Today, we live in a world where the need to capture, control, and compute real-time data is exploding, a world in which everything communicates, everything contributes and drives a demand for answers now. Consequently, analysts predict that the Internet will proliferate to more than 20 billion connected devices by 2020, which has now raised serious questions regarding its capacity for supporting this dramatic increase in data traffic. Managing this explosion of data won’t be solved through old thinking, legacy products, and traditional cloud computing. To create a future where information remains dynamic, free flowing, and relevant, we need to embrace a new direction.
Currently, the conventional wisdom is an Internet of Things (IoT) comprising dumb sensors embedded with the physical world communicating raw data through the Internet to a centralized processing facility located in the cloud. Here, in the cloud, it is imagined that huge volumes of data will be collected, stored and processed into information, which will be visualized in a user interface. Technically speaking, such an approach can be described as centralized computing. In the context of IoT, I refer to this as the Connected Thing Paradigm.
While this centralized computing approach for IoT has demonstrated value in some initial markets, inherent technical limitations exist with its long-term viability to meet key requirements related to performance, reliability and scale to achieve the grand vision of IoT. Specifically, IoT’s predicted growth in connected sensors will generate significantly more data traffic than centralized computing architectures can support for timely analysis and reporting, especially across already overburdened cellular and Internet infrastructures. Moreover, the associated expense of ever-increasing network bandwidth from connected things will negatively impact IoT’s long-term cost of ownership and thus its value proposition in many key vertical markets.
Addressing these issues will require reimagining IoT’s relationship with the cloud, a relationship in which the cloud is used for higher-level information processing (i.e. Big Data), visualization and sharing with real-time analytics occurring at the source of the data. This distributed processing architecture can be described as edge computing where analytics is migrated from a centralized point of data aggregation in the cloud to the things connected at the extremes of the network. The result is an Intelligent Thing Paradigm where real-time analytics process raw sensor data locally, thus driving significant reductions in communications traffic to the cloud and the network’s associated connection time and bandwidth requirements. As such, a more independent, intelligent network of things and thus efficient system architecture for IoT emerges with key performance benefits including lower system latency, reduced communications costs, improved security and increased quality of service.
However, implementing edge computing in IoT will not be possible with legacy technologies and will require a new category of device. In response to this need, Camgian Microsystems announced today its intent to bring the benefits of edge computing to IoT through a novel, integrated software-hardware product to be released this Fall. The product, called Egburt, will bring revolutionary edge computing capabilities to IoT supporting software defined sensor interfaces, powerful real-time sensor analytics, ultra-low power consumption and IP-based cellular communications.
In conclusion, we believe that intelligent not connected things will drive the growth of IoT. Achieving the vision of the intelligent IoT will require a new category of device and embracing a new direction for the Internet’s next generation.