Home > Workload Solutions > Data Analytics > Guides > Ready Solutions for AI & Data Analytics: Edge Analytics for Industry 4.0 with Confluent Platform > Integrating existing infrastructure
Most applications that were developed and deployed during the Third Industrial Revolution were focused on improving automation and control. In those cases, where a new automation required integration with multiple systems, the communication was implemented only for those applications. It was rare for OT to invest in standardized enterprise class platforms for integration. The only option for process integration without a true platform option was to keep adding on to an evolving set of point-to-point message-based communications.
Development of each additional integration connection was done by itself as new equipment and applications were added. This approach kept things running with the least short-term disruption but has become overly complex and cannot support the data integration needs of Industry 4.0.
Figure 3: Complex legacy integration
The manufacturing sector is not alone in confronting a complex process integration environment that has developed over decades, and that does not readily support cross-functional analytics. Also, the health care industry has historically preferred to adopt many best-of-breed software applications. That adoption has resulted in data integration challenges, especially since the advent of IoT for patient care. The best of breed solutions for different hospital functions, all linked with multiple point-to-point integration connections to manage operations and control, include:
System downtime in health care is at least as costly as the manufacturing industry. Also, the health care industry must manage distributed application and data processing with several large corporate IT centers. These IT centers are connected to many regional and local hospitals. The health care industry has also realized the potential benefits from new advanced analytics and AI applications. These applications must leverage communication across the individual edge sites and processes within a site with connections to and from central data centers.
Given the many parallels, it is worth studying the data management and advanced analytics approach that many health care companies have adopted. The industry has made significant progress over the last 20 years in reducing the complexity of application integration. Progress has also been made in the incorporation of new sources of data from IoT devices, by implementing message-based middleware. (Bellagente, P., Depari, A., Ferrari, P., Flammini, A., Sisinni, E., & Rinaldi, S. (2018, May). M 3 IoT-message-oriented middleware for Mhealth Internet of Things: Design and validation. In 2018 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) (pp. 1-6). IEEE. )
The use of message-based middleware in the industrial sector would have many parallels to health care. The typical industrial plant will have many production lines that may share some equipment, raw material inventory, packing and shipping services. Separate control systems, with isolated data management, control all these components.
Figure 4: Simple modern integration
As with health care, the switch to a central messaging or event streaming platform is incremental usually starting with two or at most a few systems. The risk of system disruption that would come with converting all integrations to run through a central platform is too great. The flexibility of the Kafka platform in handling both integration using traditional messaging and high velocity streaming has been valuable in both health care and manufacturing applications.