Optimize Industrial IoT Data with InfluxDB and AWS

By Jessica Wachtel / Use Cases
Jun 12, 2023

The modern factory’s relationship with data is experiencing a major change. Data now shapes the future rather than only telling the story of the past. The language inside the factory sounds like higher Overall Equipment Effectiveness (OEE) as the result of a shift from preventive to predictive maintenance. It could also look like expanding business goals to a new market based on impactful data-driven decisions. A change in purpose requires an update in technology. This is no frivolous update – legacy data historians lack the tools and interoperability required to derive the meaningful insight needed to revolutionize an organization’s relationship with their data.

There are several factors that go into successful predictive maintenance strategies and impactful data-driven decisions. Real-time data processing at nanosecond precision is a must. That enables real-time advanced statistical analysis, machine learning models, and artificial intelligence (AI) derived insight. The real-time insights work jointly with historical records to paint a complete picture. Amazon Web Services (AWS) and InfluxDB have a suite of tools that fit the criteria needed to modernize the Industrial IoT (IIoT) environment.

Trailblazers

Bboxx and Moxie IoT created modern solutions for longtime problems using InfluxDB and AWS.

A Bboxx system includes a solar panel connected to a battery with a set of USB and DC connectors. These clean, affordable, solar energy systems power light, radios, and low powered TVs in the developing world. Bboxx turned to InfluxDB to handle the storage, visualization, and analytics for its millions of sensor readings in real-time. With InfluxDB, Bboxx analyzes data across their entire customer base without limitations. They host InfluxDB on AWS with the remainder of their technology stack and this helps keep latencies down.

MOXIE IoT created Moxie World, a comprehensive solution that tracks the movement of factory activity and assets. Moxie World collects, stores, and analyzes industrial sensor data and displays real-time dashboards on their iOS app. InfluxDB features, such as time-stamped data collection, querying specific ranges of time, and offloading data aggregation, are some of the reasons why MOXIE IoT built InfluxDB into the Moxie World platform. Moxie uses an MQTT broker with AWS EC2 instances running in Python to handle data collection. This workflow helps first with sorting by customer followed by filtering based on preferences.

Tools and services

Telegraf

Telegraf is lightweight, written in Go, has no external dependencies and requires a minimal memory footprint. It’s perfect for collecting industrial IoT (IIoT) sensor data because you can install Telegraf on even the smallest device and it can handle the scale and volume of industrial data. Telegraf’s architecture is plugin-based with over 300 plugins available for just about every service and protocol.

InfluxDB

InfluxDB is a purpose-built time series platform, designed for large datasets for both high read and write throughput, high availability, low read latency, and durable edge-to-cloud data replication. InfluxDB is built on top of Apache Arrow in the Rust language. It’s a columnar database with greater flexibility because it separates compute and storage. InfluxDB supports SQL queries and uses Flight SQL to transfer data between clients, servers, and other systems and tools. As more solutions in the open data ecosystem adopt Arrow, users have more options to easily integrate with InfluxDB.

Connectivity and control services

AWS has a full suite of connectivity and control services that allow developers and stakeholders to control, manage, and secure connected devices at scale. AWS Greengrass is software available for local install on the device and provides local compute, messaging, data caching, and sync capabilities for connected devices.

End-to-end architecture

It all starts with data collection on the factory floor. In addition to challenges presented by the massive volumes and velocities associated with time series data, there’s also a lack of standardization between industrial protocols. Telegraf is a flexible technology built to solve just that. With hundreds of input plugins including MQTT, Modbus, OPC-UA, and SNMP, and built-in compatibility with custom plugins, Telegraf is a single solution that can handle multiple data sources.

AWS can help streamline data collection if you’re working with many different protocols. If AWS Greengrass is installed locally on the device, it can output data as MQTT. Telegraf can then ingest all factory data from an MQTT broker. If the data is already in AWS cloud, Telegraf can run on an EC2 machine, an ECS container, or EKS and collect data with the Amazon Kinesis plugin. Telegraf has optional processor and aggregate plugins that allow you to amend and transform your data. Telegraf can output directly to the cloud or to an InfluxDB edge node on a hardware gateway inside the plant.

InfluxDB OSS has has a feature called Edge Data Replication (EDR) that empowers operators on the factory floor complete autonomy to collect, visualize, manage, and store data how and where they need it. Let’s say a factory has four locations worldwide and headquarters in Chicago. Headquarters doesn’t need all the nitty gritty details from each plant. The data scientists in Chicago need aggregated, transformed data from all the factories for their advanced statistical analysis, machine learning training, and AI-derived insights.

EDR automates transfers in a durable, safe way, by using single API across the entire platform. EDR operates on the bucket level meaning operators choose which buckets to replicate. EDR writes to a queue. If there’s a reason why the data can’t get to the cloud, such as a network partition, the queue will build up rather than sending an error. Once the connection restores, the data will flush and the replicated bucket in the edge and cloud will reach parity.

Once the data is in the cloud, data scientists can review what is and isn’t working in the big picture across all factories. The data at the edge helps operators work in the here and now. The cloud tools provide the access to open data, predictive modeling, higher order analytics, and machine learning, and AI-derived insights needed to make the plans for tomorrow. Different data is meaningful across different parts of the organization, and now the tools are available to make sure the right data is sent where it needs to be.

Conclusion

Modern factories are changing to meet the demands of today. The technology is available to ease the transitions and make data-driven decisions available with ease. There are many tools and services available from InfluxData and AWS that serve IoT use cases, whether those are commercial, industrial, or anywhere in between. The tools enable users to connect to virtually any data source, process, analyze that data, and create visualization, all to derive better decision making and improved outcomes.

To get started, contact someone on our team or open your free InfluxDB Cloud account now.