"Sorry, the ice cream machine is broken... Again."

Anyone who’s set foot in a McDonald’s has probably seen the dreaded “out of order” sign taped to the ice cream machine and let out a wistful sigh. Whether you’re chasing a late-night dessert or trying to keep the kids happy, the seemingly always unavailable machine has become both a running joke and an endless source of frustration around the world.

It’s not just a joke, though. At any given time, an estimated 12–15% of McDonald’s ice cream machines in the US are out of service (mcbroken.com). In fact, they fail so often that a real-time tracker was built specifically to monitor it, letting customers check whether their local machine is working before they visit the store. In some US cities, the failure rate is as high as 30%.

But behind the memes and frustration lies an engineering problem that is surprisingly familiar to asset managers, rail operators, infrastructure owners and project teams.

It’s Not Just About Reliability

When a machine develops a reputation for being broken more often than not, most people assume it’s unreliable, poorly made or that it’s just badly designed. In reality, the problem is often much broader.

The infamous ice cream machine suffers from several challenges that Reliability, Availability and Maintainability (RAM) specialists see every day:

• Complex equipment design
• High utilisation with little downtime
• Lengthy cleaning and recovery processes
• Restricted maintenance arrangements
• Difficult fault diagnosis
• Slow return to service after an issue.

Any one of these factors can reduce availability, even if the equipment itself is technically reliable.

So, what is RAM? In short, reliability is whether an asset performs its intended function without failure. Availability is whether it’s actually there and working when you need it. Maintainability is how quickly and efficiently the asset can be inspected, repaired and returned to service when something does go wrong. RAM engineering is the discipline of balancing reliability, availability and maintainability to ensure assets are ready, working and capable of being restored quickly when things go wrong.

The Same Problem Exists in Rail

Consider a signalling failure during peak hour – maybe a failed relay in a control room. The fault itself may be relatively minor, but diagnosing the issue takes time. Once identified, specialist technicians may be required, replacement parts may not be immediately available, and recovery procedures can be complex. Meanwhile, trains are delayed, platforms become crowded, and passengers are left wondering why services are not running.

In moments like this, reliability isn’t really the bottleneck. It’s an availability problem driven by maintainability, recovery capability and operational constraints.

It’s not hypothetical, either. A 2025 review of Sydney’s rail network found peak-hour delays occurring at their highest rate since 2016, with the number of high-priority defects (assets with damage or flaws requiring maintenance) reaching an all-time record of 4,714 in June 2025, up from just over 3,800 two years earlier (Transport for NSW). A 2023 review had already found delays at their worst level in a decade, with over half of all disruptions traced to deteriorating infrastructure such as tracks, signalling, structures and electrical systems (Transport for NSW). The pattern is consistent: it’s rarely a single asset failing outright. It’s a growing backlog of faults taking longer to resolve than the network can absorb.

This is exactly the gap RAM engineering exists to close.

Why RAM Engineering Matters

Good RAM engineering looks beyond whether something works. It considers how the asset will perform throughout its operational life and how quickly it can return to service when things go wrong.

Questions typically include:

• How often will failures occur?
• How quickly can faults be identified?
• How easy is the asset to maintain?
• How long will it take to recover from failure?
• Can maintenance be completed without disrupting operations?
• Is the design introducing unnecessary complexity?

The answers to these questions influence safety, performance, customer experience and whole-of-life costs.

The Bigger Lesson for Rail Infrastructure Reliability

The next time you see an “out of service” sign on an ice cream machine (or a train!), think beyond the obvious. The issue may not be a simple equipment failure. It could be a combination of reliability, availability and maintainability challenges that have been designed into the system from the start. In other words, was the system ever designed to return to service quickly when it did break down?

Whether you’re serving soft serve or moving thousands of passengers every day, the principles remain the same: simpler systems, maintainable designs and faster recovery times lead to better outcomes.

Reliable and available aren’t the same thing and most engineering failures live in that gap.

Where’s the gap hiding in the systems you’re responsible for?

Staff Profile – Dinesh Luthra | Consultant Safety Assurance/RAM 

Related Content: 

Expertise – Systems & Safety Assurance/RAMS