VCampus has always been an ideas incubator where people share new developments and best practices. Once of the things I was always interested in was real time analytics, where data are analyzed with very little latency. This is especially important in data driven process analytics, where data are used for making decisions on process changes.
A couple of months ago we developed a prototype to executes C# code in real-time using a calculation engine. The idea was to close the gap between site specific low-level calculations that can be performed in PI Analysis and high-level data analytics that typically resides on the Enterprise level. The demo was a good start and thanks to my manager Damien O'Connor we found a project that supports the effort and were able to productize it.
The main driver for the project was PI-ACE to AF migration as there are still a lot of ACE projects out there as the product is hitting the end-of-life cycle. So naturally the engine had to be extended to fully support VB.NET. We received a lot of ACE code that will need to be migrated and interestingly a lot of the code was around batch processing. Engineers have been leveraging ACE to trigger custom batches and modifying existing structures.
During the project we also had the opportunity to work with our intern Pablo Gomez. He investigated real-time operators that can now be called as extensions to a time series operation. In addition, we worked on R and Python implementation and are unit testing the integration. The data science GUI is very different from the NET programming and tailored to allow data scientist to call their functions easily.
In summary, it is really nice to see something 1st developed in VCampus make it into the real world and allow engineers to develop ultra fast real time solutions.
Here is a screen shot of the engine in AF:
We also set up a Github page for more information: