Historians and Time Series: Add Alongside, Don't Replace, in Automotive

On October 5, 2026, IIoT World published sponsored content from InfluxData promoting a 22-page guide, "Scaling Time Series for Automotive Manufacturing." Its message is that automotive plants should add a time series layer alongside legacy historians rather than replace them. One thing should be clear from the start: this is promotional material, not independent journalism. The claims and case studies come from the vendor and have not been verified; we have also seen only the summary page, not the guide itself.

The problem described

According to the summary, the data volumes of modern assembly lines exceed what traditional historians were designed for. The issues it lists are:

• per-tag licensing turns every new sensor into a cost decision;
• proprietary architectures confine quality data in plant-level silos;
• batch export workflows slow down the real-time analytics that predictive maintenance needs;
• replacing the historian is risky, because a line stoppage can translate within hours into missed takt, JIT penalties and OEM downtime.

The proposed approach: two models

The guide proposes keeping the historian in service and growing new data capabilities in parallel. It describes two adoption models.

• Parallel feed: Telegraf, an open source collector with plugins for OPC-UA, MQTT and Modbus, sends data from new sensors to InfluxDB alongside the historian. New instrumentation, robots and launch programs thus avoid per-tag licensing.
• Historian consolidation: data from multiple disconnected historians flows into a single queryable layer, useful for cross-plant benchmarking and warranty investigations.

The example data flows span the whole process: tonnage, die temperature and vibration in stamping; KUKA/FANUC robot joint data and spot-welding current in the body shop; E-Coat tank levels, ovens and HVAC in paint; Atlas Copco nutrunners, AGV routes and end-of-line bench tests in final assembly.

Integrations and cited cases

The integrations named include PTC ThingWorx (which uses InfluxDB as its persistence layer), Kepware to InfluxDB 3, Rockwell Allen-Bradley PLC data via FactoryTalk Optix and View SE, Litmus Edge for tag mapping and asset context, and an MCP Server that gives AI tools natural-language access to production data. The cases reported by the vendor are Toyo Tires (from 100 machines at one plant in Serbia to roughly 1,000 machines across multiple sites, with 20-year retention for audits), Gotion (chosen for compression and query language) and American Axle & Manufacturing (nearly 85 sites in 18 countries).

Why it matters for brownfield plants

Beyond the product, the pattern is common in existing plants: a parallel data path, built on open protocols (OPC-UA, MQTT) into a time series store next to the historian, avoiding a rip-and-replace on validated lines certified to IATF 16949. The benefit is lower operational risk and decoupling new use cases, such as analytics and AI, from licensing constraints.

The summary does not address the trade-offs: the complexity of two systems, data consistency and governance.

What to assess in practice

Before adopting a parallel model, it is worth clarifying:

• which system is the source of truth for each tag, and how discrepancies between the two stores are handled;
• naming rules, asset context and time synchronization, so that data from different sites is comparable;
• ownership, access, retention and audit requirements for data in the new layer;
• the collector's impact on the OT network and devices, tested on a non-critical pilot line;
• an exit strategy: open formats and protocols reduce the risk of replacing one proprietary lock-in with another.

In short, the idea of adding alongside rather than replacing is sound, but the claimed benefits should be verified with a measurable pilot, not inferred from commercial material.