What are data multiplexers?

176 views

For instance, I work for the government and in their IT closet I see these big racks that are carrying multiple fiber lines which terminate into this box that says:

Nokia 1665 DMX which is apparently something called a “data multiplexer.”

I did some research but I still don’t quite understand what they do. Are they for combining data, tv, and phone signals together from the ISP?

In: 5

4 Answers

Anonymous 0 Comments

A multiplexer (and demultiplexer) is a device that combines multiple signals into one signal which can then later be split up again. And yes, it is a very vague term which can be used for a wide number of devices and software. I am not able to find any documentation I can understand on the device you have. But from what I can gather there are many slow speed inputs to it, probably one fibre from every floor, wing, department, or tenant in the building. There are then two high speed fibres from the unit that goes to a different building, maybe a datacenter or network central on the campus, base or business park. There a similar unit will split the signal into the same slow speed fibre cables. So you basically end up getting 10-20 individual network links through only two fibre cables.

The difference between a multiplexer and a network switch is that a multiplexer does not even look at the data. It will just pass the signal from one wire to another regardless of what the data is. A switch however, or even a router, will look at the data to find the destination address and send the signal in one of many different possible paths. If you think about this as a post office a multiplexer will just go around collecting out-trays into one big stack so it can be easily shipped to the post office. While a switch will look at the address on each envelope and then sort it into different inn-trays or send it to one of the neighbouring post offices.

Anonymous 0 Comments

I’m going to ignore the fact that you’re in a closet where these multiplexers are present and you don’t know what they are, because, although that’s somewhat troubling, they might be for unimportant systems or you might be someone who needs to enter that space for reasons other than maintaining them.

A multiplexer is indeed a device that is used to, in some sense, combine signals.

It doesn’t really combine them. Instead, what it does is it provides the ability to change which input signal is traveling to the output line at any given time. Let’s say you have four different signals coming into the multiplexer from, maybe, four different phone lines. But you don’t want to run for phone lines through a particular piece of conduit that’s particularly small. You’d rather run only two (I’ll get to why it has to be two in a second). One way you can do this is by installing a multiplexer.

As I said, the multiplexer is a device that switches between its various input lines and, depending on a control signal, spits one of those lines out at its output. You can think of it as literally just a switch where you select which input you want to send to the output. You might think this isn’t super useful in reducing the number of cables that you have to run through the conduit because at first it seems like all the multiplexer does, is allow you to pick which phone line is active, but you can still only have one phone line at a time.

That is technically true, but you can get around it. For example, “plain old telephone service” has a signal bandwidth of about 4,000 hertz. It turns out that, for any signal, you don’t have to watch it all of the time in order to reconstruct all of the frequency content in the signal. There is a result called the Nyquist-Shannon sampling theorem which tells you that if you’re looking at a wave, and you just take a sample at a given interval (that is, instead of trying to track the entire wave, you just only look at the value at any particular time at a regular interval) you will have enough information to uniquely reconstruct that wave as long as you take a sample at more than twice the frequency of the wave.

What this means for your switching is that you can actually combine all four telephone signals if you want to by switching between them so quickly that you’re getting enough samples for each phone line. Since you have four phone lines, each of which has a bandwidth of 4000 hertz, if you can switch between each phone line at a rate of over 32,000 hertz, that means you’ll get a measurement on phone line one, two, three, and four at 32,000/4 = 8,000 hertz each — that’s enough to reconstruct the entire phone signal. If you also have an appropriate device at the other end to identify which sections of the signal correspond to which phone line, and a device to decode the snippets back into intelligible phone traffic, then your multiplexer has just allowed you to put all four phone lines on a single cable.

The reason you need a second cable is that, at the other end, you need a way to identify which phone line is which. Even if you had a device that could separate the signals automatically without needing any information from upstream – which might not even be possible depending on what kind of data is being sent – the device would be unable to figure out which phone line was supposed to go where. The device downstream that can separate those signals (not necessarily automatically, as I pointed out) is called a demultiplexer. It does the same thing as a multiplexer, but in reverse. It takes one signal and sends it to a bunch of different outputs depending on what you tell it to do.

So, you need to run another line between the multiplexer and the demultiplexer so that they can share a control signal. That way, both the multiplexer and the demultiplexer no exactly which signal is which and when the switching is happening. When the multiplexer is outputting line 1, the demultiplexer is receiving a signal that’s also telling it “output this on line 1”. The same thing happens for line two, three, and four. So your four port multiplexer has allowed you to reduce four phone lines into a single cable that transmits the data and another cable that transmits the control signal. That might not be super useful in this example because we’re only talking about reducing four cables down to two, but imagine if you did it with 16 or 32 or 64 phone lines. You could really cut down on the number of cables you had to run.

Anonymous 0 Comments

A multiplexer is like a switch between different signals. Think for example about a home entertainment setup where you have one TV (with only one input) any many devices you want to connect to it (cable box, PS5, DVD player, etc). A multiplexer device (ex, a receiver) allows you to select an input source to feed the the TV.

I don’t know anything about the Nokia 1665 DMX specifically, but some quick searching suggests it’s like a fancy networking switch. It allows a person to connect a variety of different data interfaces to a high-speed fiber-optic network.

Anonymous 0 Comments

In my world a ‘multiplexer’ is a device that splits the light spectrum on a fiber optic cable into multiple smaller wavelengths, or ‘colors’. I mean, they are literally colors but mostly outside of the visible spectrum. What that allows us to do is take one fiber optic cable and make 100 or so separate connections.

If I use the whole spectrum for the connection, I would need one fiber strand to send data, and one fiber to recieve data. This is basic networking, a CAT6 ethernet works similarly, a number of pins send, a number of pins recieve. With a multiplexer, I could give you only one strand and you could have a send recieve pair (or multiples) on the same strand.

ISPs are interesting, I work for a US based on, we use something called hybrid coax. A coaxial cable is a cable that has a solid copper core and then a mesh copper that surrounds it that is separated by a medium that prevents crosstalk. A hybrid coax adds a fiber optic line to the middle of it. If you deliver TV over the copper, then the internet is delivered over the fiber and that fiber would have to be split into waves.