Historians: Initial Questions and Design Considerations
A practical checklist for planning Intelligence Hub pipelines with historians.
What Does This Article Cover?
This article provides essential guidance for designing reliable, scalable historian pipelines in Intelligence Hub - review these questions before you build.
Helpful Questions
Network and topology:
- Where on the network is the historian located?
- Is the historian obtaining data from one site, or is it an aggregation point obtaining data for many sites?
- Is the Intelligence Hub installed on the same network as the historian?
Scope and volume:
- How many points or tags are in the scope of the solution?
- Generally, how frequently are points or tags' values changing?
Latency, resolution, and cadence:
- What is the expected or required latency and resolution of the data in the destination system?
- Are all value changes needed, or is polling or aggregation acceptable? At what rate will data be obtained?
History, late data, and metadata:
- Is obtaining late-arriving data in scope?
- Is historical backfill of data required?
- Is obtaining point or asset metadata in scope?
- How often are new tags or points edited, and must the solution automatically reflect these changes?
For Aveva PI System Framework (AF):
- Is PI Asset Framework being used, and are the assets well organized?
- How many assets are configured in PI Asset Framework?
- How many root-level assets are configured in PI Asset Framework?
- How many asset templates are configured in PI Asset Framework?
- How many levels of asset hierarchy are configured in PI Asset Framework?
- Generally, what is the mix of PI Asset Framework asset attributes associated with PI Points versus Static Values?
General Intelligence Hub design considerations
These considerations pertain to most historian-related data pipeline solutions.
- Historian-related solutions can demand significant resources. Plan for ongoing optimization to ensure efficient resource utilization.
- The Connection Agent should be installed on the same network as the historian.
- The host resource should have at least 16 to 32 GB of RAM and appropriate processors.
- Consider optimizing JVM heap memory.
Design Approach
- Start by querying with a small number of tags or points (10s or 100s). Avoid time-consuming queries.
- Frequent short-duration (many seconds) Connection Input queries are optimal.
- Intelligence Hub Track Activity and Debug should not be used beyond 10s or 100s of tags or points.
- Pipeline polling rate should exceed the average Pipeline execution time by a reasonable safety buffer (perhaps 1.5 to 2 times).
- It may be necessary to utilize parallel Intelligence Hub Connections, multiple Connection Agents, or multiple Intelligence Hub deployments.
Intelligence Hub configuration options (performance tuning)
- Connection Output store and forward likely needs to be turned off.
- Dynamic references on Connection Outputs should be avoided, and turn Breakup Arrays off.
- “Wait for Completion” on Connection Outputs may need to be turned off.
Summary
Please utilize these questions and considerations to frame your historian integration, size resources appropriately, and choose the right query cadence and configuration options as you implement your Intelligence Hub solution.