly reliant on the data being generated by things and acted on by increasingly smart systems performing millions of decisions a day.
Performance, too, becomes an overwhelmingly important characteristic of the systems and apps that support the IIoT. When the safety of human beings may be related to whether or not a sensor’s data is received fast enough to shut down equipment before an accident, performance is not just a competitive advantage, but a critical requirement.
Security, of course, becomes paramount. A single breach of an IIoT network could lead to disaster.
Yet it is almost inarguable that the cloud is a necessary ingredient to the successful growth of any IoT initiative, industrial or not. Public IaaS cloud is often hailed as the answer to the need for scale and the cost-effective compute behind it. Public IaaS cloud enables ease of access to the millions (or is it billions now?) of things that need to report in, answer questions, obtain updates, and receive instructions from their human owners.
But for IIoT, public IaaS cloud doesn’t always make sense. The systems that make the decisions (the ones that need to scale and harness lots of compute) need to be close to the sensors and things they’re interacting with. Those things are on-premises, spread across a manufacturing floor or neatly distributed in rows of shelves in a warehouse. To force those systems to send data way out there, to a public IaaS cloud, invites undesirable delays that degrade the performance necessary to ensure faster than disaster reactions to the conditions and situations the sensors are monitoring.
Organizations appear to be showing strong preferences for those cloud models they can control – namely colo and private cloud, both on and off-premises. Public IaaS doesn’t fare as well for either market, but is definitely less likely to be chosen by those implementing IIoT systems and architectures.
Some questions remain, of course, such as whether or not healthcare – which is a significant portion of the IoT market – is “industrial” or “consumer” grade IoT. After all, it relies on sensors, some of them increasingly ‘wearable’, to monitor patients and alert medical staff to potential problems, but just as many are used on-site and monitor critical infrastructure. Surprisingly, perhaps, healthcare was not one of the industries making the biggest investments in either IoT market. Health care is spending far more on mobile apps, private cloud, and big data with equal gusto.
In our latest survey, IoT investments ranked highest for manufacturing (surprised?), technology, and telecom firms – but not enough to pull either technology into the top three for those industries. The top tech for investments for all industries was almost unanimously private or public cloud with a few outliers turning to SaaS, big data, and one notable exception that is in love with virtual desktop infrastructure (natural resources). That actually aligned well with perceived strategic impact, where Industrial IoT managed to sneak ahead of Consumer IoT, but both remained well under the clouds.
IoT in general is just getting into gear, so seeing variation in where its associated apps are being deployed is not surprising. As the two markets continue to mature and best practices emerge, we should see the gravitational pull of different types of clouds exerting more control over where such apps are deployed. IIoT looks like squarely fall more on the side of controllable clouds (private and colo) while I would predict that consumer IoT will start to naturally migrate toward more commoditized (public) cloud models. Some may even start on-premises, in a private cloud model, and depending on success (growth) may migrate or at least expand to include public cloud resources as needed for scale and to manage cost per unit to maintain.
For now, IoT seems to be landing more on-premises than off when it comes to cloud. Which may say more about cloud than it does IoT.