Ask Slashdot: Is It Time To Shrink the Ethernet Connector?
New submitter jimwelch writes: HDMI has shrunk to mini, then micro. USB has shrunk to mini, then micro. The wired Ethernet connector has not changed since 1988! On the Raspberry PI, it is the largest of the standardized connectors. Is it time to come up with a new version? What if, anything, would you like to see replace that suddenly clunky RJ-45 port? I rather like that (in theory) RJ-45 cables can't be easily dislodged, but at the same time dislike that its locking mechanism can be awfully fragile. And for that matter, I'm glad that on most of my computers so far there's been full-sized USB ports as well as full-size ethernet jacks.
Let's rewire miles of data centers for no discernible purpose.
Namely, existing infrastructure. With CAT5 and CAT6 cables everywhere, you will need some little box to convert the existing cable into a slimmer one which in turn would end with a slimmer connector.
There are far, far more RJ45 connectors in the world than USB, for example.
...gis sdrawkcab (usually not responding to ACs; don't bother posting as AC)
it has been replaced in consumer equipment by the (very small) WiFi connector
No. Leave it alone. Devices that are too thin for a standard jack are perfect candidates for a micro-USB ethernet adapter. The default assumption for RJ45 should remain as it is. No need for yet another connector to require we carry five different possible adapters and cables.
"Oh no... he found the
Ethernet is the one reliable standard that will always work, everywhere, no questions asked. And I need it. I can go on for days without eating. I can go on for hours without drinking. Without Ethernet? Good old, reliable, wired, Ethernet? What am I alive for? And don't come with your fancy "Wi-Fi" b/g/n. It never works when you need it. Airport? Conference? eduroam? It does not work! And I need it to work, this is the Internet we're talking about!
entropy happens
The standard rule applies. When a "Should x..." question is asked, the answer is no.
Any reduction would be at the expense of compatibility with everything which already exists. Modular connectors are reliable, cheap, easy to install, they work. Wired Enet is near end of it's capabilities (10G reduces the distance from 100 m to 15), so you'd be better off looking toward smaller fiber connectors as we move forward.
"National Security is the chief cause of national insecurity." - Celine's First Law
Whoever wrote this article obviously didn't do any research first.
There is already a new standard for physical ethernet cabling, calling RJ.5 (that is, ar-jay-point-five): http://www.alliedtelesis.com/videos/RJpointfive
"Allied Telesis is one of the first networking vendors to embrace the new RJ point five Ethernet connectivity standard. Built to replace the RJ-45 standard copper Ethernet connector, the new RJ point five connectors are half the size, so you save valuable space and double your port density."
They're not popular in the marketplace because the cables are uncommon and therefore expensive, and similarly the physical jacks are uncommon and therefore expensive.
One of the reasons, I think, that the RJ-45 connector has lasted so long is it's very easy to field install. A bag of cable ends and a relatively inexpensive crimp tool is all you need, and the wires are easy to insert. Making a connector that's appreciably smaller would make field installation of ends that much more difficult.
Introduce a new standard and now you'll need new cables (wall jacks to device) or adapters (cables to device) to keep new things interchangeable with existing things. That doesn't simplify anything.
That's assuming it CAN be made smaller, given the cable is unlikely to change.
=Smidge=
Sure; let's set up WIFI only for Corporations with 10,000 people on a campus! Great idea!!!
Ethernet is an attempt to use the cheapest cable possible for the longest distance possible for the highest bandwidth possible. That's part why TIA/EIA standards do not specify the cable, they specify the performance characteristics that must be met. This is why it's possible to run Gigbit over some particularly short distances over Category 3, or why it's possible to get 10G out of 6 or 5e for some short distances.
Changing the connector means that the horizontal cable gets more expensive, the jacks get more expensive, the patch cord material and plugs get more expensive.
There already has been interest in changing the connector, larger. There was a cable that put four pins on the top to attempt to electrically separate the pairs to reduce crosstalk. It didn't take off, probably because the developer didn't want to license it cheaply enough, ie, free. There were attempts at hermaphroditic cables, but they were larger and had licensing issues.
The 8P8C jack used as RJ-45 for Ethernet, RJ-48 for T1 and ISDN, and RJ-61 for telephone is not going anywhere.
Do not look into laser with remaining eye.
My opinion: leave it as is
My thinking is;
Most connected items that are portable or IoT already use WiFi now so having a smaller connector wouldn't really be a benefit.
For the larger systems like desktops and servers there would be little to no benefit from the smaller connector.
A standard RJ45/8P8C connector/jack is already about as small as you can get it and still be able to see what your doing when you install them.
Currently the tools and connectors used for CAT X cabling are completely standardized and interchangeable with most of the telephone hardware still out there. Things like the line testers and punch down tools work on both systems so I have less I need to buy and carry when in the field working with mixed systems.
All the older hardware, the Smart TV's, the server patch panels, the home routers, hubs, etc. use the full size connector. I don't think people would be happy if they bought a new router and had to get all new cables to boot.
Just some of my thoughts on the subject, I'm sure there are going to many other valid reasons for and against that other commenters will bring up.
Fiber-based Ethernet has already different kind of connectors.
bash$
What do you mean I need a terminator, the cable just plugged right in?
The beauty of the RJ-45 standard is that it has low insertion force, a positive engagement report (the click when the cable seats properly), and it is essentially impossible to put in the wrong way. It remains in place without screws, and yet releases easily. The only shortcoming it has is the fragility of the catch mechanism when pulling cables through walls or cable trays, but various manufacturers have come up with a range of boot designs to circumvent that problem. You can recognise the connector port by feel, and know the orientation blindly (ie, around back of the equipment you can't get your head behind to be able to see). Other people might disagree, but in my experience, it's the most reliable connector in common use. Maybe the RJ-11 (standard telephone jack) was, in its heyday, more commonly deployed, but probably not. I have never, ever, not once, found a panel-mounted RJ-11 or RJ-45 that had failed.
Compare with the micro USB: insertion force is high enough that it's close to the force required to plastically deform the connector when putting it in the wrong way, yet, it can easily fall out under many circumstances. There is no positive feedback on proper seating. The holes for a micro USB are indistinguishable by feel from many other ports (at least to me). There is no retention mechanism other than friction. The connectors are very fragile, and nearly impossible to join to the cable in the field (read: you can't make your own cables). The insertion count lifetime is quite low, and I've worn out quite a few of them myself. It's a poor standard.
The folks designing the RJ-45 and its sister standards were frelling brilliant. The people designing the more recent stuff ... not so much.
Put my fist through my alarm clock with its ding-dong death inside my ear. - The Blackjacks.
USB is horrible. Every display port since VGA has been horrible Pick a standard and STOP.
These are the current USB:
USB 3.0 type C, USB 3.1 Type C, USB type A, USB 3.0 Type A, USB 2.0 Micro, USB 3.0 MicroB 10-pin, USB 2.0 mini-B 5 pin, USB 2.0 type B, Apple lightning
These are the current video cables:
HDMI, HDMI micro, HDMI mini, DVI dual, DVI-D single, DVI-I dual, DVI-I single, Display Port, Mini-Display Port, VGA (d-sub 15 , PC-RGB),, Thunderbolt, 3-wire component, 5 wire componen video, composite video, s-video, 13w3, 5 BNC RGB.... Cant think of more.
This rubbish has to stop. I have probably 50 cables in various card and places where computers and phones live to deal with this horrible mess.
You know how much waste this creates having to chuck cables all the time or keep piles of old ones around to deal with the fact my phones,DSLRs, camcorders currently have micro, mini and type-c all at the same time.
This is stupid, wasteful and out of control. And now with Type C there is horrible problems with getting rapid charging even from 2A chargers with certified cables (the expensive ones).
Please, for Pete's sake, please dont muck with ethernet or do it exactly one more time at most. Its sickening to think of the billions of miles of CAT5e and CAT6, even some CAT3, that would be obsoleted if the 8P8C plug was EVER changed.
Legalize the constitution. Think for yourself question authority.
Have a look at some of the shenanigans NIC makers had to do with PCMCIA cards: https://en.wikipedia.org/wiki/...
Oh, I'm not saying it wasn't awesome, but it was fragile as hell.
I'm presuming that there's a diminishing law of returns there at some point. But, and this is a question - I simply do not know, is there really any reason we can keep expanding to having multiple twisted pairs all sending/receiving with well-timed off-sets so that it was a bit round-robin-esque and perhaps multiple simultaneous connections? A bit like poor-man's broadband where you multi-linked a couple of dial-up connections...
I imagine it'd be fragile as hell but I don't see why it's not realistic - to a certain point where it becomes impractical. I'm thinking something akin to the trunks or OCs running into buildings and the likes - only more specifically aimed at a closer to the desktop, such as the router, the cable to the NIC even, etc...
I know a bit about networking - the fundamentals. I've done some work with it because I had to but it's vital to point out that I've not touched anything professionally (really) since about 2000. I don't really see a reason why a thumb-width cable full of twisted pairs of copper can't be brought to the unit or even to the end-point. It's more resilient than fiber is, in the physical sense. Hell, with error correction it could just send notice and then function in limited capacity as it routes around failures. At least it can in the picture that I have in my head. ;-) (The picture in my head may not actually fit reality as well as I'm thinking.)
What am I missing? Why is this not done? Why is there no CAT-55, CAT-60, or the likes? Things already support multiple streams. Even if they didn't, splitting stuff into that wouldn't be too damned hard. I imagine that it'd be a bit frail, potentially. I imagine that scaling might be a problem but that's what error correction is for. I imagine that error correction is going to add a bunch of computational overhead and that there's a diminishing return at some point but - are we at that point? Is there anything I'm missing that prevents us from going further?
Why am I fixated on copper? More so, why am I fixated on copper when the rumor mill assures me that my *very* remote area is going to get a fiber service? 'Cause I've seen copper on the ground, bent by trees, and blown completely off multiple poles and buried under ice and *still* had reasonable throughput. They'll be hanging the fiber from the poles and not laying it in the ground - yes, they're going to and yes they already do. I'll take something slower but more reliable. Oh, I'll be jumping on fiber the minute it arrives. I'll also be keeping DSL and figuring out a way to automatically fail-over to DSL when the fiber inevitably goes down.
"So long and thanks for all the fish."
I don't know if I'd say "STOP." That seems less than intelligent to say. I'd say stop doing so for trivial reasons. I'd say stop doing so without good reasons. I'd say stop doing so for proprietary reasons. But, I'd not say stop doing so entirely.
Why? I speculate that improvements will be had in the future. I say that gains will be made and technology will advance even further. I'd rather not prevent that. I'd rather not force them to not make new standards. I'd rather they had the freedom to make newer, better, and different.
I just wish they'd not do so for trivial reasons.
Make sense?
"So long and thanks for all the fish."
All ethernet devices of the last 10+ years can autonegotiate between straight and cross-over wire configuration. So with wires on the connector organized as tx+ tx- rx- rx+ (4-wire example), a flippable connector would not require any changes in the electronics. And it would get rid of the impedance-killing central pair split in today's standard connector.
Of course, the question is whether you could make a flippable connector that's field-crimpable with simple tools.
Avantslash: low-bandwidth mobile slashdot.
I might be one port but things dont work right. I know, I have the entire bevy of these connectors, both USB and Video, and USB speed and behavior has been terrible since 3.0. Speed isnt right, amps isnt right. You can look on Amazon reviews on cables - massive complaints. There is a Google engineer "Benson Leung" that reviews cables as a hobby and checks them to see if they violate USB spec. Even with Benson approved cables the ports often do not charge correctly. Sometimes its 1A even on a 2A port. Sometimes its 2A. Its almost never does the 100 watts / 20 volts / 5A - unless you buy a charger. And type-c chargers are bloody expensive and hard to find. You need a Benson approved one and test it to see if it really rapid charges.
So yes, while type C might be closer to the ideal its a stupid joke how much they cost and how little benefit they provide over a regular stinky 2.0 micro cable these days.
So do I want more unversiality like 8P8C / "RJ45" - hell yes. Id rather wait a bit longer for 10GBASET to become practical. But wait, since there are literally TEN 10GbE standards - that would be 10 different transceivers that can be 10GbE - making fixed port SFP+ and 10GBASET restrictive and SFP+ twinax expensive and SFP+ other types of transceivers very expensive...
You know how much the cables and transceivers cost in a DC? I've seen figures in the 30% range.
You know much more expensive 10GbE because of the number of lunatic standards we had and how much cabling cost?
You see the advantage of CAT5e/CAT6 was it was useful for a VERY, very long time - since 1991. That reduces waste and cost and keeps wiring in the walls good. It was more universal then power adapters, other anything today. It was and is the entire world standard, along now with WiFi - ABGN - I've never had an issue connecting with WiFi or wired ethernet. Ever.
But the diarrhea mess of Apple and phone cables I keep in a backpack at all times. Joke. Custom power cables for laptops. a thunderbolt Ethernet, a DVII, a DVID, a display port, HDMI, mini-displayport, a VGA, type C-C, type A-C, a rapid charger, lightning-cable - all this trash being lugged around. But there is one thing I have one of. Ethernet and the wifi (wherever that is buried.within the laptop). I have a gaggle of dongles and a gaggle of USB cables and chargers because these standards suck horribly.
Legalize the constitution. Think for yourself question authority.
Nobody's talking about changing the wiring or the standard, as far as I can see. What people are discussing is a new standard for connectors at the end of the cable. This would not obsolete any existing equipment, as the RJ45 connector is perfect in data centre installations etc. The new standard would be targeted at portable end-user devices, things that the RJ45 connector was never designed to cope with anyway.
Got them moderator blues I blieve I walk out the do', With these mod-points I been gettin', I 'most never post no mo'