Key takeaways
- Design the network around the traffic it must carry, including legacy voice, serial and alarm signals, before choosing equipment.
- Rings with physically diverse paths and fast protection switching keep SCADA data flowing through a fibre cut.
- Hybrid multiplexers let new Ethernet services and old low-speed interfaces share one platform, avoiding a rip-and-replace.
- Teleprotection and other latency-critical traffic should drive the design, not follow it.
Start with the traffic, not the technology
Remote SCADA networks fail on assumptions more often than on equipment. Before choosing a topology, list everything that has to cross the network and what each service needs: bandwidth, latency, jitter and availability. A typical industrial network carries far more than telemetry, including voice, CCTV, alarms from dry contacts, engineering access and, in utilities, protection signalling.
A metro railway project we delivered shows the range. One STM-16 SDH backbone had to carry SCADA telemetry, security camera feeds, passenger information displays, legacy telephony and dry-contact alarms from every station back to headquarters, on a single converged platform, with one management view for the operator.
Build resilience into the topology
A point-to-point link is a single point of failure. Rings give every node two ways to reach the control room, so one fibre cut or equipment fault does not isolate a site. In SDH networks, two protection mechanisms are common. Subnetwork connection protection (SNCP) protects a path by sending traffic both ways round the ring and letting the receiving end pick the healthy copy. Multiplex section protection (MSP 1+1) duplicates the line signal on a protection fibre and switches at the receiver if the working line fails.
Protection switching in SDH is normally designed to restore service within about 50 milliseconds (ITU-T G.841), fast enough that SCADA polling and most voice calls ride through without noticing. Protection only helps if the two paths do not share a fate, though. A ring laid in one duct is one dig-in away from failure, so physical route diversity, with separate cable paths, entry points and power, is part of the design rather than an upgrade.
On an oil field network in the Middle East, we used this approach to connect dispersed fields across desert terrain: SNCP ring connectivity and MSP 1+1 line protection on an STM-4 backbone between the two largest sites, with multiple ring protection at the satellite sites. A fibre or line fault does not interrupt the data operators rely on.
Bridge legacy and modern interfaces instead of replacing them
Remote sites rarely arrive on a clean sheet. Older fields and substations still run low-speed voice and data interfaces: DS0, E1, FXS, FXO, E&M and serial links to RTUs. Replacing all of them to reach Ethernet is costly and risky, and often unnecessary. Hybrid multiplexers terminate legacy interfaces and modern Ethernet on the same platform, so each site connects the way it needs to. If you are deciding between SDH and a packet network, see SCADA over MPLS, SDH or radio.
The same oil field network did exactly that. Large, high-investment fields got high-speed access, while older sites kept DS0 interfaces for legacy voice and telemetry equipment, all on one platform that also standardised monitoring and maintenance. Plan for headroom too. The backbone equipment supports an upgrade path to STM-16 without hardware replacement, so growth does not mean a new network.
Teleprotection: the most demanding traffic on the network
Line current differential protection (87L) compares the current at each end of a line over a communications channel and trips if the two do not add up. That makes the channel part of the protection scheme. Delay must be low and stable, and the send and receive paths should match. A fault in the channel can cause a false trip or a failure to operate, which is why this traffic should set the design requirements for the whole network.
IEEE C37.94 defines the optical interface between protection relays or teleprotection equipment and digital multiplexers, carrying N × 64 kbit/s signals. Where equipment is not standard, the interface itself becomes part of the project.
For an East Asian power utility, 87L relays could not communicate because too little fibre was available. Rather than build new fibre, we specified a hybrid multiplexer with a custom C37.94 optical interface card that connected straight to the relays and carried protection messages over the existing E1 and SDH network. Planning, design, procurement and commissioning were all completed remotely, and the utility restored its protection scheme without installing new cable.
When there is no fibre: radio and microwave
Pipelines and remote assets often run where no telecommunications backbone exists. At Karbala Refinery, no communications infrastructure covered the pipeline corridors, so we built the network with UHF/VHF radio links and microwave backhaul, giving every measurement point a low-latency path to the control room. Radio is also a sensible second path for a fibre route, because it fails in different ways. Choose the technology from distance, line of sight, bandwidth and availability targets, and allow for fade margin and licensing early.
Management and security
A resilient network still has to be run by people. Prefer a management system sized to the network. On the metro project, the operator could configure and monitor every node from one location without a large, complex network management system. Treat security as part of the design too: segment networks into zones and conduits as described in ISA/IEC 62443, keep protection and control traffic apart from general traffic, and restrict remote access.
A checklist before you specify
Before issuing a specification, check that you can answer these:
- Have you listed every service, legacy interfaces included, with its latency and availability target?
- Does every site have two diverse paths to the control room, and is the diversity physical rather than only logical?
- How quickly does protection switching restore service, and have your SCADA polling and teleprotection design been checked against it?
- Can legacy interfaces terminate on the same platform, and is there upgrade headroom?
- What is the plan where no fibre exists, and where a fibre route needs a second path?
- Who will manage the network, and with what tools?
- How will it be commissioned and tested, including remotely if the team cannot be on site?
If you are planning or upgrading a network for remote operations, our team designs, integrates and supports mission-critical communication systems from fibre route planning through commissioning. See more in our delivered projects, or contact us to discuss yours.
Frequently asked questions
Is SDH still used for SCADA and utility networks?
Yes. SDH remains widely deployed in utilities, rail and oil and gas because it offers deterministic latency, mature protection switching and long-lived equipment, and it carries legacy TDM interfaces natively. Packet technologies such as IP/MPLS are also common, and many networks run both. The right choice depends on your traffic, your existing assets and the skills you have to run it.
What is the difference between SNCP and MSP 1+1?
SNCP protects an individual path: traffic is sent both ways round a ring and the receiver selects the healthy copy. MSP 1+1 protects a whole line between two nodes by duplicating the signal on a working and a protection fibre and switching at the receiver on failure.
What is IEEE C37.94?
IEEE C37.94 is the standard for an optical fibre interface that connects teleprotection equipment to digital multiplexers, carrying N × 64 kbit/s signals.
How fast does SDH protection switching need to be?
SDH protection architectures are typically designed to restore traffic within about 50 milliseconds (ITU-T G.841). What your application can tolerate depends on its own timing, so protection schemes and polling intervals should be checked against the network’s switching and delay figures.
