# Polling vs Subscription in OPC UA

## A Practical White Paper for Industrial Automation Systems

### Using C# and the TRAEGER OPC UA SDK

---

# Executive Summary

OPC UA (Open Platform Communications Unified Architecture) provides multiple mechanisms for obtaining data from industrial devices and systems. The two most common approaches are:

1. **Polling** (Client-driven data acquisition)
2. **Subscription** (Server-driven event/data notification)

Choosing the correct approach is critical for:

- System scalability
- Network performance
- CPU usage
- Latency
- Real-time responsiveness
- Reliability
- Application complexity

This white paper explains both methods in depth using practical C# examples based on the TRAEGER OPC UA SDK.

The document is intentionally written in a Markdown + AI/RAG friendly format:

- Small sections
- Explicit headings
- Short code samples
- Semantic explanations
- Direct pros/cons tables
- Minimal ambiguity

---

# 1. Introduction

Industrial systems increasingly depend on reliable real-time communication between:

- PLCs
- Digital Front Ends (DFEs)
- Print Controllers
- HMIs
- MES systems
- SCADA systems
- Cloud analytics platforms

OPC UA has become one of the dominant communication standards for Industry 4.0 because it provides:

- Platform independence
- Secure communication
- Rich data models
- Reliable messaging
- Scalable architectures
- Cross-vendor interoperability

Two primary approaches exist:

| Method | Direction | Typical Usage |
|---|---|---|
| Polling | Client asks repeatedly | Slow-changing data |
| Subscription | Server pushes updates | Real-time monitoring |

---

# 2. Polling vs Subscription

## Polling

Polling is a repetitive read operation where the OPC UA client periodically requests values from the server.

### Basic Polling Flow

```text
Client --> Read Request --> Server
Client <-- Value Response <-- Server
```

### Polling Characteristics

| Characteristic | Description |
|---|---|
| Client driven | Client controls timing |
| Simple | Easy to implement |
| Predictable | Fixed intervals |
| Higher network load | Many repeated requests |
| May miss fast changes | Depends on interval |

---

## Subscription

Subscriptions allow the server to automatically notify the client when values change.

### Basic Subscription Flow

```text
Client --> Subscribe --> Server
Server --> Data Change Notification --> Client
```

### Subscription Characteristics

| Characteristic | Description |
|---|---|
| Server driven | Server detects changes |
| Efficient | Minimal traffic |
| Event based | Notifications only when needed |
| Lower latency | Near real-time updates |
| More complex | Requires lifecycle handling |

---

# 3. C# Polling Example using TRAEGER SDK

```csharp
using Opc.UaFx.Client;

var client = new OpcClient("opc.tcp://localhost:4840");

client.Connect();

while (true)
{
    var value = client.ReadNode("ns=2;s=Machine/Temperature");

    Console.WriteLine($"Temperature: {value.Value}");

    Thread.Sleep(1000);
}
```

---

# 4. C# Subscription Example using TRAEGER SDK

```csharp
using Opc.UaFx.Client;

var client = new OpcClient("opc.tcp://localhost:4840");

client.Connect();

client.SubscribeDataChange(
    "ns=2;s=Machine/Temperature",
    (sender, e) =>
    {
        Console.WriteLine($"Temperature changed: {e.Item.Value}");
    });

Console.ReadLine();
```

---

# 5. Polling Pros and Cons

## Advantages

| Advantage | Explanation |
|---|---|
| Simple implementation | Easy for beginners |
| Full timing control | Client decides frequency |
| Deterministic behavior | Fixed update cycles |
| Easy debugging | Explicit reads |
| Good for snapshots | Useful for reports |

## Disadvantages

| Disadvantage | Explanation |
|---|---|
| High network traffic | Constant reads |
| Wasted bandwidth | Reads unchanged values |
| Higher CPU usage | Repeated processing |
| Poor scalability | Large systems overload quickly |
| Missed fast events | Changes between polls lost |

---

# 6. Subscription Pros and Cons

## Advantages

| Advantage | Explanation |
|---|---|
| Low network usage | Updates only on change |
| Real-time notifications | Immediate updates |
| Scalable | Efficient for many nodes |
| Reduced CPU load | Less repetitive work |
| Better responsiveness | Event-driven model |

## Disadvantages

| Disadvantage | Explanation |
|---|---|
| More complex | Lifecycle management required |
| Requires reconnect logic | Must restore subscriptions |
| Harder debugging | Asynchronous behavior |
| Queue management needed | Burst updates possible |
| Server resource usage | Server tracks subscriptions |

---

# 7. Performance Comparison

| Category | Polling | Subscription |
|---|---|---|
| Network Efficiency | Poor | Excellent |
| CPU Usage | Higher | Lower |
| Latency | Depends on interval | Near real-time |
| Scalability | Poor | Excellent |
| Complexity | Simple | Moderate |
| Event Capture | Weak | Excellent |

---

# 8. Industrial Best Practices

## Use Polling For

- Static configuration
- Scheduled reports
- Startup validation
- Rarely changing values

## Use Subscription For

- Alarms
- Telemetry
- Live dashboards
- Sensor monitoring
- Real-time systems

---

# 9. Hybrid Architectures

In modern industrial systems, polling and subscriptions are often combined.

| Data Type | Recommended Method |
|---|---|
| Temperatures | Subscription |
| Alarms | Subscription |
| Configuration Data | Polling |
| Historical Statistics | Polling |

---

# 10. Final Recommendation

## Prefer Subscriptions When

- Real-time monitoring matters
- Systems are large
- Data changes frequently
- Scalability is required

## Prefer Polling When

- Data changes slowly
- Simplicity matters
- Small systems are involved
- Scheduled reads are sufficient

---

# References

- https://documentation.hybridhelix.com/sdfe/opc-ua-interface-guide-smart-connectivity
- https://docs.traeger.de/en/software/sdk/opc-ua/start
- https://github.com/Traeger-GmbH/opcuanet-samples