Configuring Redundant Fault-Tolerant Plant Bus Networks in SIMATIC PCS 7

A redundant, fault-tolerant plant bus in SIMATIC PCS 7 requires two independent optical or electrical rings constructed with managed industrial Ethernet switches (such as SCALANCE X-200 or X-300) configured with Redundancy Managers (RM). A single ring is merely fault-tolerant against a single cable break, whereas a double ring provides full redundancy. In an S7-400H system, each subrack houses one CP 443-1 communicating with one bus segment, while each OS Server requires two CP 1623 cards to maintain independent dual connections. The Terminal Bus must never be configured in NetPro, and direct cable links between separate redundant rings are strictly forbidden.

Plant Bus vs Terminal Bus Network Segregation

SIMATIC PCS 7 architectures enforce strict physical or logical segregation between the Plant Bus and the Terminal Bus:

Plant Bus: Transmits real-time deterministic automation communication between AS controllers (S7-400H / CPU 410-5H) and OS Servers / Engineering Stations. Communication utilizes ISO-on-TCP (RFC 1006) or ISO transport protocol over S7 Named Connections. The Plant Bus is fully configured, routed, and parameterized within STEP 7 NetPro.
Terminal Bus: Transmits visualization data, client-server SQL database queries, and operating commands between OS Servers, OS Clients, Central Archive Servers (CAS / Process Historian), and BATCH servers. The Terminal Bus operates under standard Windows TCP/IP networking (SIMATIC NET PC Station) and is NOT configured in NetPro.

Single Ring (Fault-Tolerant) vs Double Ring (Redundant)

Network availability levels are formally classified in PCS 7 documentation:

Fault-Tolerant Plant Bus (Single Ring): Consists of a single physical ring topology governed by High-Speed Redundancy Protocol (HRP) or Media Redundancy Protocol (MRP). If one fiber or copper link breaks, the Redundancy Manager (RM) opens the ring segment to restore communication within 300 ms (or <30 ms with HRP). However, a secondary break or switch failure partitions the network into isolated islands.
Redundant Fault-Tolerant Plant Bus (Double Ring): Consists of two completely separated, parallel optical rings (Ring 1 and Ring 2). Each automation station and server maintains independent interfaces to both rings. This topology withstands simultaneous cable breaks, switch power outages, and component failures without any loss of process data.

Controller and Server Interface Hardware Distribution

The hardware allocation between controllers and servers follows high-availability design principles:

S7-400H Subracks: In an S7-400H fault-tolerant pair (Subrack 0 and Subrack 1), each subrack contains one CPU and one CP 443-1 communication processor (6GK7443-1EX30-0XE0). Subrack 0 connects to Ring 1, while Subrack 1 connects to Ring 2. Because the two half-CPUs synchronize continuously via fiber-optic sync modules, the AS station achieves full redundancy across the two rings without needing four physical CP cards.
OS Server Stations: Each redundant OS Server requires two industrial communications processors, such as CP 1623 (6GK1162-3AA00) or CP 1613 A2. Port 1 of Card A connects to Ring 1, while Port 1 of Card B connects to Ring 2. The SIMATIC NET S7-REDCONNECT software package coordinates bumpless channel switchovers.

Managed Switch Requirements and Ring Redundancy Rules

Adhere to the following network engineering rules:

No Unmanaged Switches: Unmanaged switches like SCALANCE X108 (6GK5108-0BA00-2AA3) cannot evaluate redundancy frames or manage ring state. Constructing a ring with X108 switches creates an immediate network broadcast storm.
Managed Switches Mandatory: Use SCALANCE X-200 (6GK5208-0BA00-2AA3) or SCALANCE X-300 managed switches. Exactly one switch in each ring must be designated as the Redundancy Manager (RM), while all other switches in the ring are Redundancy Clients (RC).
Never Bridge Redundant Rings: Connecting an Ethernet cable directly between Ring 1 and Ring 2 is strictly forbidden. Bridging the two rings creates circulating loops that defeat the redundant separation.

Plant Bus Network Architecture Comparison

The table below details topology characteristics and hardware requirements:

Topology ClassificationRing CountSwitch HardwareAS 410 / S7-400H HardwareOS Server HardwareMaximum Protection LevelStandard Bus0 (Linear/Star)SCALANCE X-100 / X-2001 CP 443-1 per AS1 Standard NIC / CP 1623None (Single point of failure)Fault-Tolerant Bus1 Ring (HRP/MRP)SCALANCE X-200/X-300 with RM1 CP per AS (or 2 in H-system)1 CP 1623 per ServerSingle cable cut toleratedRedundant Fault-Tolerant2 Rings (Double Ring)SCALANCE X-200/X-300 (2 RMs)2 CP 443-1 (1 per H-subrack)2 CP 1623 + S7-REDCONNECTMultiple cuts & switch outages tolerated

Verification

Verify redundant plant bus configuration and communication health:

In STEP 7 NetPro, verify that the AS400H station shows two S7 connections under Industrial Ethernet (one per CP 443-1).
Open the Web Based Management (WBM) of each SCALANCE switch. Confirm the ring status is Active and verify the Redundancy Manager reports normal state.
On the OS Server, open Communication Settings; verify both CP 1623 cards show Operational.
In WinCC OS Runtime, launch Channel Diagnosis (WinCCChnDiag.exe). Confirm that Named Connections displays status OK with both redundant paths established.
Disconnect a network cable from Ring 1 during a planned test; confirm that WinCC continues displaying live process values without alarm interruption.

Limitations and Operational Precautions

Review the following network precautions:

S7-REDCONNECT Licensing: Using dual CP 1623 cards for redundant S7 communication requires an active SIMATIC NET S7-REDCONNECT license (6GK1706-0HB00-0AC0) on each OS Server station.
Do Not Route Terminal Bus via NetPro: Configuring the Terminal Bus in NetPro can cause incorrect routing tables and disrupt S7 Named Connections.
Reconfiguration Time Budget: Ensure total ring switch count does not exceed 50 switches to preserve HRP recovery within 300 ms.

Engineer FAQ

Should the Terminal Bus be configured inside STEP 7 NetPro?

No. The Terminal Bus connects OS Clients to OS Servers via standard Windows TCP/IP networking and must not be configured in NetPro. Only the Plant Bus (AS-to-OS communication) is configured in NetPro.

Why is a single network ring not classified as 'redundant fault-tolerant'?

A single ring is only 'fault-tolerant' against a single cable break. If a second failure occurs (such as a switch power failure or second cut), the ring fragments into disconnected segments. True redundancy requires two independent rings.

Can an unmanaged switch like SCALANCE X108 be used in an optical plant bus ring?

No. Unmanaged switches do not support redundancy manager protocols (HRP/MRP). Connecting an unmanaged switch into a ring topology creates a continuous network loop and immediate broadcast storm.