The Intelligent IoT Edge: The Next Step in Enterprise Infrastructure
The Internet of Things (IoT) is rapidly pushing enterprise infrastructure outward to a new, broadly distributed network edge. To prevent latency in a world of autonomous cars, smart cities and always-on connectivity, data and data processing must be kept close to the IoT device. Much of the infrastructure needed for IoT will be unmanned and yet tasked with making crucial decisions. How can IoT service providers ensure data systems housed on the edge will make the correct decisions continuously for thousands, if not millions of data flows in a wide range of situations?
The answer lies in a combination of artificial intelligence (AI), machine learning, cognitive computing and other smart technologies already fueling the transition to a digital services model. By deploying systems that can learn from past data patterns and adapt to new situations, industries ranging from automotive and electronics to insurance and retail will be able to maintain complex IoT workflows for asset and facilities management, product development and a host of other functions, with little or no manual oversight.
Essential Enterprise Infrastructure
RCR Wireless notes that machine learning is enabling new ways to optimize IoT. To manage immense loads of data and quickly receive, analyze and distribute it, edge-based processing needs to know what to keep, what to ignore and what to forward to central authorities. In a traditional setting, teams of programmers tell computer systems what to do — but in IoT, the computers themselves take on that role.
An intelligent edge will also reduce the costs of deploying IoT infrastructure. According to Enterprise Innovation, companies in the Asia-Pacific region that relied on centralized architectures for their IoT models experienced enormous expenses with little or no revenue. By pushing loads to the edge, they dramatically cut costs while improving latency, enhancing performance and scaling up operations to generate revenue.
The IoT edge must become more intelligent to meet the demands of smarter IoT devices. To that end, IBM recently teamed up with Singaporean messaging firm Unified Inbox to devise an intelligent IoT messaging platform to equip human-to-machine conversation, according to Techseen. The system uses the Watson platform to support natural language and conversational intelligence on Unified Inbox’s UnificationEngine platform. Users gain voice-driven interfaces to smart devices at home and in the office, connecting to more than 20 international messaging platforms ranging from legacy email and SMS services to the latest apps and chatbots like WhatsApp and Skype. Smart cities can also deploy this system to optimize transportation, events and emergency warning applications.
Perhaps the most crucial function AI offers the edge is enhanced security. With such large volumes of free-flowing, largely unstructured data passing between user devices and enterprise infrastructure, IoT dramatically broadens the attack-vector geometry to sensitive data and systems. According to Digital Trends, researchers at MIT calculate that machine learning can predict 85 percent of cyberattacks and reduce false positives. This is achieved by improving threat detection, identification analysis and the ability to automate many of the error-prone functions of data management that increase the risk of compromise.
An intelligent edge is the only viable way for IoT to fulfill the promises users are expecting from their new smart devices. The data loads are too vast, the speed and dynamism of the workflows too great and the complexity of network connections and analytical relationships too intricate for manual processes.
Going forward, the IoT edge will not simply be another layer of infrastructure but a thinking, reasoning member of the business team.