Manage Learn to apply best practices and optimize your operations.

Manufacturers’ database quandary on the path to IIoT transition

For most manufacturers rapidly undergoing digital transformation, IIoT success means monitoring the quantity of materials, assessing the condition of production machines, and providing key production data to workers who can in turn act quickly to ensure product quality and production efficiency.

These benefits of IIoT enablement are becoming even more important as production increasingly shifts to dispersed locations. For many manufacturers, the operational monitoring, maintenance and troubleshooting no longer reside in a single physical location.

Ensure a successful IIoT implementation

Manufacturers’ IIoT deployments will collect millions of pieces of streaming data coming from hundreds or thousands of different data points each minute. This process of collecting data helps to detect irregularities and communicate issues to floor workers accordingly. It’s a massive data undertaking, and the time sensitivity inherent to a successful IIoT undertaking puts a huge responsibility on fast and efficient data management.

The energy sector has undergone a similar transformation within the past decade, as the rise of renewables has increasingly decentralized the energy generation. With this shift from a handful of large power plants to hundreds of thousands of individual producers, such as wind turbines or solar panels, sensor data has had a key role in optimizing power production.

For example, a change in the wind speed creates considerable fluctuations in the amount of energy generated, which connected systems can detect in real-time to trigger appropriate responses such as throttling other energy producers, activating temporary storage and changing sales prices as necessary.

But data management requires a decentralized recording of millions of data per minute across a wide variety of formats, enrichment of the resulting time-series data with additional information, real-time analysis, and instant data transfer for visualization and actionable response.

From both a capacity and performance perspective, conventional databases become no match for these data demands. Pure SQL-based databases are too sluggish, while NoSQL databases lack the required functionality and convenience to really be used all that effectively. At the same time, larger and traditional databases simply aren’t cost effective in dealing with complex time-series data.

In contrast, newer databases designed for time-series data combine what’s good about relational databases and what’s good about non-relational databases, delivering SQL functionality built on a more flexible NoSQL backbone. Furthermore, time-series databases offer distributed data storage, shared-nothing architecture and cloud service delivery that makes them ideal for manufacturers’ IIoT applications; even in the most demanding environments.

The distributed shared-nothing architecture means that every node of the database system is the same, making it possible to increase the performance of the overall system simply by adding new nodes while bolstering availability.

Use SQL and NoSQL in tandem

For manufacturers, the ability to tap into the best components of SQL and NoSQL enables them to continue to use their existing applications with SQL interfaces, and to pragmatically leverage their existing knowledge of those solutions. Most time-series databases offer extensive interfaces at both the data and application levels as well, allowing manufacturers to implement complex data management solutions in-line with the specific needs of their IIoT systems.

For example, machine and sensor data can be read into and enriched in the database through streaming gateways in popular formats such as MQTT, JSON and others. At the application level, data can be visualized, monitored, analyzed and delivered in real time, helping optimize the all-important manufacturing metric of overall equipment effectiveness.

The success of manufacturers’ IIoT implementations requires more than just the right array of sensors: it depends crucially on deploying a data management strategy able to wield collected data productively and optimally. Looking forward, with emerging technologies such as AI and machine learning requiring systems able to harness enormous quantities of data, modern time-series database architectures offer to pave both a solid foundation for today’s IIoT use cases and a clearer path for scale.

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.

Data Center
Data Management