What Data Looks Like in Each Storage Type

One of the clearest ways to understand why we need three different storage systems is to look at the data itself. The same machine can produce the same signals, but the way the data is stored and represented changes depending on the system.

1. UNS – Real-Time State in Context

UNS organizes data in a folder tree based on the ISA-95 hierarchy (Enterprise → Site → Area → Line → Cell → Asset). Each node publishes a JSON payload that contains the machine’s live data points and metadata.

Example path:

HighByte/Portland/Production/Line1/Mixer/Motor001

JSON Payload:

{
  "AssetID": "Motor001",
  "Current": 24,
  "Speed": 1530,
  "Vibration": 70,
  "State": "Operating",
  "MaterialID": "MAT006783",
  "TankFillPct": 86.3,
  "MixTime": 1045,
  "Timestamp": "2025-08-29T10:15:23Z"
}

• • It resembles a live broadcast of what’s happening now.
  • Perfect for HMIs, dashboards, MES orchestration, Andon boards
  • Not stored long-term — short-lived state + context

2. Historian – Machine Memory as Time-Series

In the Historian, the same tags (Current, Speed, Vibration) are stored as time-series values with timestamps.

Example table:

Timestamp	Tag	Value	Unit
2025-08-29 10:15:23	Current	24	A
2025-08-29 10:15:23	Speed	1530	RPM
2025-08-29 10:15:23	Vibration	70	mm/s
2025-08-29 10:15:24	Current	25	A
2025-08-29 10:15:24	Speed	1532	RPM

• • Feels like a heart monitor — capturing data second by second
  • Used for trends, SPC charts, golden batch comparison, predictive maintenance
  • Stored for 3–7 years at high resolution

3. Lakehouse – Enterprise Record of Production

In the Lakehouse, the machine’s data is integrated with business context (order, material, cost center, operator). It looks like a wide relational table used for analytics.

Example production events table:

Order ID	Asset ID	Material	Qty Produced	Scrap	Start Time	End Time	Operator	Shift	CostCenter
WO12345	Motor001	MAT-001	500	5	2025-08-29 08:00:00	2025-08-29 10:00:00	John Doe	Shift A	CC-1001

• • Feels like a business record of production history
  • Used for BI dashboards, forecasting, compliance reports, AI models
  • Stored for 7–10+ years with flexible schemas

Why It Matters

• • UNS shows what is happening right now (live JSON in context)
  • Historian shows how it changed over time (time-series memory)
  • Lakehouse shows what it means for the business (records + analytics)

Together, these three perspectives turn raw machine signals into insightful information for operators, engineers, and executives alike.