The Land Transport Authority (LTA) has concluded that the mysterious signalling faults on the Circle Line (CCL) were due to faulty signalling hardware on a rogue train. It did take a while to get sorted out. This reminds me of the old days of 10Base2 Ethernet. You merely needed one bad connection on the network to severely impact, or completely disrupt, the entire network.
I didn’t have the opportunity to actually work with 10Base5 networks, although remnants of it were still around when I got into a networking job. But 10Base2 is old enough. Many people now don’t know about 10Base2, or perhaps don’t remember about 10Base2. Basically, it’s an Ethernet network, not dissimilar to the Ethernet we use for our wired computer network today, that’s based on thin coaxial cables. That’s the kind of cables you have on your StarHub cable TV (which will be slowly phased out).
The 10Base2 cabling has to run to all user locations, and a T-junction is needed to tap-off signals from the main cable to feed to a network device. If any of these junctions are loose, or broken, or if the terminating resistor is loose or removed, the entire network goes down.
Even in the just slightly newer era of 10Base-T networks, before bridges and switches became prevalent, the entire Ethernet network is just one fragile system. If any device on the network generates spurious signals, the entire network is disrupted.
Does this sound disturbingly like the CCL signalling problem? A faulty train caused signal interference that affected the line.
The problems with defective Ethernet devices that disrupted a 10Base2 network happens often enough. We were quite proficient in quickly isolating such defects.
The advent of network switches made made Ethernet far more robust, and a bunch of various other technologies pretty much made Ethernet just plug-and-play. I’m sure seasoned professionals understand the underlying fundamental workings of Ethernet, they know MDI and MDI-X, they have grappled with auto-negotiation nastiness, and a whole lot of other curious bits of Ethernet that few people understand nowadays.
I work, in my day job, with people who can be considered extremely computer savvy. Yet, from time to time, I need to remind them that an Ethernet switch needs electricity to work. Heck, even those specialising in network technology need the occasional reminder about basic network knowledge.
Can you believe it that people have called me, on an IP phone system, to announce with an extreme sense of urgency that “the entire network is down!”, “not just me, but many other people have also confirmed that nothing is working!”, etc. They called me using an IP phone, to my IP phone, which runs on that same network they declared is down. I just ask them, how is it the network can be down when you’re using it to call me?
To many people, the network is just magic. No, perhaps not actually magic, but it’s this thing that just works, and they don’t know how. They’re just end-users, and how the network works is someone else’s problem.
I don’t know if this is, perhaps, how our MRT is run. SMRT, SBS Transit, and LTA, maybe they’re just users of systems they bought from someone else. They don’t really know how the system works. The signalling system, for example. It’s supposed to just work. When it doesn’t, everyone’s stumped, and they have no clue.
As perhaps the LTA has realised too late, we are lacking in engineering talent. It’s not just engineers that we require, it’s the expertise that we need. One shouldn’t assume that simply having more engineers onboard will satisfactorily address the need.