From OPC UA to TigerData Cloud: Idako Performance Test

Overview

This post documents a performance test case for Idako — a store & forward data collection solution by One Way Automation — connected to a cloud-hosted TigerData instance. The goal was to evaluate how well Idako handles high-frequency industrial data at scale: collecting OPC UA process values and reliably forwarding them to the cloud with minimal latency and no data loss.


Test Setup

Data Source

The simulated process data was generated using the OPC UA Demo Server by Unified Automation, a widely used reference implementation of the OPC UA specification. The server was configured to expose 20,000 variables, each updating every 200 milliseconds — equivalent to a 5 Hz sampling rate across the entire tag set.

This produces a theoretical throughput of:

20,000 variables × 5 updates/second = 100,000 values/second

Data Destination

The collected values were forwarded to a TigerData cloud instance. TigerData is a time-series data platform designed for industrial IoT use cases, providing ingestion, storage, and analytics for high-volume process data.

Collection and Forwarding Layer

Idako acted as the middleware between the OPC UA server and TigerData. Idako is a lightweight, edge-deployable application that:

  • Subscribes to OPC UA data sources
  • Buffers values locally (store) in case of network interruption
  • Continuously forwards buffered data to the target cloud platform (forward)

The store & forward architecture is critical for industrial deployments where network connectivity cannot be assumed to be constant. Idako ensures no data is lost during connectivity gaps — values are persisted locally and forwarded once the connection is restored.

Idako Configuration

The following screenshots illustrate how Idako was configured for this test.

Logged Variables — the list of 20,000 OPC UA variables subscribed and collected by Idako:

Idako — Logged Variables configuration

TigerData Connection Settings — the time-series database destination configuration, specifying the TigerData cloud instance endpoint and credentials:

Idako — TigerData TSDB configuration

Advanced JSON Settings — the low-level JSON configuration for the TigerData connection, exposing additional tuning parameters for the ingestion pipeline:

Idako — TigerData TSDB advanced JSON configuration

Instance Configuration — assigning the TigerData cloud connection as the forwarding destination for this Idako application instance:

Idako — Instance configuration with TigerData destination


Hardware and Network Configuration

OPC UA Demo Server — Windows Machine (Edmonton, Canada)

Component Specification
RAM 64 GB
Logical processors (threads) 16
CPU utilization ~57%
Memory utilization ~42 MB
OS Windows

The OPC UA Demo Server is relatively lightweight on memory (~42 MB), with the CPU load of 57% reflecting the work of maintaining 20,000 subscribed variables and serving change notifications at 200ms intervals.

Idako — oBox-M Industrial PC (Edmonton, Canada)

Component Specification
Hardware oBox-M industrial PC
RAM 32 GB
Logical processors (threads) 12
CPU utilization ~5% (across all 12 logical processors)
Memory utilization ~640 MB
Deployment Docker container

Idako runs as a Docker container on the oBox-M, an industrial-grade edge PC offered by One Way Automation. At ~5% CPU utilization across 12 logical processors, Idako's footprint is minimal, leaving substantial headroom for other workloads on the same device.
Idako connects to the OPC UA Demo Server over a local 1 Gb Ethernet network, and to the internet — and therefore TigerData — over a 1 Gb fiber connection.

TigerData Cloud Instance — Oregon Data Center, USA West

Component Specification
vCPUs 4
RAM 16 GB
CPU utilization ~55%
Memory utilization ~31% (~5 GB)
Location Oregon, USA (West Coast)

The TigerData cloud instance is ingesting and storing data received from Edmonton, Canada over the public internet. At 55% CPU and 31% RAM utilization, the instance handles the ~98,000 values/second ingestion rate with capacity remaining, and without becoming a bottleneck in the pipeline.

Network Topology Summary

[Unified Automation OPC UA Demo Server]  (Windows, Edmonton, Canada)
              |
         1 Gb LAN
              |
       [Idako on oBox-M]  (Docker, Edmonton, Canada)
              |
         1 Gb Fiber
              |
         Internet
              |
    [TigerData Cloud Instance]  (Oregon, USA West)

Results

The test was run long enough to accumulate over 6 billion values in the TigerData instance. The screenshot below shows the Idako statistics panel captured during the run.

TigerData Stats — Idako store & forward performance

Key Metrics

Metric Value
Data collection rate 98,390 values/second
Forwarding rate to TigerData 98,445 values/second
Total values stored & forwarded > 6,000,000,000
Values in local buffer (pending) ~8,000
Estimated forwarding delay < 1 second

Analysis

Collection rate of 98,390 values/second is very close to the theoretical maximum of 100,000 values/second, indicating that Idako keeps up with the full OPC UA subscription load with minimal overhead.

Forwarding rate of 98,445 values/second slightly exceeds the collection rate. This is expected behavior — when the forwarding pipeline catches up after any brief lag, it processes queued values slightly faster than new ones arrive, draining the local buffer.

Local buffer size of ~8,000 values is the most telling metric. At a collection rate of ~98,000 values/second, 8,000 buffered values represent roughly 80 milliseconds of backlog — well under one second. In practice, the forwarding delay is effectively real-time for industrial monitoring purposes.

6+ billion total values confirms the test ran over an extended period and that Idako maintained stable throughput throughout — there is no evidence of memory leaks, performance degradation, or data loss over time.


Conclusion

Idako demonstrated robust performance in a demanding store & forward scenario:

  • Near-theoretical throughput at ~98,000 values/second from a 20,000-variable OPC UA source at 200ms update intervals
  • Negligible forwarding lag — the local buffer stays at a fraction of a second's worth of data under steady-state conditions
  • Long-term stability confirmed by accumulating over 6 billion forwarded values without issue
  • Reliable delivery to TigerData cloud with collection and forwarding rates staying tightly matched

For industrial deployments requiring high-frequency data collection from OPC UA sources and reliable cloud forwarding, this test confirms that Idako combined with TigerData is a capable and production-ready stack.


Resources