Data and insights are valuable assets for any company with data driven decision making permeating every aspect of our lives.
Historically, businesses collected data from different IoT devices and sensors, gathered them in a central point such as a data lake or data warehouse and performed analysis to gain insights. What if organizations can remove the step of data centralization/integration and skip to the analysis phase? This approach is called “edge analytics” and enables organizations to achieve
- autonomous machine behavior
- higher levels of data security
- reduced data transfer costs
What is edge analytics?
Edge analytics is a method of analytics that performs analysis at non-central components of the system, such as sensors, switches and various connected devices. In other words, insights are gained closer to the devices where the data is gathered, rather than relying on a central location that can be thousands of miles away.
With the emergence of IoT technologies, there is an increasing amount of data that organizations collect and want to analyze. but transmitting data to a central point and insights to the edge back and forth takes time. Edge analytics provides fast and decentralized insights from data sets collected from the edge of the network.
Why is it important now?
It enables faster decision making, especially in cases of low bandwidth. Given businesses’ increasing reliance on automated, data driven decision making, edge analytics is an area of significant investment by tech giants.
Industries such as retail, energy, security, manufacturing and logistics can benefit from quick decision making enabled by edge analytics. For example, an autonomous vehicle needs to make a split-second decision about braking when it encounters an obstacle on the road. The decision speed requirement, in that case, is far faster than any cloud computing solution can provide. Businesses are rapidly deploying sensors and smart devices at the edge of the network that helps analyze data faster.
Edge analytics is not only for making decisions within milliseconds. Data collected from various devices and sensors is increasing rapidly. In cases where there is limited bandwidth between the server and the edge device, the speed of data transfer may not be sufficient even for less time sensitive applications.
Edge analytics is an area of significant investment by tech giants:
- In Jan/2020, Apple acquired an edge focused AI start-up called Xnor.ai. Apple plans to run deep learning analytics models on edge devices such as phones, IoT devices, cameras, drones, and embedded CPUs.
- Both Google Cloud and AWS have edge IoT focused products
How does it work?
The general workflow of edge analytics tools follows this pattern:
- Sensors or devices at the edge collect data
- Analytics capabilities within the devices enable performing analysis at the edge
- If the device needs to take action, it does so relying on the results of the analysis. For example, Rulex is a vendor that provides autonomous operational decisions with real-time analytics at the edge.
- Relevant data is transmitted from the edge to the cloud so businesses can see the big picture by aggregating summarized data (in case of bandwidth constraints) from thousands of devices
The figure below shows the working principle of IBM IoT Edge Analytics in a global hotel chain. A microphone is equipped with Watson analytics tool and it analyzes the tone of customers’ voice whenever a customer interacts with the receptionist. After the analysis, hotel management gains insights and designs actions to increase customer satisfaction.
For a more technical guide on how edge analytics works, IBM has a series of videos showing how businesses can leverage from IBM Edge Analytics solution:
How is edge analytics different than regular analytics?
Edge analytics have similar capabilities as regular analytics applications, except for the place where the analysis is performed. One major difference is that edge analytics applications need to work on edge devices that can have memory, processing power or communication constraints. These applications are optimized to work within these constraints.
When should businesses prefer performing analytics at the edge?
For edge analytics to be advantageous for businesses, organizations should know the answers to the following questions:
- Does analyzing data in real-time improve productivity for your business?
- Do you need a solution that can scale over time?
- Does your business in an industry that requires fast responses when an unexpected change happens?
If your answer is yes to the above questions, edge analytics tools are what your business needs for data analysis.
What are the advantages of edge analytics?
Advantages of edge analytics include:
- Faster, autonomous decision making since insights are identified at the data source, preventing latency
- Lower cost of central data storage and management since less data is stored centrally
- Lower cost of data transmission since less data is communicated to the central data warehouse
- Better security/privacy since the most granular data such as video footage is not stored or communicated
What are edge analytics use cases?
- Retail customer behavior analysis: Retailers can leverage data from a range of sensors, including parking lot sensors, shopping cart tags, and store cameras. By applying analytics to the data collected from these devices, retailers can offer personalized solutions for everyone with the help of behavioral targeting.
- Remote monitoring and maintenance for various industries: Industries such as energy and manufacturing may require instant response when any machine fails to work or needs maintenance. Without the need for centralized data analytics, organizations can identify signs of failure faster and take action before any bottleneck can arise within the system.
- Smart Surveillance: Businesses can use benefit from real-time intruder detection edge services for their security. By using raw images from security cameras, edge analytics can detect and track any suspicious activity.
What are the pitfalls to avoid?
We have observed two common pitfalls to avoid when organizations decide to put money on edge analytics.
- Security: Cloud environments are designed with security in mind because breaches on the cloud are quite costly for the business. However, edge security is also important because some edge devices make decisions about real-world behavior of machines. Breaches can result in the sabotage of equipment, other costly machine errors or at least misinformation.
- Maintenance: Some edge analytics systems share only their output with the cloud due to bandwidth or storage constraints. Then, businesses have no chance to review the raw inputs that led to the analyses that are shared with the cloud systems. Therefore, they need to make sure that inputs are processed with the latest analytics software, relying on outdated models can lead businesses to make decisions on wrong information.
What are the edge analytics tools?
Some edge analytics tools include:
- AWS IoT GreenGrass
- Cisco SmartAdvisor
- Dell Statistica
- HPE Edgeline
- IBM Watson IoT Edge Analytics
- Intel IoT Developer Kit
- Microsoft Azure IoT Edge
- Oracle Edge Analytics (OEA)
- PTC ThingWorx Analytics
- Streaming Lite by SAP HANA
Interested in learning more about analytics that is changing the way we do business? Check out our research for a wide range of analytics related articles:
If you still have questions, please don’t hesitate to contact us:
How can we do better?
Your feedback is valuable. We will do our best to improve our work based on it.