Edge computing helps overcome the curse of network latency. The origin of edge computing can be traced back to the nineties when Akamai introduced Content Delivery Networks (CDN). A CDN usually uses edge nodes closer to users to prefetch and cache web content. These edge nodes can also perform some content customization, such as adding location-relevant advertising, and more importantly, CDNs save bandwidth through caching.
In edge computing, the data gathering and processing occur at the application source or device, or at least nearby. This proximity could range between the hand-held or local devices and appliances, to point-of-sale or physical units distributed across locations. In fact, the concept of CDN is generalized by edge computing by leveraging the cloud computing infrastructure.
Edge Computing: Why Now?
The 5th generation of cellular network technology or 5G is expected to have 10x higher bandwidth than the current 4G technology. However, even with 5G, sending data across vast distances will involve network latency unless the 5G connected end-devices have higher computing power. This is exactly why edge computing is the key to realising the profoundly transformative potential of 5G, as both computing power and network bandwidth are mandatory to reach the 10x speed.
Moreover, Gartner predicts that by 2020, about 5.8 billion IoT sensors will be used in the enterprise sectors, including automotive. However, with the growing number of IoT sensors generating terabytes of data, there is a felt need for fast, reliable, secure, and cost-effective data processing within or nearby these sensors or edge devices. This need has created a common IoT approach of edge computing software communicating with IoT cloud platforms. But, again, the latest 5G-enabled IoT use cases mandating lesser than 20 milliseconds of turnaround time along with high throughput data volumes demand more of both – computing power & network bandwidth.
This combined push from cloud services, pull from IoT, and a shift from being data producers to being data consumers has catapulted the edge computing market to be worth $34 billion by 2023, as predicted by IDC.
Leveraging the Edge
Edge computing enables connectivity with the sensors and actuators across a multitude of industry protocols that facilitate different styles of device communication and supports the most common protocols. Resultantly, various use cases of edge computing that are rapidly emerging include video surveillance, image recognition, connected vehicles, autonomous driving, industrial automation, voice recognition, AR/VR media engagement, blockchain and fog computing.
Edge computing is also enabling drone management for unmanned maintenance and virtual fraud detection for retail, banking, entertainment, and more. Apart from all this, edge computing’s scalability, versatility, and reliability also make it an attractive proposition for companies implementing digital transformation. With faster data processing speeds for improving efficiencies, productivity, speed to market, and customer experience, edge computing also expedites digital transformation based on where a company strategically chooses to employ edge devices.
Edge computing has become a very important factor driving the adoption of technologies such as IoT, which is further enhanced by offerings like AWS Greeengrass, Azure IoT Edge, from the big players. However, the most innovative applications of edge computing, along with other new-age technologies, are emerging to make an average human’s life better, safer, and more comfortable.
This empowerment is not just here, but here to further evolve in the form of multiple devices and applications. Some examples include secure smart homes, resilient power grids, safer remote surgeries, automated vehicle insurance, and futuristic wearables with functionalities that are only limited by the technologist’s imagination.
Advantages of Edge Computing
Eliminating Latency – Applications expected to give near realtime performance, cannot wait for a data processing request to return from the cloud before taking action. Hence, sensor data from safety-critical industrial operating and control systems are processed using edge gateways to eliminate this network latency and achieve a quicker response time.
Portability – The software for edge solutions are usually portable to different hardware platforms. This prevents them from being tied to a particular vendor’s hardware or software.
Cost-Effective – In edge solutions, the relevant data that is sent to the cloud can be filtered out based on some thresholding applications. This filtering reduces the cost of data transmission over the network, cost of cloud storage, and the processing cost of irrelevant data.
Operational Autonomy – Edge computing provides the ability to have local storage and local computation, thereby making solutions autonomous in their operations. This allows the device to continue functioning even if it is not connected to the network. Advanced analytics algorithms or machine learning models, which can learn from limited data, can be efficiently deployed on edge gateways to facilitate efficient computing on smaller data sets.
Security – Starting from permission-based access control and encrypted communication to certificate management and easy integration with existing solutions – typical edge computing improves the security and privacy of an IoT application.
Moreover, by managing the amount of sensitive, private information transmitted through a network and the number of sensors and actuators connected to the Internet, potential security attacks are minimised.
By Sindhu Ramachandran, Principal Architect, QuEST Global