Internet Of Things, commonly called IoT, refers to the connection of daily devices like cars, home appliances, and industrial devices to the internet. As Gartner predicts IoT will reach 26 billion connected devices by 2020. One of the challenges in IoT is to capture, analyze and gain insights from data from this massive volume of devices effectively and efficiently. Most Internet Of Things Devices equipped with sensors have two parts- firstly, a front-end application or a device like Coffee Maker; Door lock (generally closer to the consumer) and secondly, a cloud where the data from the device goes and is then processed to build context (generally remote to the consumer). The first one is referred to as “IoT Edge” and the second one is called the “IoT Cloud.” The activities happening on the IoT Edge part is known as “Edge Computing.”

Microsoft, CISCO, IBM, Dell, and many startups are championing Edge computing; now that represents a shift in IoT implementation architecture. In Edge computing, data intelligence happens on the Edge instead of the Cloud which resulted in localizing certain kinds of data analysis and decision-making. Edge computing enables quicker response times, resolves network latency and reduced data traffic by sending only selective data to the cloud. Edge computing helps to achieve better efficiencies.

Two major components in Edge computing are Edge computing hardware and Edge computing software. The processing power of the Edge hardware plays a crucial role in determining the Edge software capability. Usually, a high computing device like a desktop or server or custom Edge hardware is a good choice. Edge software typically comprises of components that are required to establish the connection with the device sensors like Wifi, Bluetooth, ZigBee, Z-Wave and components needed to extract and store data from the device sensors like database.

The third components are necessary to perform data analysis like data analytics and machine learning. Edge Analytics and Edge Machine learning have been around for a few years now. There are many open source and proprietary software in the market. However, they had limited use for the last few years due to the complexity of including them in low compute power legacy IoT network solutions. AWS Edge and Azure Edge are a couple of prominent Edge software from Amazon and Microsoft. Apache Kafka and Scikit are open source implementations of Edge Analytics and Machine learning respectively.  

Of late, the Edge software package that has been making waves is Edge AI. Edge AI is Artificial Intelligence capability on the IoT Edge and is also a prime place for more innovation, as  adding these advanced software capabilities into a limited computing power resource is considered to be technically challenging. The goal of Edge AI is to understand, learn and act on the data from the IoT devices without any human intervention.

Edge Software is taking center stage in IoT applications due to its ability to offer lower network latency, faster reaction times on the IoT data and lower cost of operating on the data without much support of cloud computing. Because of this reason, we will start seeing them more in legacy and new IoT applications. We will see all IoT platforms, both of established players and startups embracing Edge software into their platform offerings. The hyper-scale cloud players like AWS, Microsoft and Google have been slow to enter the Edge software space; solely because Edge can have negative effect on their IoT cloud revenues. However, the technology and business case for the Edge software, as discussed above, has become prominent recently, that nobody can afford to ignore the role of Edge. Hence, this is forcing the IoT cloud players to explore newer revenue models that includes both IoT Cloud and IoT Edge Date intelligence revenue streams rather than riding only on the IoT cloud revenue stream.  

Edge Data Intelligence has become a key part of Enterprise IoT strategy already and there is lot of technology advancements happening. This will continue to have a positive impact on the IoT applications as the industry gets into a mass adoption phase.