Cat5 / 5e / 6 Wiring Idea

[Longhair]

In today's world of going wireless, it seems that the wired section (people who want fast, reliable, secure speeds) have been left on the wayside when it comes to new ideas and technology. Hopefully someone will pick up this idea and run with it to the point where I get a small, obscure footnote in history

The idea: Make a T-switch that works the same way as a 5, 8, ... 24 port switch. Remove the need for an external power supply, all the LEDs that tell you it is connected, etc. and have it be the same size as a regular wall jack that is currently out now.

What this will do is remove the need to have an indiviual cable per Cat 6 wall jack. Currently, it looks like this:

Code: [Select]


|-----------------| ---------------------------- Wall Jack 1
|- Modem / Router-| --------------------------------------------------- Wall Jack 2
|-----------------| --------------------------------------------------------------------------------- Wall Jack 3


The ideal solution would be for it to look like this:

Code: [Select]


|-----------------|
|- Modem / Router-| ------------------- Wall Jack 1 ----------------------- Wall Jack 2 ---------------- Wall Jack 3
|-----------------|



[Psi]

it introduces problems though
Such as;

- lots of traffic collisions as computers speak at once and have to wait for the line to be clear.
- Slowdown when node A wants to talk to node B at full network speed while at the same time node C wants to talk to node D at full speed also
- Breaks in the chain cause everyone to lose their connection.

Old coax 10base2 networks were like that, a big daisy chain of coax.

What advantages do you see a system like this having?

[alm]

I assume each wall outlet / switch would implement a real 802.3 switch, so collisions should not be an issue. I see this as mainly useful for SOHO environments, were 90% of the traffic is between a workstation and the router anyway, so bandwidth should not be an issue. Most SOHO environments don't even have wall outlets, just some loose cables.

For larger environments, the bandwidth will be an issue. Connections between switches often use faster link speed (eg. GE if FE downstream, or aggregated GE/10GE if GE downstream), you would need to run this to each wall outlet. The backplane of most switches is a lot faster than the link speed. Management is also mandatory in those environments, which would probably make costs and size prohibitive.

One issue I see is power. Implementing PoE at the router seems like the ideal solution, but is relatively expensive (proper 802.3at, not the proprietary 'lets put some DC voltage on a random pair' that some manufacturers implement). Needing a wall wart for each wall outlet would suck.

[Psi]

ah, i missed the mini switch per wall jack bit.
that makes more sense

[mitpatterson]

So...Correct me if i'm wrong, basically what you want to do is put a small SOHO switch in each outlet, just strip the LED's? cause to my knowledge you can't really can't run it without some sort of power.

P.S. If you have any more specific standards questions or anything, i have taken a 2 year Cisco(Routers and Switches) course, as well as I'm an A+ Certified Depot Technician.

[PetrosA]

From an installer point of view, I wouldn't recommend a system like this to any customer. With the current designs, if one cable fails or gets damaged, only one endpoint is affected and with CAT 5e you have back-up pairs to use. In a daisy chain system any damage will affect the rest of the chain and could possibly take out the entire network. That's a reliability nightmare. It certainly wouldn't sell in large office environments where there can be hundreds of cables networked. In smaller SOHO environments, I'm not sure what your selling point would be. The cost of the wall jacks would probably be equivalent or higher than the extra cable needed to run home runs to each computer/printer etc. I can't comment on a CAT 6 environment - I haven't seen one installed anywhere.

[Longhair]

My house just had some remodelling done and we added some RJ45 wall jacks to clean up all the cables laying around. The amount of cable going from the router to every jack is a lot even for a SOHO. New houses being built today are having the RJ45 (Cat 5/5e/6) wall jacks placed in just like RJ11 (telephone) wall jacks and power outlets.  So this idea would be better implemented in new homes and businesses.

Current RJ45 wall jacks are completely passive so this is a redesign of it to make it active. For the sake of typing, let's call this the "T-Switch Wall Outlet" or "TSWO".

Power over Ethernet will be the power source so no need for wall warts.

For trouble shooting, there will be a multi-color LED on the front. Not in use (no plug): Yellow. In use and working properly (plugged in): Green. Problem with the TSWO (something failed): Red. Problem with the cable: No light. Also, if a cable does fail taking down everything after it - say in between TSWO 4 and 5 - a pair of adapters (which also needs to be designed) would could be plugged in at TSWO 4 & 5 and act as a Y-Splitter allowing for a patch cable to be put in place until the faulty cable is replaced.

To keep problems with businesses and high traffic down to a minimum, let's put a maximum number that can be used together - say 24 (a number I pulled out of the air). TSWO 1 would be plugging into the #1 port of the router / switch and TSWO 2 upto 24 would be daisy chained. TSWO 25 would be plugged into the #2 port of the router / switch and TSWO 26 upto 48 would be daisy chained... Now if there were 240 workstations on a single floor, instead of having 10 24-port switches, 10 things plugged into the wall outlet and miles (or kilometers) of wiring, there would only be 1 x-port switch, 1 thing plugged into the power outlet, and not so many miles (or kilometers) of wiring.

Add in some firmware like current routers, managed switches, firewalls, etc. have to give the administrator (or parent) control of the individual TSWOs - network only, blocked websites, etc. If a company really wanted to, they could sell "add-on" applications - think iPhone - to get the most out of each market or individual switch.

If you were building a new house or business and this product was on the market, wouldn't you want it to be standard equipment in the new house or business?

[mitpatterson]

My house just had some remodelling done and we added some RJ45 wall jacks to clean up all the cables laying around. The amount of cable going from the router to every jack is a lot even for a SOHO. New houses being built today are having the RJ45 (Cat 5/5e/6) wall jacks placed in just like RJ11 (telephone) wall jacks and power outlets.  So this idea would be better implemented in new homes and businesses.

Current RJ45 wall jacks are completely passive so this is a redesign of it to make it active. For the sake of typing, let's call this the "T-Switch Wall Outlet" or "TSWO".

Power over Ethernet will be the power source so no need for wall warts.

For trouble shooting, there will be a multi-color LED on the front. Not in use (no plug): Yellow. In use and working properly (plugged in): Green. Problem with the TSWO (something failed): Red. Problem with the cable: No light. Also, if a cable does fail taking down everything after it - say in between TSWO 4 and 5 - a pair of adapters (which also needs to be designed) would could be plugged in at TSWO 4 & 5 and act as a Y-Splitter allowing for a patch cable to be put in place until the faulty cable is replaced.

To keep problems with businesses and high traffic down to a minimum, let's put a maximum number that can be used together - say 24 (a number I pulled out of the air). TSWO 1 would be plugging into the #1 port of the router / switch and TSWO 2 upto 24 would be daisy chained. TSWO 25 would be plugged into the #2 port of the router / switch and TSWO 26 upto 48 would be daisy chained... Now if there were 240 workstations on a single floor, instead of having 10 24-port switches, 10 things plugged into the wall outlet and miles (or kilometers) of wiring, there would only be 1 x-port switch, 1 thing plugged into the power outlet, and not so many miles (or kilometers) of wiring.

Add in some firmware like current routers, managed switches, firewalls, etc. have to give the administrator (or parent) control of the individual TSWOs - network only, blocked websites, etc. If a company really wanted to, they could sell "add-on" applications - think iPhone - to get the most out of each market or individual switch.

If you were building a new house or business and this product was on the market, wouldn't you want it to be standard equipment in the new house or business?

Thats a great idea in theory, but the problem is, you have, in your example, 23 workstations on a single switch port, and most switch ports are either 100Mbps or 1Gbps(10Mbps if you have old equiptment) so lets assume the 1Gbps, 1000 Mbps divided by 24, your down to at MAX 41.66Mbps per workstations, now for just intner thats fine, but if you want to do any data transfer that is not very good at all. heck, wifi(N) can be faster than that(assuming only a few devices, i won't get into shared media with wifi). And i still don't understand what is going to be doing the "smart" stuff in each jack, If cabling is that big of concern here is what you do:

Its called a hub and spoke technology(or "Star" sometimes), you put your main router/switch in the middle of the building(or close) then you run a line out to each corner lets say(or each department or what have you) there you place another switch, and then connect all your workstations in that area. here is an example, but replace the hub with a switch(hubs suck) and the workstateions with another switch, and then basiically at each switch, you would have the same thing again:


Here is a image showing the differnt topology's, one of the ways is the extended star, Now for really big netowrks you use the "hierachical" desing, you put your "core" layer stuff at the top, the things that can switch a LOT of data, then you have you "distribution layer" which has servers(that the whole netowork needs) you sometimes also apply your ACL(access contol list) here, then you have your "access" layer on the bottom, these are where your smaller less powerful switches go, and then not shown are the workstations connecting to those


In summary, nice idea, but hard to make very practical

Also, in regards to your last line, it depends on if your technology was "standard" or not, if it wasn't then NO, i would use what is standarized, and the odds of this being standarized are slim, the IEEE is who would normally standardize this stuff, and they are slow, look how long "N' was a "draft" standard, by time it got though them we would be on like wireless "AEJGHSJ"(a random set of letters showing many revisions)

[Uncle Vernon]

My house just had some remodelling done and we added some RJ45 wall jacks to clean up all the cables laying around. The amount of cable going from the router to every jack is a lot even for a SOHO.

Cat5E is around $80 for 305m, Cat6 not a lot more, most of the cost is in the fittings and hardware, your design would massively reduce the functionality with very little cost benefit.

New houses being built today are having the RJ45 (Cat 5/5e/6) wall jacks placed in just like RJ11 (telephone) wall jacks and power outlets.  So this idea would be better implemented in new homes and businesses.

older RJ11 cabling will generally be parallel connected and suitable only for PSTN use.

If you were building a new house or business and this product was on the market, wouldn't you want it to be standard equipment in the new house or business?

Frankly no. The current push is for structured cabling the RJ45run are available to be configured for use in a variety of functions, not just Ethernet data. Structured cabling can carry A/V, security, access control, data, PSTN phone, in house VOIP,  home automation etc. Your solution while not without merit would restrict cable use to Ethernet data alone.

[Chasm]

That Ethernet cabling is also used for other purposes than Ethernet has been mentioned.
After a remodeling we use it for Telco stuff (DSL, ISDN), RS232 (Remote control for a ventilation system), 1-Wire (some "because I can" remote weather sensing), the door bell runs on it, and also Ethernet.
It is easier to push at least one Cat-7 twin cable into every room from a central location than to use different stuff. Less confusion for the guy installing it, who is btw. more expensive than the cable. Why Cat-7 you may ask, simply to have some reserves. - I don't want to replace that cable anytime soon.

It is a bit of a tangent, but as a rule of thumb the bandwidth that Telcos (have to) deliver to their customers increases by a factor of 1000 every 10 years.

Ethernet does not grow as fast, but still you can do a lot of stuff with many legacy 10BASE-T installations to increase the bandwith without replacing all the cables. - Basically unless it is was build on existing phone infrastructure in the first place like it happend in large parts of the US.
"Simple" stuff: Exchange the 10MBit Hub for a 100MBit Switch. Install a Switch with Gbit uplink and replace only that link.

Cable is usually cheap -again the guy installing it is more expensive- unless you have so much of that you get problems to route it through your cable ducts. And there the answer is usually to push relatively stupid  "aggregation" switches deeper into the field.

[RayJones]

Standards exist for a reason, and twisting standards about usually ends up biting you on the backside.

10base-2 cabling was a nightmare to manage and extend.
Whenever you needed to add points, the entire network segment was brought down.

We even had the issue of somebody mistaking the 10base-2 coax cable for a signal cable within a radar. The radar functioned perfectly well, but it took a couple of hours to find out why the network was down though.

Serialising cable CAT5/6 cable runs returns you back to the mess of 10base-2.
I would not go back there.

[Longhair]

Let's take the visors off for a moment...

What if we don't use CAT 5/5e/6/7/847132 cable but something different? For example - fibre optic cable for example.
What if we don't worry about the cost? The goal is to make the product / system popular enough so it drives the cost down to where everybody could afford it. Remember the days you couldn't by any PC for under $2500 and today they are giving them away?
What if we don't care about standards and make our own? Music theory was build around music, not the other way around like today.

[PeterG]

If it aint broken, why fix it?

[Uncle Vernon]

Let's take the visors off for a moment...

What if we don't use CAT 5/5e/6/7/847132 cable but something different? For example - fibre optic cable for example.

same result, higher cost.

What if we don't worry about the cost?

That's what economists said pre-GFC. Unless you've found a money tree or a sugar momma cost is always going to be a factor.

The goal is to make the product / system popular enough so it drives the cost down to where everybody could afford it. Remember the days you couldn't by any PC for under $2500 and today they are giving them away?
What if we don't care about standards and make our own? Music theory was build around music, not the other way around like today.

I can see what your getting at, but honestly the problem isn't the cable. The structured cable approach offers a much more flexible solution. Running all your cables back to a central location/s offers the greatest degree of future proofing.

[Longhair]

If it aint broken, why fix it?

[PetrosA]

If it aint broken, why fix it?

You can't compare a servo control system to a computer network...

[Longhair]

If it aint broken, why fix it?

You can't compare a servo control system to a computer network...

Why not?

• They improved something that has been around for a very long time to work better even though there was nothing wrong with it.
• There is different technology being implemented to achieve the same thing in the end.
• The need for multiple wires going from the receiver to the servo has been lowered by a lot with an added side effect of dramatically shortening the trouble shooting process.

So why is it so wrong to apply the same logic to a computer network?

[baljemmett]

• The need for multiple wires going from the receiver to the servo has been lowered by a lot with an added side effect of dramatically shortening the trouble shooting process.

So why is it so wrong to apply the same logic to a computer network?

I take it you've never the dubious pleasure of running a 10base2 / 10base5 Ethernet network, or any similar topology?  Moving to star topologies and the ilk also dramatically shortened the trouble shooting process, with an added side effect of not rendering entire segments of the network totally dead whenever a user decides to practice office feng shui...

[RayJones]

Having flown RC models in the past this S-bus system concerns me greatly.
If you lose a connection close to the Rx, you have definitely lost your aircraft.

Also I would suspect that the classic PWM stream (most likely in digital form) is simply being relayed on beyond the Rx instead of decoding it at the Rx.

Yes it is an innovation, but reliability is paramount, especially in today's litigation crazy world.

As stated already, a star system is used to minimise disruption to other clients (servos) when connections fail.
I would have thought this safety aspect alone would outstrip the apparent GAS presented in that video.

[PeterG]

There are 3 main issues with the system being proposed here.

The first problem i see is length of cable. I don't see it working well over 50m+ lengths. I am sure it works well over the 40cm max it is designed for.

The second is the issue of noise, being a purely serial system, it does not appear to be very robust.

The third issue is with overall throughput. How is this system going to perform with 20+ devices all wanting to talk to each other at Gbit speeds? Current star configurations allow 2 devices to communicate at full bandwidth without a reduction in bandwidth being seen by other devices.

I would have a look here for more info about current 802x technology. http://www.ieee802.org/

The use of Cat5/6 allows for a robust signal that can run over long lengths (100m+ in some cases).

My 2 Cents.

[williefleete]

if what i am seeing is right you want to use one length of wire with active nodes at each computer, this is looking rather similar to old thicknet which has a trunkline and you core out a section and hook in a module that connects to a DB15 connecter or thinnet with coax, T connectors and 50 ohm terminators at each end, and is an arse to diagnose if a section of cable goes OC
do we really want to go BACKWARDS with network technology

[Longhair]

Having flown RC models in the past this S-bus system concerns me greatly.
If you lose a connection close to the Rx, you have definitely lost your aircraft.

Isn't that also the case if your Tx is unable to communicate with the Rx (brown out or someone else on the same frequency) without the S-Bus system?

[Longhair]

Why is everybody so closed minded and is only able to say "it won't work" or "it will fail because..." ?

I thought that engineers were suppose to push forward through existing limitations to come up with new and exciting solutions.

I do not have the skill set to implement my idea but this is how I would go about it.

• Step One: Educate myself at the task at hand (learning standards & limitations).
• Step Two: Make a list of what problems may arise and needs to be addressed.
• Step Three: Repeat steps one and two if necessary.
• Step Four: Try.
• Step Five: Fail.
• Step Six: Try again until I pass the current fail point.
• Step Seven: Move forward until next failing point and repeat step six.

My whole outlook at life is "I'd rather fail trying than fail not trying" when it comes to doing something new and interesting.

[PeterG]

Nobody is saying your idea is wrong, we are saying it was done a while ago. It had many flaws and has evolved to the modern networking topography.

I would have started by googling 'network topography' to see the pros and cons of each type.

Just my 2 cents.

[PetrosA]

Why is everybody so closed minded and is only able to say "it won't work" or "it will fail because..." ?

I thought that engineers were suppose to push forward through existing limitations to come up with new and exciting solutions.

I do not have the skill set to implement my idea but this is how I would go about it.

• Step One: Educate myself at the task at hand (learning standards & limitations).

This is where you're at right now. You have a whole gob of people who deal with networks telling you that this "standard" has already been tried and the limitations have been discovered and it was an epic fail. A new iteration isn't going to change what caused the failures of the design, so further trials aren't going to help much.

...Unless you really want to be the next Icarus and try super glue or some 3M double sided tape instead of wax