Of all the challenges facing the development of IoT infrastructure, one of the most difficult is the need to establish real-time data gathering an analysis.
Not only does this require high-speed, high-power systems and processes throughout the data chain, it also requires a level of data handling that was unthinkable just a few short years ago. The fact is, very little of the IoT data stream is actually necessary for real-time operations, so IoT-facing infrastructure will have to devote a good deal of its resource capability simply identifying and extracting these crucial bits from the reams of run-of-the-mill data.
Think of all the data that will be generated from a smart city, says Crystal Valentine, VP of technology strategy at MapR. Much of the data involving traffic management is of the macro sort – used for long-term planning and basic functionality like red-light coordination and vehicle rerouting. But some of it requires split-second action, as in variant behavior that places lives at risk or potential mechanical failure that could impede vehicle flow. Establishing the processes to properly handle both offline and online data will come to define the early stages of IoT development.
It’s hard to see how this will happen without artificial intelligence, says Customer Think’s Thomas Wieberneit. The data volumes are too large and the speed requirements are too fast for even an army of technicians. With each data set capable of combining with other sets in an unlimited variety of ways, only a thinking, learning algorithm has the capacity to parse the information in ways that are useful and within a reasonable timeframe. This isn’t about using AI to do more with less, both in terms of systems and people, but in applying the only technology that can bring true value to the IoT investment.
Technology alone won’t produce real-time IoT performance, however. As Built.io’s Nishant Patel notes, establishing proper data integration patterns is also crucial for smooth communications to and from connected systems. The most common approach consists of data stream processing and rule-matching, usually fostered on an integration Platform as a Service (iPaaS) or Mobile Backend as a Service (MBaaS) architecture. This provides a high degree of modularity for various systems and use cases, and can extend across multiple mobile, web and IoT topologies.
For real-time performance, of course, you’ll need high-speed processing at multiple points in the data chain. A company called Kinetica hopes to tackle this challenge with an IoT database platform built around graphical processing units (GPUs). The typical GPU, after all, has upwards of 4,000 cores compared to 32 for a general-purpose CPU, so by going with an all-GPU configuration the company says it can crunch numbers 100 times faster using about one-tenth the hardware. Born out of the U.S. Army’s Intelligence and Security Command, the Kinetica platform employs deep learning, geospatial visualization, location-based analytics, and a host of other tools for applications ranging from connected cars to inventory management.
Even if the application does not depend on real-time data and analytics, the enterprise still has a vested interest in generating results as quickly as possible. In today’s fast-moving economy, opportunities come and go at the speed of a mouse-click or as quickly as customers can get out the door. Putting data into action sooner rather than later will provide the competitive advantage that organizations need to remain relevant, and profitable, in a digital universe.
With a real-time IoT infrastructure in place, the enterprise has a handle on not just what is happening now, but is likely to happen in the near future.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata and Carpathia. Follow Art on Twitter @acole602.
