TCP/IP
Initialism of Transmission Control Protocol/Internet Protocol, a standard for computer network communication, used in particular for the Internet.
TCP/IP: the foundational protocol pair that runs industrial networks
TCP/IP is not one protocol but two working together. IP (Internet Protocol) handles routing, breaking data into packets and sending them across networks to their destination addresses. TCP (Transmission Control Protocol) runs on top of IP and ensures those packets arrive in order and without loss, automatically resending any that go missing. Together they form the backbone of nearly all modern industrial network communication, from factory automation systems to remote monitoring and diagnostics.
IP works at layer 3 of the OSI model, assigning unique addresses (IPv4 uses 32-bit addresses, IPv6 uses 128-bit) to devices and determining how packets traverse routers and network segments. TCP operates at layer 4, establishing connections between applications, managing flow control, and handling retransmission of corrupted or lost data. Other protocols like UDP also ride on top of IP but skip TCP's reliability guarantees in favor of speed, making UDP common in real-time systems where occasional loss is acceptable but latency is critical.
Industrial deployment and variants
In industrial settings, TCP/IP runs over Ethernet cables and wireless links, forming the basis for protocols like Modbus TCP, EtherCAT, and PROFINET. Factory networks typically use private IP address ranges (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16) to isolate control systems from the public internet. However, this isolation can create security risks if devices expose services without authentication, and real-time control applications demand careful network design because TCP/IP offers no deterministic latency guarantees.
The protocol pair is connectionless at the IP level (each packet is independent) but connection-oriented at the TCP level (a virtual circuit is established before data flows). This mismatch means applications must account for network delays, packet reordering in rare cases, and the overhead of TCP's handshake and acknowledgment scheme. In legacy systems running fieldbus protocols like PROFIBUS or Modbus RTU, TCP/IP gateways bridge to Ethernet, allowing those older devices to speak to modern networks.
Failure modes include network congestion (TCP backs off and retransmits, raising latency), routing loops (causing packet loss until resolved), and duplicate packets arriving if acknowledgments are lost and retransmission occurs. Many industrial TCP/IP installations now add redundancy layers, such as Parallel Redundancy Protocol (PRP) or High-availability Seamless Redundancy (HSR), to eliminate single points of failure in mission-critical control loops.