Troubleshooting network speeds is the ethos of my job description within friend circles and family, but do it for a few years and a few realizations dawn quickly. For one, the blinking lights of a network switch are often a distraction from the real, physical limitations hiding in plain sight. It didn't take me long to prioritize high-quality structured cabling in my desk setup so I could put longer hours in without buffering. However, I've noticed several peers treating cabling as an afterthought to the shiny metal box that is the switch.

We love to blame the switch. It is, after all, the busiest hub of operations and second only to the modem in importance of operation. When speeds drop, we assume the switch is overheating, or a port has gone bad, or perhaps the packet buffer is overflowing. But in my experience, the switch is often just the messenger delivering the ‌news about your cabling infrastructure. You see, I could sip a thick shake through a coffee stirrer, and a switch can have all the 10Gbps uplink capability in the world, but it cannot fix the physics of copper wire adhering to specs from the 90s.

The hierarchy of speed

And why everyone blames the router or switch first

Ethernet cables are typically the fastest way to connect devices on the same network, be it the critical wired backhaul between mesh routers, or a gaming PC client connected to the main Wi-Fi Access Point (AP). While Wi-Fi 6E and the upcoming Wi-Fi 7 promise speeds that rival physical connections, they are subject to the laws of radio interference in ways that a shielded copper cable simply isn't. For raw throughput and low latency, there is nothing as simple as an RJ45 connector and a Cat6 cable I plug in.

However, there are several reasons why you might not be getting the best speeds imaginable, even on Ethernet, and that starts raising eyebrows, too. I might plug in a device expecting a Gigabit handshake, only to be shackled to a 100Mbps uplink, or worse, experience intermittent packet loss that makes online gaming a stuttering nightmare. When this happens, standard debugger instinct tells me to look at the active electronics in the chain and start from there.

Network switches are an important component of a busy network, like a small business or home office, where you need multiple Wi-Fi APs, a fast hardline internet connection to several devices, like a NAS, and a failsafe for critical equipment. In such environments, a switch is a logical addition that effectively expands your network's capacity, allowing devices to communicate directly without clogging up the router's CPU. Since it sits atop the topology, it draws the most suspicion when things go sideways.

Unless you have legacy infrastructure handed down for years, the network won't have a cheap 100Mbps Fast Internet switch that would throttle gigabit Ethernet. That said, I've been in situations where I ruled out ports gone bad due to power surges, configuration mismatch where the switch and client device can't agree on a duplex setting, Quality of Service (QoS) settings deprioritizing traffic, and the switch just running hot and unventilated. There's a lot that can go wrong, but the culprit turned out to be the cables, creating a bigger bottleneck than the switch and other network devices.

There is a nonzero chance your cables are to blame

Don't think about running old cables when moving house

The issue is rarely that the cable is completely borked, and fails a simple continuity test. More commonly, faulty cables work just enough to establish a link, but fail to maintain the signal integrity required for high speeds. I don't handle network troubleshooting professionally, but I've still seen my fair share of aggressive bends, staples driven through jackets, and terminations that look like they were done in the dark. These physical deformities introduce crosstalk and interference, forcing the network equipment to re-transmit packets, which tanks your effective throughput.

Poor speed on wired LAN in corporate settings opens up a whole other can of worms. Well-meaning network admins tend to buy a huge spool of Cat 5 with a 100-pack of RJ45 connectors and a crimper for that custom-length cabling. However, within the decade, a switch to Gigabit Ethernet reveals that Cat 5 was a poor choice, and all that needs to be redone is at least Cat5e spec. Even if you get a Gigabit link light, the cable frequency rating might not support sustained transfer rates without errors.

With the obvious stuff handled, one might still forget critical cables and patch lines on the switch reliant on an older spec, untouched for years, and forgotten because they never caused problems. I’ve spent hours debugging a "slow switch" in a friend's new home, only to realize the uplink cable between the switch and router was a 15-year-old patch cord he carried over from his old place because "wires don't go bad, right?" Well, it looks like an Ethernet cable and clicks like one, but it strangles the entire network's bandwidth.

Wired networks done right and done well are mutually exclusive

In networking, most hardware shackled together works, but not always optimally. That takes tinkering, and it's all too easy to hook cables up and not touch them again until you need to move house or something catastrophic happens. We assume that copper doesn't "go bad" like a battery, but connectors oxidize, retaining clips snap (reducing contact pressure), and cables get crushed behind drywall.

We strongly suggest labeling your cables so identification and replacement are easy, and regularly update them if you have the opportunity when upgrading your router, switch, APs, and other components. It is a small investment that pays huge dividends in terms of reliability down the road. Cabling plays an important role in a fast home network, and it is easy to ignore it and blame the complex components of the home network instead. When troubleshooting, check everything, and that includes cables.