This has bugged me for a very long time… (first world problems, I know)… what do you actually call those networking cables that connect a router to a computer/modem/switch/etc.?
CAT5/6 cable? (But those are specific categories of this type of cable)
CAT cable? “Category cable” just sounds strange
Patch cable? But they’re not only used for patching devices together, and they’re not all twisted pair
Ethernet cable? But Ethernet is just a protocol, and can use coax or fiber too.
Networking cable? LAN cable? A bit generic, since FireWire, phone lines, NVLink, etc. also exist.
RJ45 cable? Well, that’s the connector at the end, and they’re apparently more properly called 8P8C anyway…? If you buy them in bulk spools, it’s up to you put the right connectors on their ends.
It’s so odd that one of the most common cables in the world has no easy generic name, just a bunch of almost-precise-enough semi-synonyms. But I suppose the same could be said of most other cable types too, like a “USB-C cable” (which is really like twenty different standards that happen to look the same but are rarely interoperable), a “power cable” (which can have many different gauges, shielding, and connectors)… maybe we’re just bad at naming cables in general?
Another vote for “Ethernet cable”. I know there are technically different kinds, but if anyone says “ethernet cable”, everyone else will know that that’s what’s meant, and so, by the standards of actual language, that’s that thing’s name.
I’ve always called them ethernet cables. For home computer/router use don’t they default to RJ-45? I’ve never called them an RJ-45 because I don’t have other cables like RJ-25s or RJ-62s laying around. I haven’t had a landline in years either so no other connector types.
“Ethernet cable” works just fine today where these things are ubiquitous and there’s no ambiguity about what it refers to, namely one of 7 categories of Ethernet connector cables with RJ45 connectors. In the old days it might not have been clear whether one was referring to the thick coaxial backbone cable, to a transceiver cable (formally, an Attachment Unit Interface (AUI) cable), or to 10base2 thin coax. Today “Ethernet cable” pretty much means only one thing.
I did/do this as well, or 10base T Ethernet cable. This was primarily to differentiate between it and 10base 2, which was common back when I actually spent any time creating cables.
Ethernet cable. Even in the presence of Ethernet-carrying cables that don’t have an 8P8C jack.
For instance, the computer I’m on has two cables that carry Ethernet. One of them has an Ethernet cable plugged in. The other has OM3 optical fiber carrying 10GBASE-SR.
With overhead, 1Gb tops out at about 100 MB/s. But even a very basic RAID setup like on my file server can hit 500+ MB/s on spinning discs, and if I used SSDs instead they’d easily saturate even 40Gb Ethernet, let alone 10Gb.
The whole setup cost like $250 (I ran the fiber myself), so it didn’t exactly break the bank.
Never did figure out where these dips came from, but it kinda doesn’t matter at these speeds:
Ah, if you’re running a high-performance file server, that makes sense.
By way of historical perspective, although Ethernet as originally introduced to the mass market had a theoretical performance of 10 Mbits/s and was typically much less due to CSMA/CD collisions, the original Ethernet first operated at Xerox PARC was actually 2.94 Mbits/s. How far we’ve come!
One of the reasons the 10Gb setup was so cheap is that I used components (10Gb SFP transceivers) that were basically obsolete junk from datacenters. Even 5 years ago, 40Gb was borderline obsolete, and 10Gb was almost given away for free (I think it was $6 per transceiver)!
Now you’ve got me reminiscing, and I should add that my comment about CSMA/CD collisions wasn’t meant as a criticism, but it was an upper limit to performance in most situations. I was deeply involved in these controversies when I worked for DEC and we constantly faced misleading hype from IBM, sometimes even disguised as “scientific papers”, touting the supposed superiority of token ring. The reality is that when IBM recognized the value of LAN technology, they got into it, but with the strict stipulation that the technology must be “anything but Ethernet”, because damned if they, the mighty IBM, would deign to lower themselves to join the DEC-Intel-Xerox Ethernet alliance. So they got token ring. And today we know which one survived, and which one didn’t.
Not that DEC was completely blameless. They made some missteps. Being highly driven by the spirit of engineering excellence, DEC initially rejected the idea of Ethernet over unshielded twisted-pair (UTP – aka, telephone wiring) and heavily promoted 10base2 as the “modern” Ethernet. It was only at the insistence of a large European customer that DEC developed 10baseT adapters at all.
Another mistake they made was insisting that any Ethernet architecture that was faster than 10 Mbits/s was so fundamentally different from Ethernet that it couldn’t be called “Ethernet” at all. Instead, they promoted their commitment to FDDI. But, although FDDI is still used in some telco infrastructure, as a generalized LAN technology it went nowhere because even then, 100 Mbits/s Ethernet was just around the corner.
