Living on the Edge — Next wave of IoT transformations
With the hype around Edge Computing, Internet of Things (IoT), and Machine Learning at its peak, we are at the cusp of the next evolution of cloud computing. The first decade of cloud computing was driven by migrating on-premise workloads to the cloud. Now, with the proliferation of IoT devices and the sheer velocity and volume of data generated by these devices, it has also increased the need for streaming all this data to the cloud efficiently. Efficiency is needed in terms of ensuring the lowest possible latency, reduced costs for data storage and analytics and the ability to scale linearly as more and more IoT devices are added. What we need to accomplish these are cloud capabilities to be brought closer to where data is generated. Edge computing is increasingly becoming an answer to that need. The next decade will see a shift towards workloads being migrated from cloud data centers to edge locations. That doesn’t mean migrating workloads to the cloud is going away anytime soon. But the rise of Edge computing is creating new possibilities for IoT applications and also setting the stage for the next round of infrastructure evolution. It’s creating a new paradigm for where workloads reside.
What is Edge Computing?
The edge in edge computing can be thought of as the world of connected devices and gateways sitting on the field and acting as a local proxy to its corresponding cloud hub. These devices and gateways have the ability to do advanced on-device processing and analytics which is what the computing part is all about. This ability also makes edge computing a critical component of an IoT application. An IoT application delivers real value when the data generated by the things can be used to generate intelligent insights and drive actions. Bringing cloud-like capabilities closer to devices opens up new avenues for developing intelligent applications. It allows us to decide whether to put data someplace centrally for analytics or to bring analytics to where the data is. What I would consider as the true Internet of Things lies at the intersection of Artificial Intelligence, Cloud, and Edge Computing.
Primary Drivers
The need for Machine Learning close to the source of data generation is a powerful driver for the rise of edge computing capabilities for IoT applications. Devices not only need to run complex analyses quickly, but they also need to do so while consuming very little resources.
“By 2022, Gartner predicts 75% of enterprise-generated data will be created and processed outside a traditional centralized data center or cloud . Currently that number is around 10%”
Processing at the edge allows complex stream processing and machine learning to be run with minimal latency. For instance, a healthcare monitoring system tracking patient health data in real-time. On-edge-device machine learning can enhance patient care greatly allowing practitioners to react more quickly by detecting and being alerted about anomalies sooner. Edge devices can act as smart gateways by aggregating patient data locally and running complex analytics. The edge device can also execute basic commands or adjust medical device settings even if internet connectivity fails.
Security and privacy can also be improved with Edge computing by keeping sensitive data within the device. For example, healthcare companies can leverage Edge computing to help protect patient privacy by keeping the data at the source rather than sending Protected Health Information (PHI ) to the cloud.
Edge Computing delivers real value in a variety of IoT use cases. It can help respond and make decisions in near-real time by facilitating the lowest latency between the device data and decision. Also, when working with a massive amount of data produced by devices, having the ability to analyze and filter the data before sending it upstream can lead to huge savings. Not all IoT data generated by the devices are required for cloud analytics. Data can be processed locally and only the data meaningful for further analysis can be sent upstream to the cloud.
Key players
Edge computing involves moving workloads away from the cloud, so it’s not surprising that the three big cloud vendors are doing all they can to have those workloads in their eco-system:
Amazon — AWS Greengrass builds on the company’s existing IoT and Serverless offerings to extend AWS’ capabilities to edge devices.
Microsoft — Microsoft has Azure IoT Edge which allows containerized cloud workloads; such as Azure Cognitive Services, Machine Learning, Stream Analytics, and Functions to run locally on devices from a Raspberry Pi to an industrial gateway using Azure IoT Edge.
Google — Google announced two products for developing and deploying smart connected devices at scale: Edge TPU and Cloud IoT Edge. They are still in the alpha testing phase currently.
What’s next?
Edge computing is the next natural shift towards a more distributed architecture that has taken us from mainframes to PC’s, to the internet, to the cloud. There is a huge opportunity for businesses to learn how to benefit from the available distributed computing power — tapping into the capabilities of cloud intelligence and edge intelligence with IoT platforms connecting the two. Edge devices are becoming more powerful every day with advanced compute capabilities. That is setting the stage for the next wave of IoT transformations.
