So fiber optic is used top transfer information around the world, including this post, between my computer and the SE server, wherever that is. But as I read, that info is transmitted through light pulses, on/off indicating 1/0 in terms of bits. I understand how we can transmit a simple message from source to destination (e.g. morse from lighthouse to ship)

But electronic information is complex and is routed through common pathways (fiber optic). Therefore, such communication must include metadata. For instance, this post is sent to SE by me. But only this post and not the other thing I am doing in another tab of my browser. Thus, there is the information itself as well as the metadata about it (where, to whom, beginning/end, date, etc). How can simple light pulses transmit both data and metadata?

Transmission of info, afaik, is done through some translation mechanism, converting complex info into binary data. But how is metadata processes? How does the converter know when a piece of info begins and ends and where to send it?

  • 2
    $\begingroup$ There are many different layers to data transmission, and the metadata and data are combined in some of them, and separated in others. For example in good old RS-232 communication, the stop and parity bits are not 'data' per se, but are part of the protocol as metadata. $\endgroup$
    – Jon Custer
    Apr 21 '20 at 19:58

As mentioned by Jon Custer in the comment, there are many layers with different protocols in computer network communication. The protocol is for example WWW for web browser which is acting on the highest level (so-called application protocol). Data and metadata from your browser are sent to lower layers and in the end they appear on physical layer (optical fibres, wires or EM radiation). At this layer the data are encoded to light or electrical pulse symbolising zeros and ones (I will not dive into modulation of signal and other technical subtleties) and then transmited. At this layer, there is no difference between data and metadata. All are just bits represented by some physical quantity - light intensity or polarization, voltage etc.

The physical layes does not understand data, it only transmits them. When the data appear in the receiver, a context is given to them by higher layers (and communication protocols working on these layers), information is unfolded and data and metadata can be distinguished.

In short, the light pulse does not distinguish between data and metadata as both are simply somehow encoded data (zeros and ones in binary communication).

  • $\begingroup$ Thanks. Are you saying that there are different physical layers of data transmission? That the pulses are multidimensional? Have different attributes? Perhaps data is in the light itself and metadata in the intensity, duration, or some other physical property of the light? If not, and it's literally just one dimension, on/off, then I still don't get how the different layers are transmitted using the same, on/off mechanism. $\endgroup$
    – luchonacho
    Apr 21 '20 at 21:49
  • 1
    $\begingroup$ Light has wavelength, amplitude, phase, polarization even in single-mode fibre. In multimode fibre there are also modes. There are a lot of dimensions and a lot of ways of representing them. Light in fibres, as an information carrier, is not that different to radio-waves in coaxial fibres. The main difference is that in coaxial cables the electromagnetic waves are strongly confined (wavelength much larger than cable cross-section), whereas in fibres the waves are weakly confined (wavelength about 100 smaller than fibre-cross-section). $\endgroup$
    – Cryo
    Apr 21 '20 at 23:10
  • $\begingroup$ @luchonacho: There is one physical layer and there is no diference between data and metadata. $\endgroup$ Apr 22 '20 at 6:22
  • $\begingroup$ So how can the receiver distinguish between data and metadata? $\endgroup$
    – luchonacho
    Apr 22 '20 at 14:24
  • 1
    $\begingroup$ @luchonacho They define protocols. For example in Ethernet (not sure about optical Ethernet, this is electrical Ethernet): the first bits are 101010101010 which helps the receiver lock onto the clock speed. Then, instead of 1010 it goes 1011 and that means this is the end of the locking-on part. The next 48 bits are the address where the data is going. The next 48 bits are the address it came from. And the next 16 bits say what other metadata there is. $\endgroup$
    – user253751
    Oct 9 '20 at 18:00

The trick is that the system is designed in layers, and the data for one layer includes the metadata for the next layer. Here's how a hypothetical system could work:

Layer A: So the fiber-optic transmitter - the physical layer - only knows how to transmit a 1, transmit a 0, or turn off.

Layer B: The next layer understands that if the transmitter was off and then turns on, that's the start of a packet, and if it was on and turns off, that's the end of a packet. And it knows how fast the 1s and 0s are supposed to arrive and how to lock onto the same timing as the transmitter. This makes packets of 1s and 0s.

Layer C: The next layer looks at the packets of bits from Layer B and understands that the first several bits of the packet are just a guide to lock on the timing and can be ignored. And it understands that the last several bits of the packet are a checksum and it ignores the packet if the checksum is wrong. And it understands that the number of bits must be a multiple of 8.

Layer D: The next layer understands that the first 48 bits are the destination address, the next 48 bits are the sender address, and the next 16 bits identify the next layer after this one (the path splits here depending on the type of packet).

And so on.

Notice that everything is data in Layer A. In Layer B, the transmitter being off is metadata. It processes that fact but doesn't tell Layer C. In Layer C, the first and last bits in the packet are metadata. It processes those bits but doesn't tell them to Layer D. In Layer D, the addresses and the next layer number are metadata. It processes those but doesn't tell them to Layer E. And so on.

Obviously, each layer needs to be designed so that it knows where the metadata is. Layer D knows that the first 112 bits are always metadata, for example, and it knows the format of those 112 bits.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.