Introduction
Welcome to this presentation on OPC redundancy. Today we’ll be covering the problem we’re trying to solve, the solution, and an introduction to OPC Expert redundancy. We’ll walk through the architecture and the three redundancy modes: hot, warm, and cold. At the end, I’ll let you know how to get the webinar recording and where to download the software.
The Problem: Why OPC Redundancy Matters
In automation, systems are operating 24/7 and are expected to work 100% of the time. OPC server failures are very possible. A user might shut down Windows unexpectedly, kill an application, or the software itself might fail due to a bug. Hardware failures are also a factor, as are resource constraints: another application consuming too much CPU, low memory, or a very busy hard drive. And in some cases, there’s simply no network access between your application and the OPC server.
In all of these cases, redundant data sources provide a reliable communication path. If the primary path fails, the system moves over to the secondary automatically.
Some applications already support redundancy natively. If your application connects directly to OPC servers and handles failover on its own, you don’t need anything else. But if your application does not currently have redundancy, you can retrofit it with OPC Expert and benefit from redundancy immediately, with very little or no data loss. You can automatically move communication from a primary to a secondary path, and you can do this on existing applications.
What the Redundancy Solution Must Do
A redundancy solution needs to detect any kind of failure: network issues, hardware problems, software crashes, or data that has stopped changing when it should, or changed when it shouldn’t have. After detecting a failure and failing over to the standby path, it also needs to restore the original connection automatically once it’s available again. Throughout all of this, it must communicate the exact communication status to users clearly. And it must run as a Windows service so it starts automatically, requires no manual login, and is restarted by Windows if anything goes wrong.
OPC Protocol and Platform Support
The solution supports OPC UA, OPC Classic, DCOM, and OPC tunneling. It supports all security profiles, connections to both local and remote OPC servers via DCOM, TCP/IP, HTTP, and HTTPS, and full Windows compatibility from Windows XP through Windows 11 and from Windows Server 2003 through Windows Server 2025.
OPC Expert Architecture
When you download OPC Expert, here’s what’s inside. It includes an OPC DA client and an OPC UA client for data ingestion. That data flows into OPC Expert’s intelligent engine and comes out via OPC DA and OPC UA servers. This means you can have OPC DA coming in and OPC UA going out, or vice versa, or DA in and DA out. OPC Expert handles all of it within the engine.
Beyond OPC, OPC Expert can also output data via Excel, MQTT (with a built-in broker including publisher and subscriber), SQL databases, Azure, Amazon AWS, and REST API. There are many ways to get data in and out of OPC Expert, and all of this is contained within the single product.
For this webinar, the focus is on OPC in and OPC out, specifically adding redundancy capabilities to applications that don’t already have them.
How Retrofit Redundancy Works
Your application connects to a local OPC server that acts as a mirror of the remote server. The application doesn’t need to worry about network connections or failures because OPC Expert handles all of that. OPC Expert then connects directly to the primary and standby OPC servers via OPC UA or OPC DA. If tunneling is required for any reason, OPC Expert includes a built-in tunnel as well. Everything needed to bring redundant data into your application is included within OPC Expert.
Redundancy Modes
Hot Redundancy
In hot redundancy, both the primary and secondary OPC servers are actively sending data to OPC Expert at the same time. Both servers are collecting data from the PLC simultaneously. The key advantage is zero data loss: if one server fails, data continues to flow from the other with no interruption. The disadvantage is that both servers are drawing data from the PLC at the same time, which some PLCs or architectures cannot support. Hot redundancy also works with or without tunneling.
Warm Redundancy
In warm redundancy, there is a primary and a standby. OPC Expert is connected to the standby server and monitoring it, but it is not requesting data from it. The PLC is only serving one OPC server at a time, which means roughly half the traffic. The disadvantage is the potential for some data loss: it takes a small amount of time to detect a failure and then additional time for the standby server to start collecting data. Failover can take five to ten seconds depending on how quickly the standby OPC server responds. If that window is a concern, hot redundancy is the better choice. Warm redundancy also supports tunneling.
Cold Redundancy
In cold redundancy, the standby OPC server is not running at all until the primary fails. It may not be activated, or the machine may not even be powered on. Only when the primary path fails does OPC Expert start and connect to the standby. The advantages are reduced traffic and compatibility with architectures that do not support two OPC servers running simultaneously. For example, Honeywell TDC 3000 and Honeywell Experion only allow one OPC server running at a time, making cold the only viable option. The disadvantage is a longer failover time, as the standby server must start up before data can resume. If data continuity is critical, warm or hot redundancy is preferable. Cold redundancy also works with or without tunneling.
Comparison Summary
Hot redundancy is recommended whenever the architecture supports it, as it provides zero data loss. If hot is not possible, warm is the next best option. If warm is not possible, cold is supported as well. OPC Expert supports all three modes; you simply select the one you need and OPC Expert executes it.
Live Demonstration: Hot Redundancy Setup
The demonstration uses three computers: one for the application, one acting as the primary OPC server, and one as the secondary. The following steps walk through how OPC Expert was configured from scratch.
Installing OPC Expert as a Windows Service
When you download OPC Expert, you receive a single file. When you run it, it extracts all required files and folders automatically. On each server machine, you right-click OPC Expert and select Run as a Service, then choose to run it under a system account. Once confirmed, OPC Expert registers itself as a Windows service. The GUI is no longer needed after this point; the service runs in the background. A system tray utility is also available for quick status monitoring. OPC Expert also provides a command-line interface, so the entire setup can be scripted with a batch file for environments with multiple installations.
Configuring Redundancy
On the application computer, OPC Expert is launched and the two server computers are added: one labeled as primary, one as secondary. Setting up redundancy is done by dragging the primary OPC server and dropping it onto the secondary. That single drag-and-drop action creates the redundancy configuration. You then select the mode (hot, warm, or cold) and choose whether to use tunneling. After clicking OK, OPC Expert creates the necessary services and local OPC servers automatically. The item names visible through the local OPC server are identical to those on the source servers.
Diagnostic Information
OPC Expert includes extensive diagnostic data. The most relevant items during a redundancy test are the server action for the primary, the server status for the primary, the server action for the secondary, and the server status for the secondary. Under normal operation with both servers connected, these will show an asynchronous status with no active remediation action.
Fault Injection Tests
Several failure scenarios were demonstrated. When the network cable was disconnected from the primary, OPC Expert immediately began attempting to reconnect, retrying continuously until the cable was reconnected. Data continued to flow throughout. When the cable was disconnected from the secondary, all real-time data continued arriving from the primary with no delays or interruptions. When the OPC Expert UA server process was killed on one machine, Windows detected the service had stopped and restarted it automatically. OPC Expert reconnected once the service was back online. When the service was stopped from the service manager rather than killed, Windows did not restart it automatically, but OPC Expert continued retrying the connection and reconnected as soon as the service was manually restarted.
In a final test, both OPC servers were killed simultaneously. Data flow briefly stopped, then resumed immediately once OPC Expert autonomously restarted the servers. Killing only one OPC server at a time had no visible effect on data flow whatsoever.
Additional Tools and Training
For automated troubleshooting, OPC Rescue is available at opcrescue.com. OPC Expert is available at opcexpert.com and supports trending, bridging, tunneling, archiving, and redundancy, among many other capabilities.
Hands-on training classes are available for those who want to earn a Certified OPC Professional designation. The curriculum covers configuring OPC Classic with DCOM and OPC UA, setting up security, troubleshooting OPC systems, applying security best practices, and configuring firewalls and user accounts. Each participant works with two computers and connects over a live network. Instructors alternate between getting communication working and deliberately breaking it, reinforcing troubleshooting skills in a practical setting.
Upcoming workshops are available as live remote classes with a live instructor. These are interactive and not recorded. Public classes include participants from multiple companies. Private training is also available for teams within a single organization. All classes can be delivered remotely or onsite.
What You’ll Receive After This Webinar
An email will be sent with a link to the webinar recording, a link to download OPC Expert, and links to OPC redundancy examples showing how the configurations work in practice. You’ll also receive an invitation to the next webinar covering OPC bridging, which addresses how to transfer data between OPC servers, including between OPC DA and OPC UA servers or across multiple servers with large point counts. That webinar is scheduled for August.
For questions, contact info@cti.com or call directly.
Questions and Answers
