With the arrival of the Internet of Things (IoT) and the predicted surge in the number of connected devices, edge computing has become increasingly necessary. The architectural standard for edge computing is called Multi-access Edge Computing, or MEC. Edge computing is the concept of bringing real-time, high-bandwidth, low-latency access to the user by putting computing power at the edge of the network.
IoT represents all devices that are connected to the internet and can communicate with other connected devices through wireless networks and embedded sensors. Research from Business Insider projects over 55 billion IoT devices will be in use by 2025 as compared to 9 billion in 2017. Research by Cisco predicts that by 2022 there will be 14.6 billion machine-to-machine IoT connections.
Edge computing will continue to play a major role in IoT functionality as the IoT market grows. The low latency and reliability edge computing provides are requisites for most IoT use cases. When automated vehicles communicate with each other about hazards on the road, they need the low latency provided by the nearby edge network to spread information fast enough to avoid crashes. Additionally, a smart factory can’t afford to stop production because the larger network has gone down. If it takes advantage of edge computing, the interconnected system of machines can keep running.
Edge computing alleviates the demand for access to the cloud from IoT devices.
How IoT Uses the Edge
An IBM blog post mentions that the edge not only helps applications handle poor network connectivity, but also decreases the volume of movable data. That means reduced traffic on the network and reduced cost of transmission.
The IoT network can benefit from data aggregation in the edge, according to a GSMA white paper. Data analytics can be done more efficiently because data is not replicated over multiple systems. Since the data at the edge is specific to the enterprise it serves, there is less data to sift through, which reduces latency and allows for faster decision making. Data pre-processing has the potential to make data sets more digestible for machine learning.
With edge computing, industries can maintain sensitive data from IoT communications at a localized source and only send non-sensitive information to the cloud for processing. VMware’s Project Liota, for example, allows for control of data transmission from devices to the cloud or data center. It is open enough to be used in many ways, including keeping confidential information local.
In a blog post, Microsoft describes the benefits for edge computing when internet connection is not reliable, or for use cases in remote areas, where edge computing allows for manufacturing equipment and smart devices to continue operation. Edge computing is attractive in remote locations like ships, planes, and rural areas for its quick analysis of data. Equipment failures can be detected without access to the cloud, since data from sensors can be analyzed locally.
Edge computing is multifaceted and can be used for many things. However, most conversations about edge revolve around IoT. Exponential growth is expected in the number of IoT devices that demand consistent connectivity and low latency — the very functions edge computing promises.
Updated April 2019