What we can expect to see from edge computing in IoT in the coming years
Data is an enterprise’s most valuable asset, and IoT devices constantly gather vast amounts of data from their surroundings. Only recently have organizations given real thought on how to capitalize IoT device data, especially edge IoT device data with data sourced elsewhere. For example, agriculturists collect weather data from IoT sensors and use data points over time with historical weather data to determine when to plant and harvest crops. Agriculturists can also use IoT data to determine when and how much to irrigate and fertilize crops for the best yield year over year. Edge computing’s ability to network IoT devices together and analyze data creates a more complete picture of data across devices, supplemental data and analysis from upstream sources.
Edge computing is steadily gaining popularity within IoT because of improvements in how data can be captured, analyzed and used. Today, around 10% of organization-generated data is created and processed outside traditional data centers or the cloudGartner predicts this figure could reach 75% by 2025. With 90% of data currently at the core, it makes sense that devices send edge data back to the core by default. With edge data this will all change.
As edge computing becomes a more integral part of modern IoT data management, here’s what we can expect to see in the coming years.
Edge computing adoption will expand as localized data becomes the majority
With so much data at the edge, latency concerns will begin a shift away from the costly, insecure and slow process of sending data to the cloud and back for analysis and storage. Latency is one critical factor that will drive the need for richer applications and more advanced analytics that require local data management at the edge. Another critical factor will be avoiding the storage of data that could become a security vulnerability and ensuring prompt decisions made at the point of action. Due to these factors, we will see significant increases in adoption of resources that can handle compute, analysis, supporting data management and networking over the next few years.
Edge computing creates a better understanding of performance, durability and use patterns in relation to end users, consumers and employees. Better comprehension results in increased and improved usage, user retention, streamlined operations and improved ROI over time. Increased adoption will save businesses time, money and compute resources and encourage better data use.
Data harvest and use rates will grow
Organizations will extract more value from IoT data because of the ability to deploy stronger analytics and underlying databases for management. Enhanced management includes lower power compute resources, smaller form factors, higher durability and the plummeting cost of underlying hardware.
As a result, organizations will generate more data more quickly and must collect data faster, more efficiently and closer to the source. They can use edge computing to pull insights from the data and only send what is needed to the cloud, which will cut costs and inefficiencies. For example, several lidar signals must be collected and analyzed to support self-driving cars. This data is important for real-time decision-making for the car. Data scientists and engineers can also use metadata describing the lidar array’s data and vehicle’s decision-making patterns across millions of cars to develop a more accurate decision-making algorithm set. They can download the algorithm as updates to continually improve safety and reduce accidents per million miles of driving. Data scientists and engineers can collect metadata from proper tools and procedures, including cloud data warehouses and artificial intelligence or machine learning platforms, such as Tensorflow. Processed historical data can provide the fuel to further tune machine learning and other artificial intelligence algorithms so that more IoT data can deliver more actionable insights, leading to better monetization of IoT data.
The onset of 5G networks will increase efficiency
5G networks promise to provide huge bandwidth and peer-to-peer interactions between devices to create richer shared information and analysis performed at the edge without latency of using back-end systems to act as arbitration and central analysis for edge operations.
With higher network bandwidth, many industries — including communications, media and entertainment, logistics and transportation, healthcare, manufacturing, education and smart cities — will see industry shifts take place. Secure localized groups will join in 5G-enabled augmented reality and virtual reality games, meetings and other localized peer-to-peer scenarios. In other cases, such as smart homes or smart hospital rooms where local integration and decision-making capabilities are essential, multiple connected devices will share information and perform analysis more quickly and seamlessly with 5G bandwidth.
Organizations will continue to collect, harvest and use IoT data at an increased rate, proving edge computing is a key tool for streamlining the process. There will be increased interaction between local devices and gateways as consumers and organizations integrate 5G, and these interactions will create new use patterns. Interactions will have the potential to improve, and even radically change, decision-making by devices and people at the edge.
As volume and velocity of IoT data generation and collection increases, so will inefficiency in streaming information to the cloud or data centers for processing. Innovations in edge computing will streamline this process sooner than we think. Local device and gateway compute resource capabilities coupled with networked artificial intelligence and significant increases in bandwidth will open new possibilities for how edge devices and gateways will interact with each other and the people using them. In the coming years, using IoT data at the edge will further prove to be an important innovation with enormous promise to unlock new capabilities across industries.
All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.