Tune PI Tag Compression Settings Automatically
“A tag tuning tool for the OSIsoft PI system that recommends optimized compression settings based on evaluation and analysis of real time signal characteristics of PI tags.”
COMPRESSIONINSIGHT 3.0 NOW AVAILABLE!
A Data Driven Approach to Compression Setting for the OSIsoft PI System:
As experts in data analysis, Pattern Discovery Technologies understands that capturing accurate data in your PI System is a must. To manage the performance of the OSIsoft PI System, there are several tuning parameters that govern the collection and storage of data. Most significant among these are the parameters that govern the compression of the data captured at the interface and again when it is stored in the archive. If these settings are not correct, the archived data will not accurately reflect the original signal, or conversely, will store unnecessary volumes of data. Since the PI Data Archive is the fundamental source for retrieving historical records, the exception and compression settings have a significant impact on the quality of the data stored and the performance of downstream applications relying on it. “Getting it right” is critical to the delivery of meaningful, useful information.
Particular attention should be paid to defining the Exception deviation (ExcDev) and Compression deviation (CompDev) settings in the PI System configuration. This traditionally requires an intimate knowledge of the process combined with PI system administration capabilities. Given the overwhelming and time-consuming task of defining thousands of tags, users often rely on static “rules of thumb”, default settings or simple cut and paste exercises. Sadly, this may not take into account the true characteristics of a specific signal, and a “set and forget” philosophy does not consider the dynamic nature and behavior of the data. Getting these parameters right and keeping them tuned throughout the life of the tag is critical to the quality of the data stored in the archive.
Download Your Free Trial Today
To aid with this, Pattern Discovery Technologies has developed an automated tag tuning and tag behavior identification software for the OSIsoft PI System called CompressionInsight™. The foundation for CompressionInsight is an Information Measure we call “Data Fidelity”. The software analyzes an uncompressed data feed to the PI server (raw snapshots) and compares this signal to recreated archived values (as currently configured in the PI System). Using principles founded in Information Theory, we determine how much information would be retained in the PI archive and express this as a percentage of the raw snapshot information. Then, using this same measure, we proceed to make an optimized recommendation for revised compression and exception parameters considering the balance between Data Fidelity and data storage.
The goal is to store data efficiently and effectively without sacrificing accuracy. This establishes a high degree of confidence in the data stored while minimizing performance issues due to bandwidth and data storage. This optimization methodology allows us to make automatic recommendations for both over-compressed tags (removing too many data points) or under-compressed tags (retaining too many data points).
With CompressionInsight you will be able to:
- Evaluate existing compression configurations for data fidelity
- Review automatic recommendations that maximize data fidelity and minimize data storage
- Tune thousands of PI tags concurrently
- Create detailed summaries of compression analysis for review and updating
- Review graphic visualizations of compression analysis results for individual PI tags
- Identify tags demonstrating tag behavior issues that do not respond well to compression
- Decrease time spent managing PI System tag configurations
- Enhance confidence in archived PI tag data
CompressionInsight makes PI Tag Tuning Easy
OSIsoft’s PI System is the industry standard for managing real-time data and events and provides an infrastructure that has been specifically designed to collect and store large volumes of time series data. However, the user of the PI System must balance between collecting enough data to be useful but not collecting too much which could impact downstream applications of the data. Proper compression configurations ensure that performance (both from a data storage and data access capacity) is maximized and redundant data is minimized. But without a feedback loop, how do you get it right?
CompressionInsight uses an unbiased data driven approach which allows the user to evaluate these compression settings using the dynamic nature of the original signal. It then proceeds to make automatic recommendations which optimize the fidelity of the data while taking into account the data volume going to the archive. Patented analytics are used to measure the amount of information captured in the archive for a given set of PI tag compression settings. Thousands of PI tags can be analyzed concurrently, making the process of PI tag tuning effortless. CompressionInsight is integrated with the OSIsoft PI System so PI tags can be automatically updated from the software or the information can be shared in report format for collaboration and updating offline using PI Datalink.
CompressionInsight ensures that downstream analysis tasks can be performed with confidence, knowing that the compressed data accurately reflects the characteristics of the original PI tag signal.
Download Your Free Trial Today
To get more information on CompressionInsight, you can view our brochure online, download a copy below, or visit our partners, OSIsoft, for more information about PI tag tuning.