System Integration

Calgary OSIsoft PI Experts and Calgary OSIsoft AF Experts. This is a deep area; we begin by differentiating data integration from application integration. Data integration is an essential need for all organizations so that data in their disparate data sources does not remain siloed. There are three main forms of data integration used by organizations: replication (copying the data as-is), consolidation (transforming source data into a condensed format for shared access), and virtualization (keeping the data in place but allowing access from other places).

Application integration is required when two or more applications must share information to enable or facilitate a certain business process. Both forms of integration may use similar techniques and software, but the intent of application integration is different than that of data integration. When we think of transactional data processing, we are thinking of application integration. Whereas when we think of moving large amounts of data for historical data analysis, we are talking about data integration.

Traditionally, the functionality of an application has been self-contained. However, as the scale of data grows (due to many trends in the industry, such as big data and IOT, and as companies try to take advantage of real-time data analytics, the singleton application design is evolving into distributed data integration patterns. We separately discuss this subject below.

From an operational data standpoint, organizations require both data integration and application integration. Oftentimes, due to data security and performance reasons, companies prefer to migrate data to another data store rather than allowing direct access to the historian (preferring data consolidation or replication as opposed to virtualization) although this is not always the case. As more sensor data is collected, companies are also increasingly looking at big data frameworks especially for the ability to scale streaming data, and to leverage cloud-based analytics.