Machine learning and other forms of artificial intelligence will likely infiltrate all levels of the IT infrastructure stack, but some architectures will take to it more readily than others. And while it is tempting to view AI in terms of the changes it will bring to the data center, the more imminent and profound impact will be on the Internet of Things (IoT).
Particularly on the edge, AI offers the only viable means of assessing and coordinating the data flows from massive numbers of devices – many of them imbued with their own levels of intelligence – to produce results that are both meaningful and timely. Much of the storage and processing of IoT workloads will take place on the edge, so it only makes sense that it will incorporate intelligence as a core asset.https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=i
According to AutomatedBuildings.com founder Ken Sinclair, a new generation of edge controllers is poised to deliver on the promises of autonomy for everything from cars to washing machines and, yes, even buildings. A key application is pattern recognition and response, which is what allows an autonomous car to identify a dog or a person in its path and react accordingly. In the same way, an autonomous building can determine that there is no one at home at the moment and it is safe to adjust lighting and temperature to lower levels. Eventually, these learned actions become autonomous largely through the exchange of data between the intelligent device and the intelligent edge.
As a matter of preference, it seems that most organizations would prefer to implement intelligence on the IoT edge faster than on the centralized Big Data analysis engines that will do much of the heavy number-crunching, says Datanami’s George Leopold. A recent survey by GlobalData found that applying intelligence to things like enhanced insight and decision-making ranked dead last in IT’s list of priorities. Instead, the focus is on deploying machine learning on the edge to automate operations and reduce costs. Indeed, with increased data democratization, even long-standing centralized business intelligence platforms are starting to cede ground to smaller, more targeted approaches to data analysis, such as SQL query, predictive data modeling and auto-generated discovery visualization.
Intelligent IoT edge solutions are likely to emerge on industrial infrastructure more quickly than in consumer or enterprise settings. As Pipeline’s Susana Schwartz notes, Industrial IoT is expected to affect nearly two-thirds of global GDP, spanning industries as diverse as agriculture, health care, finance and transportation. Since the edge will form the crossroads between wired and wireless networks, IT and manufacturing, and a host of other platforms, a high degree of openness and third-party coordination is required. And the best way to do that is to drive as much intelligence into edge infrastructure as possible – something that cloud service providers will be eager to do in order to differentiate themselves from competitors.
Already, we are seeing the hyperscale cloud giants push intelligence onto their edge platforms. Amazon recently tapped NXP Semiconductors to integrate the Amazon Greengrass architecture onto the Layerscape Intelligent Gateway system. The goal is to securely extend the AWS Cloud to local devices where it can collect and analyze data closer to its source. With NXP’s technology, Greengrass can support functions like real-time data gathering and simultaneous management, analysis and storage.
As a greenfield deployment, the edge also has an advantage over legacy infrastructure when it comes to supporting artificial intelligence. With no retrofitting involved, the edge can be built around intelligent operations from the processor up. And since most edge facilities will be unmanned, basic IoT functionality will depend on the ability to know what to do without being told.
Eventually, of course, the rest of IT infrastructure will catch up, but for the time being the enterprise can continue to rely on human operators for centralized facilities where the speed of operations is still within manageable limits.
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.