Industrial Networking Protocols Comparison | TCP/IP, MODBUS, CAN, MQTT Selection Guide

1. Industry Pain Points & Technical Evolution Background

Industrial networking protocols are the core carrier of data interaction between field PLCs, sensors, gateways, and cloud platforms. With the upgrading of industrial automation from single-point control to full-network interconnection, diversified networking protocols coexist on-site, bringing a series of prominent technical pain points that restrict system stability:

• Blind protocol selection leads to performance mismatch: Many engineering teams uniformly adopt Ethernet TCP/IP for all industrial transmission scenarios. Low-speed equipment brings excessive protocol overhead, while high-frequency sampling equipment faces bandwidth bottlenecks, resulting in increased network latency and bus load rates up to 85%.

• Multi-protocol mixed networking causes docking failure: Traditional field buses (CAN/MODBUS) and Ethernet protocols (TCP/IP/MQTT) have different frame structures and addressing mechanisms. Unconfigured protocol conversion leads to data parsing errors and intermittent equipment offline.

• Lack of parameter standardization leads to poor stability: Unreasonable baud rate, frame length, and retransmission parameter configuration of networking protocols cause frequent packet loss, data disorder, and link flapping in strong electromagnetic interference industrial environments.

• Legacy and new equipment protocol iteration conflicts: Legacy industrial equipment relies on low-speed field buses, while new intelligent equipment adapts to high-speed Ethernet protocols. Incomplete protocol compatibility leads to high renovation costs and difficult system iteration.

Driven by the demand for high-speed, low-latency, and high-reliability industrial interconnection, mainstream networking protocols have formed a clear hierarchical and scenario-based differentiated pattern. Mastering the core differences and deployment specifications of each protocol is the key to solving industrial network instability.

2. Core Technology & Underlying Architecture Analysis

Current industrial mainstream networking protocols are divided into two major systems: field bus protocols and Ethernet Internet protocols. Field buses represented by CAN, CAN-FD, and MODBUS RTU are suitable for short-distance on-site device control; Ethernet protocols represented by TCP/IP, MODBUS TCP, and MQTT are suitable for long-distance transmission and cloud interconnection. Each protocol has an independent underlying frame structure, transmission mechanism, and performance parameters.

Protocol Profile Overviews

1. TCP/IP Protocol Stack: Layer 3/Layer 4 core Ethernet protocol, with IP responsible for 1500Bytes MTU routing segmentation and TCP responsible for three-way handshake reliable transmission, supporting max 1Gbps bandwidth, suitable for cross-network segment large-data transmission.

2. MODBUS Protocol: Divided into RTU (serial port) and TCP (Ethernet). MODBUS RTU features a simple frame and low overhead, with a maximum single-frame data of 256Bytes; MODBUS TCP is encapsulated based on TCP, with standard port 502, which is the most common PLC control protocol.

3. CAN/CAN-FD Protocol: Automotive and industrial dedicated field bus. Classic CAN supports max 8Bytes single-frame data and 1Mbps baud rate; CAN-FD upgrades to 64Bytes single-frame data and 8Mbps data segment rate, with strong anti-interference and non-destructive arbitration features.

4. MQTT 3.1.1 Protocol: Lightweight application-layer IoT protocol based on TCP, adopting a publish-subscribe mode, with an ultra-small 2Bytes minimum header, low power consumption, and low overhead, specially used for industrial equipment cloud data reporting.

Multi-Dimensional Technical Parameter Comparison

The following technical parameter comparison table intuitively sorts out the core differences of all mainstream industrial networking protocols, covering key engineering indicators:

Core Protocol Interaction Rule: Industrial multi-protocol networking relies on gateway protocol conversion. Field bus data (CAN/MODBUS RTU) is encapsulated into Ethernet protocols (TCP/MQTT) through gateway parsing to realize cross-layer and cross-network data interconnection, and all conversion processes follow international standard protocol frame specifications.

3. Typical Engineering落地 (Deployment) Solutions

Solution 1: On-Site PLC Control Networking (MODBUS + CAN Hybrid Scheme)

• Applicable Scenario: Workshop production line PLC control, motor drive control, low-frequency switch signal acquisition, and other on-site short-distance control scenarios.

• Deployment Architecture: Adopt MODBUS RTU (115200bps baud rate) for master-slave polling of conventional PLC analog quantity control; deploy a CAN bus for high-priority equipment interlock control to utilize the multi-master arbitration mechanism and avoid single-point failure. The gateway unifies on-site bus data and converts it into the MODBUS TCP protocol for upper computer monitoring.

• Actual Engineering Effect: The on-site control bus load rate is controlled below 40%, the control command response delay is stabilized within 20ms, the equipment interlock misjudgment rate is reduced to zero, and the compatibility of old and new control equipment is fully realized.

Solution 2: High-Precision Sensor Data Acquisition (CAN-FD + TCP/IP High-Speed Transmission Scheme)

• Applicable Scenario: Industrial vibration, temperature, and pressure high-frequency array sampling, high-precision testing equipment data real-time transmission scenarios.

• Deployment Architecture: Use a CAN-FD bus for on-site sensor data collection, adopt 1Mbps arbitration segment + 8Mbps data segment dual-rate transmission, and make full use of 64Bytes ultra-long single-frame data to eliminate data fragmentation. The gateway encapsulates CAN-FD data into standard TCP/IP packets with 1500Bytes MTU for long-distance transmission.

• Actual Engineering Effect: The data sampling frequency is increased by 75% compared with the classic CAN scheme, the data fragmentation rate is reduced to zero, and the long-distance transmission packet loss rate is $\le$ 0.01%, which meets the high-precision and high-frequency data acquisition requirements of industrial testing.

Solution 3: Industrial Equipment Cloud Interconnection (MQTT + TCP/IP Cloud Transmission Scheme)

• Applicable Scenario: Remote equipment status monitoring, production data cloud reporting, remote parameter debugging, and other IIoT cloud networking scenarios.

• Deployment Architecture: Based on a TCP/IP Ethernet underlying bearing, adopt the MQTT 3.1.1 lightweight publish-subscribe protocol for cloud data interaction. Set a 30s heartbeat packet cycle, enable the QoS1 message confirmation mechanism, and realize stable uploading of equipment logs, operating status, and fault data.

• Actual Engineering Effect: The cloud transmission protocol overhead is reduced by 60% compared with a pure HTTP scheme, the remote data online rate reaches 99.98%, and the remote debugging response delay is controlled within 35ms, realizing efficient and stable interconnection between industrial field equipment and cloud platforms.

4. Selection & Deployment Best Practices (Expert Troubleshooting Guide)

Summarized from thousands of industrial multi-protocol networking debugging cases, follow these three core protocol selection and deployment standardized specifications:

1. Scenario-Based Protocol Strict Selection Specification: For low-speed on-site PLC logic control, prioritize MODBUS RTU/TCP; for high-real-time industrial and vehicle interlock control, select CAN/CAN-FD; for cross-network large data transmission, adopt TCP/IP; for equipment cloud interconnection, you must use the lightweight MQTT protocol to avoid resource waste.

2. Multi-Protocol Mixed Networking Isolation Rule: When field bus and Ethernet protocols coexist, deploy industrial protocol gateways for isolated conversion and prohibit direct transparent transmission of heterogeneous protocol data. Uniformly standardize baud rate, MTU, and heartbeat parameters to prevent frame confusion and bus parsing exceptions.

3. Anti-Interference Parameter Optimization Specification: In industrial environments with strong electromagnetic interference, reduce the CAN-FD data segment rate from 8Mbps to 5Mbps appropriately, set the TCP retransmission timeout to 200ms, and enable the MQTT QoS1 confirmation mechanism to effectively avoid data bit errors and packet loss caused by environmental interference.

5. Frequently Asked Questions (FAQ)

Q1: What are the core differences between mainstream industrial networking protocols, and how do I distinguish application scenarios?

A: Field buses (CAN/MODBUS RTU) are suitable for short-distance, low-latency on-site equipment control; Ethernet protocols (TCP/IP/MODBUS TCP) are suitable for high-bandwidth local area network transmission; MQTT is dedicated to cloud remote interconnection. CAN-FD is preferred for high-frequency sampling scenarios, while classic CAN and MODBUS are suitable for conventional low-speed control scenarios.

Q2: Why do multi-protocol mixed industrial networks often suffer from data parsing errors?

A: Different networking protocols have independent frame headers, verification mechanisms, and addressing logic. Direct mixed transmission without gateway protocol conversion will cause equipment to fail to identify frame formats, resulting in parsing errors and data loss. Heterogeneous protocol networking must be equipped with professional protocol conversion equipment for data encapsulation and parsing.

Q3: Which protocol has the strongest anti-interference performance for industrial on-site networking?

A: CAN-FD has the best comprehensive anti-interference performance, relying on differential signal transmission and 21-bit high-precision CRC verification, which adapts well to high-power frequency converters and strong electromagnetic interference environments. This is followed by classic CAN, MODBUS TCP, and TCP/IP, while the MODBUS RTU serial protocol has the weakest anti-interference ability.

Q4: How can I solve high bus load and network congestion in industrial networking?

A: Optimize protocol matching by replacing low-efficiency protocols with high-speed CAN-FD for high-frequency data. Additionally, split bus segments to reduce single-node load, adjust TCP retransmission parameters and the MQTT heartbeat cycle to minimize invalid protocol overhead, and shield redundant polling instructions from the MODBUS master station to reduce the overall bus occupation rate.