Home > Workload Solutions > Data Analytics > Guides > Ready Solutions for AI & Data Analytics: Edge Analytics for Industry 4.0 with Confluent Platform > Traditional approaches
The manufacturing industry is trying to increase their analytics capability by extending the deployment of familiar technology for application and data integration. These technologies are based on traditional data messaging and queuing technology, primarily MQTT.
The advantages of MQTT include:
The design of MQTT predates the massive deployment of data generating and consuming devices that are encountered when implementing Industry 4.0. There are some significant limits to the technology. The ability to work asynchronously for long periods of network unavailability limit its usefulness for near-real time applications. Many enterprise applications have limited support for MQTT, hindering its ability to be a common data exchange technology. Many MQTT broker implementations have limited scalability, and therefore do not handle systems with many clients or large bursts of messages.
A second approach for data management and integration is based on a hub and spoke approach. The hub of the data storage and access is frequently called a data historian. PR Newswire predicts that the data historian software market will grow to $1.2 billion by 2023. (https://www.prnewswire.com/news-releases/data-historian-market-worth-1-271-1-million-by-2023-845765143.html) The data historian software category was also developed long before the advent of Industry 4.0. There is a wide array of vendor offerings in the market with only a small number with a significant market share. These products were typically evaluated, purchased, and operated by the OT organization without significant integration with IT systems. This lack of integration capability is the single largest challenge with data management strategies that are built around data historians. Many of the same challenges that IT has had implementing large-scale data lakes also apply to data historians.
The market and products have evolved over time. Data Historians have expanded into many different industries and applications where time series data is prevalent. They have evolved from single site, on-premises implementations to multisite cloud-based implementations. An ecosystem of System Integrators and software providers has emerged to implement these systems and create value added applications on top of Data Historians.
As with any software application category that is over 35 years old, pundits have predicted the demise of Data Historians since before the advent of Industry 4.0. The consensus that is emerging is that Data Historians will remain important to process engineers and OT for a long time. However, they are unlikely to solve any of the new challenges being surfaced in the move to developing more advanced analytics. Better data integration between OT and IT silos is required. It must be dynamic, and better integrated with technologies from the world of big data and data science.