Rail telecoms resilience: why lifecycle costs trump initial bids

Rail operators often discover too late that the cheapest bid at tender stage is not the cheapest network over its lifetime. In rail telecoms, where one platform or signalling failure can affect safety, punctuality and passenger confidence, the real test is not the sticker price of the hardware but how the system performs over years of service.

Nokia’s Benoît Leridon argues that the telecoms layer acts as the railway’s central nervous system, carrying signalling, train cont...

rol, CCTV and passenger information. Industry practice commonly treats 99.999% availability as the benchmark, a level that leaves only minutes of downtime each year. That standard is difficult to sustain if procurement decisions prioritise initial capital expenditure over resilience, maintainability and lifecycle cost.

The strongest networks are designed so that a single fault does not interrupt services. That means redundancy not only in the obvious places, but across the full architecture: hardware, fibre routes, switching and power. In rail signalling, one familiar approach is the blue-and-red model, in which parallel devices and separate paths allow traffic to keep flowing if one side fails. Rail engineering specialists have long stressed the same principle: critical control and communications systems need diversity, hot standby capacity and redundant power supplies if they are to stay online through faults and maintenance.

Operations matter just as much as design. A network can be technically robust and still be expensive to run if faults take too long to repair. Regional spares, local maintenance capability and the speed at which a technician can reach the right site all shape mean time to repair. The difference between a short intervention and a prolonged outage can erase any saving made by choosing cheaper equipment in the first place.

That is especially true when equipment has been designed for office IT rooms rather than railway environments. If a failed module requires a chassis to be unracked, cabling stripped out and the system rebuilt, what should be a simple swap becomes a service-affecting intervention. By contrast, units built specifically for rail use allow failed components to be replaced with far less disruption. Leridon’s point is that redundancy should be treated as a core procurement criterion, not a premium extra.

Planned maintenance brings its own penalty. Metro systems often have only a narrow overnight engineering window, and every change must be reversible if something goes wrong. Mainline operators face different but equally expensive constraints, because work on signalling networks can require a validation run before passenger services can resume. In both cases, the cost of intervention is driven not just by parts and labour, but by access to the network and the need to prove that it is safe to reopen.

The challenge has grown as railways have brought together two very different classes of technology. Safety-critical applications such as train control and signalling usually have long lifecycles and strict assurance requirements. Non-safety systems, including CCTV, passenger displays, IP telephony and trackside sensors, are more akin to commercial IT: they need regular updates, are more exposed to cyber risk and often evolve faster as networks expand. According to recent academic work on rail maintenance and system design, network topology, redundancy and service impacts are increasingly being modelled together rather than treated as separate questions, precisely because one change can ripple across the whole system.

That is one reason physical segregation is attracting more attention. Running separate networks for safety-critical and non-safety applications sounds expensive, but the overlap between the two is often smaller than operators assume. Leridon suggests that the shared infrastructure can account for only about a fifth of the total, making the added capital cost easier to justify when set against fewer maintenance windows and lower operational disruption. Swiss Federal Railways is frequently cited as a strong example: it has kept dedicated networks for critical and non-critical applications in place for more than a decade, with redundancy embedded throughout. Its reputation for punctuality is often linked to that long-term engineering discipline.

The broader lesson is that rail procurement should be judged through total cost of ownership, not unit price alone. Consultants and researchers in the sector have repeatedly reached the same conclusion: lifecycle cost, reliability and service performance need to be balanced together if operators are to avoid hidden expense later. A modest saving on a router or switch can quickly be overtaken by one emergency callout, a missed maintenance window or a failed software upgrade.

For operators, the practical questions are straightforward: how much redundancy is built in, how often will software need to be updated, how long will a repair take, and what does every intervention cost once access, testing and service disruption are included? The cheapest option on paper is not necessarily the most economical in practice. In rail telecoms, durability, recoverability and maintainability are the features that ultimately determine value.

Source: Noah Wire Services