Inferensys

Glossary

Modbus TCP

Modbus TCP is a variant of the Modbus serial communication protocol that uses TCP/IP as its transport layer to facilitate master-slave communication between industrial electronic devices over Ethernet networks.
Modern WeWork hardware lab area with product team collaborating around AI device prototypes, 3D printer in background, dramatic industrial lighting with product sketches on glass walls.
PROTOCOL DEFINITION

What is Modbus TCP?

Modbus TCP is the Ethernet-based implementation of the widely adopted Modbus protocol, encapsulating serial communication frames within TCP/IP packets for industrial automation.

Modbus TCP is a variant of the Modbus family that uses TCP/IP as its transport mechanism, enabling master-slave communication between industrial electronic devices over standard Ethernet networks. It encapsulates a standard Modbus frame—consisting of a Modbus Application Protocol (MBAP) header and a Protocol Data Unit (PDU)—directly into a TCP segment for transmission.

Unlike serial Modbus variants, Modbus TCP eliminates the need for traditional checksums by relying on the underlying TCP stack for error detection, and it introduces a Unit Identifier to address devices behind a gateway. This protocol is foundational to Operational Technology (OT) environments, facilitating data acquisition from Programmable Logic Controllers (PLCs) and remote terminal units within modern SCADA architectures.

PROTOCOL ARCHITECTURE

Key Features of Modbus TCP

Modbus TCP adapts the legacy serial protocol for modern Ethernet networks, combining simplicity with TCP/IP reliability for industrial automation.

01

TCP/IP Encapsulation

Modbus TCP encapsulates the standard Modbus Protocol Data Unit (PDU) within a TCP/IP frame, replacing the serial checksum with the TCP stack's built-in error detection. The standard Modbus Application Protocol (MBAP) header adds a Transaction Identifier and Unit Identifier to route messages across Ethernet networks. Communication occurs over port 502, the IANA-assigned standard port for Modbus TCP traffic.

Port 502
Standard Port
02

Client-Server Model

Unlike the master-slave terminology of serial Modbus, Modbus TCP operates on a client-server architecture. The client initiates requests, and the server processes them and returns responses. A single client can communicate with multiple servers simultaneously, and multiple clients can access the same server. This model supports concurrent transactions through unique transaction identifiers, enabling parallel data exchange across the network.

Concurrent
Transaction Support
03

Function Code Operations

Modbus TCP retains the standard function codes for data access and control. Key operations include:

  • Read Coils (FC 01): Read discrete outputs
  • Read Holding Registers (FC 03): Read 16-bit analog values
  • Write Single Register (FC 06): Write to a single holding register
  • Write Multiple Registers (FC 16): Write a block of registers These function codes are inspected by deep packet inspection systems for anomaly detection.
04

Connectionless Communication

Modbus TCP is fundamentally connectionless at the application layer. Each request-response pair is independent, with no session state maintained between transactions. The TCP layer handles reliable delivery, but the Modbus protocol itself does not require persistent connections. This design simplifies implementation but requires stateful whitelisting in security architectures to validate the logical sequence of commands against the industrial process state.

05

Data Model Addressing

The Modbus data model defines four primary data tables accessible via function codes:

  • Discrete Inputs: Read-only single-bit values from field devices
  • Coils: Read-write single-bit outputs
  • Input Registers: Read-only 16-bit analog measurements
  • Holding Registers: Read-write 16-bit configuration and control values Each table supports up to 65,536 addresses, providing extensive data mapping for complex industrial systems.
MODBUS TCP CLARIFIED

Frequently Asked Questions

Clear, technically precise answers to the most common questions about the Modbus TCP protocol, its security implications, and its role in modern industrial control systems.

Modbus TCP is a variant of the Modbus serial communication protocol that uses TCP/IP as its transport layer to facilitate master-slave communication between industrial electronic devices over Ethernet networks. It works by encapsulating a standard Modbus Protocol Data Unit (PDU) within a TCP/IP packet, adding a 7-byte Modbus Application Protocol (MBAP) header that includes a transaction identifier, protocol identifier, length field, and unit identifier. The master (client) initiates a request, and the slave (server) responds. The default TCP port is 502. Unlike Modbus RTU, Modbus TCP does not use a CRC checksum, relying instead on the TCP/IP stack's error detection. This allows for high-speed, routable communication across standard IT infrastructure, making it the dominant protocol for integrating PLCs, HMIs, and SCADA systems in modern automation.

PROTOCOL COMPARISON

Modbus TCP vs. Other Industrial Protocols

A technical comparison of Modbus TCP against DNP3, OPC UA, and IEC 61850 across key operational and security dimensions relevant to SCADA anomaly detection.

FeatureModbus TCPDNP3OPC UAIEC 61850

Transport Layer

TCP/IP over Ethernet

TCP/IP or Serial

TCP/IP, HTTPS, or SOAP

Ethernet (Layer 2) or TCP/IP

Communication Model

Master-Slave (Client-Server)

Master-Remote with Report-by-Exception

Client-Server with Pub/Sub

Client-Server with GOOSE Peer-to-Peer

Built-in Authentication

Built-in Encryption

Time Synchronization

Data Modeling

Flat Register Map (Coils, Holding Registers)

Flat Point Index with Data Types

Object-Oriented with Semantic Modeling

Abstract Logical Nodes and Objects

Typical Latency

< 10 ms

< 100 ms (polled)

< 50 ms

< 4 ms (GOOSE)

Primary Use Case

Simple monitoring and control in factories and buildings

Telemetry and control in electric and water utilities

Cross-platform industrial interoperability and IT/OT integration

High-speed protection and automation within substations

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.