What has my fibre optic internet (transmits using light frequency) have to do with electricity ? Hydro good out miles/kms away – not affecting our household, yet my internet goes out, why?
In: Technology
Fibre itself is just a light-tube, there are no electronics.
At each end are lasers and light detecting circuits. Powered by electricity.
Often the fibre run is to a local cabinet where it’s converted to electrical signals and then put down a different, better type of fibre that can travel many kms. That fibre then ends in… more electronics that do the same, and those are connected to things that do the same, and so on, around the entire world.
But your house fibre – if indeed it is fibre, because there’s a lot of misselling – is only a small local run. The actual fibre that connects you and a thousand other users back to a central hub will be local to you, and powered by electricity.
This is one of the things I have to explain to employers on a regular basis, for even “local” networks (i.e. a few hundred metres in either direction). Because the switches we used to deliver networking to the copper cables in your office and the wireless access points are almost always connected to each other by fibre (fibre doesn’t allow electricity down it, so you use fibre connections between buildings to prevent different electrical phases from allowing current to flow between the equipment and potentially starting fires or damaging equipment).
But they are powered by electricity, so if you have too many things in the critical network paths that are all reliant on power, things go off when something in the middle goes off.
In many workplaces, I have actually enforced that we do this:
– Switch A is connected to Switch B with fibre.
– Switch B is connected to Switch C with fibre.
AND
– Switch A is connected to Switch C with fibre.
This means that if the equipment running Switch B goes off, there is still a fibre-only-path between A and C. So long as A and C stay powered, they will stay connected to each other, even if B is off.
And this usually works out very simply because the fibres between A and B, and those between B and C will already have – say – 16 individual fibres in them.
I run 8 of those fibres directly to B, where it connects to the switch and converts the signal.
I run the other 8 directly into the 8 of the other fibre that’s going to C. Now there is a path that is entirely fibre between A and C and B can turn off without affecting it.
But still if A or C turn off, we lose the data in the fibre because one end isn’t powered to send/receive the data. So what you do is just extend this scheme in madly-connected ways.
A -> B -> C -> D -> A.
As well as B->D, A-C, etc.
Now any one switch going off makes no difference to the rest of the network because they’re all connected via fibre paths to some other switch which IS powered.
Every time I inherit a network, it’s designed like your local ISP network. One connection from you to some equipment, and that cabinet is reliant on a connection to more equipment and so on, with little resiliency or redundancy in it.
And every time I explain to my team what we need, and start building out, by the time I leave a site, every points is heavily connected to each other INCLUDING using fibres to just bypass certain switches. It costs almost nothing, in the grand scheme of things.
Because fibres are amazing and can connect you to equipment that’s literally kilometres away… but it’s all dependent on both ends having power to read/send the signal over the fibre.
Latest Answers