In recent years, there has been a strong desire for manufacturing innovation around the world, and Smart Factories are at the center.

Smart factory is the combination of “information and communication technology (ICT)” and “production and manufacturing technology”, which is a traditional manufacturing technology.

Through combining technologies including Internet, Big Data, Cloud computing, Intelligent Robotics and Cyber-Physical system (CPS), the Smart Factory becomes a production system in which equipment and its parts are interconnected to one another.

Smart factories are different from simple factory automation. Factory automation is automated and optimized only for each unit process and it is not linked between processes. However, smart factories connect all data across all processes and facilities in the plant. In addition, it is possible to integrate sensor data with the production and management data of ERP and MES, so that management plans can be directly applied to the process equipment in the factory.

In smart factories, energy and materials are optimized for plant operations. This saves resources and improves facility operation rate by preliminary inspection through precautionary maintenance for equipment malfunctions. It controls plants that are automatically linked to production volumes and management plans. However, processing the large amount of data generated from sensors in factories can be challenging.

We will review the recent trends and related technologies from the data storage point of view for storing tag data generated in smart factories.

Smart Factory

Implementation Technology

1) Connectivity
Various sensors such as temperature, pressure and vibration are attached to manufacturing facilities while IoT technology is applied to data collection in real-time. Existing manufacturing facilities already include various sensors, actuators and PLC devices. Also, data communication protocols such as Wife, Zigbee, and BLE are available. In recent years, communication methods are being standardized as OPC UA.

2) OT (Operation Technology)
OT refers to collecting and monitoring the data generated by the equipment at the facility, primarily in the factory. OT utilizes historian or Real Time Database (RTDB) to collect and store data as well as monitor in real time. It can alert personnel in case of data that is outside normal ranges.

3) IT (Information Technology)
The IT department stores various data from the business systems such as MES, PLM and ERP and performs big data analysis. Rather than from a unit facility, it collects and analyzes data from the entire plant or even the entire company. Hence, the amount of data collected is enormous and massive processing techniques such as Oracle Exadata or Hadoop are needed. In addition, IT conducts correlation analysis and multi-dimensional analysis using expert analysis tools such as R and SAS. They also perform the advanced analysis through machine learning.


Latest Trends

Recently, only limited amounts of data is collected from unit facilities and monitored for changes in trends while the number of sensor devices has been increasing. Eventually, both the amount and speed of the data to be processed will increase tremendously.

1) Increase in the number of sensors
As the number of facilities and sensors increases the amount of collected data increases as well.

2) Enhanced scan rate
When the equipment data is fetched, the data is read at about 500msec, that is 2 times per second. However, there has been an increase in demand for analyzing data by reading the data more precisely. Thus, in recent years, 10 msec and even 1 msec resolution may be required. 1 msec means that the amount of data is exponentially increased by reading and collecting 1,000 tags of data per second.

3) Storage period and capacity increase
Until now, due to the storage capacity limit, the storage period of the facility data was less than one month. Most data is deleted without backup so it is difficult to keep the data for historical analysis. Also, even if the factory is applying machine learning, it is not easy to acquire proper learning modeling because the amount of data to be learned is absolutely small.

However, in recent years, with the decline of hardware storage costs and the technologies of big data, data storage period and capacity could be infinitely increased.

4) Combination of OT and IT
Traditionally, data for real-time monitoring that occurred at the manufacturing site was collected and used only by the business department. It was not transmitted to the IT department, so it was difficult to comprehensively analyze the entire company plan when issues occurred.

In recent years, IT departments tend to receive facility data through OT departments, or directly collect and analyze facility data on a separate channel from the OT department.

With the convergence of the OT data and IT, we are facing a trend which attempts to discover the business values that have been traditionally overlooked.


New Solutions required

1) High-speed data collection and storage
As the speed and amount of data to be collected increases, data storage which is capable of collecting and storing data at very high speeds of more than one million tags per second is required.

2) Scalability
Most systems pre-calculate CPU, MEMORY, and storage before they are used. However, there are some cases where more hardware resources are needed due to the change of business environments. In particular, as smart factories increase the number of sensors and the amount of data, system expansions are required. The multi-node cluster configuration will fulfill this need.

As smart factories increase the number of sensors and the amount of data, system expansions are required. Click To Tweet

3) High Availability Service
Most of the production facilities in operation at the manufacturing sites are always running 24/7 and tag data generated here must be collected and stored accordingly. It is necessary to cope with server failures. Also, a HA configuration is needed.


Time Series DBMS, Machbase for Smart Factories

Machbase is a high-speed time series DBMS that stores 2 million tag data per second while compressing and storing in real time. In addition, it is possible to create indexes in real time and query the data input simultaneously.

Machbase enterprise edition also contains multi-node clusters in which data storage and queries are distributed.

In addition, HA has replicas in Active-Standby configuration, so it is possible to fail over the system failures as well as HA. Also, since Machbase supports traditional RDBMS interfaces, DB engineers can easily start to use it without learning new techniques or DB languages.



The smart factory is a plant where sensors are installed in factory facilities and machines. It also collects and analyzes data in real time. It monitors all aspects of the factory and analyzes them and controls itself according to its purpose.

In recent years, as the number of sensors and data have increased, a fast and scalable solution for collecting and storing data has become necessary.

Attempts have been made to establish new business values that have not been found by collecting all the data and analyzing them in connection with existing systems in IT. Machbase database is a data solution suitable for storing and retrieving tag data generated in the facilities of smart factories.

Edge Computing with High Performance Time-Series Database 1. Era of Smart

Leave a Reply

Your email address will not be published.