View Single Post
  #10  
Old September 30th 18, 07:32 PM posted to uk.telecom.broadband
R. Mark Clayton[_2_]
external usenet poster
 
Posts: 526
Default Openroach turbo master baiting too busy to install fiber

On Sunday, 30 September 2018 18:02:48 UTC+1, 7 wrote:
Interesting - 10Gbps fibre modems are as opposed to for 1Gbps.

I have no data on that.


I looked it up. 7 seems to have prices for 1k off.


Can they use a diffraction grating to extract multiple beams from the
same fibre?

It's possible. Whether it's cost effective is another matter.


A diffraction grating costs pence, however setting it up accurately would
be difficult.

It may be that all the fibre modems are using the same colour anyway.


They don't need to be colour sensitive. However this solution has
problems in the uplink direction - splitting a beam in the field is
probably pretty simple - it could be done in a passive module. Merging
them again is very difficult, so the uplink would either be lower
bandwidth or by other means (copper).


As 7 sort of points out this is a frequency division multiplex. It might
be easier to statistically multiplex 'bundles' of data for the same switch
at the exchange and distribute these to high speed switches at the cabinet
- these would unbundle for 10Gbps switches to say 100 users or
alternatively just lay 8 core fibre (as at present AIUI).


That is because you are an Internet troll and have no idea which is cheaper
if any.
None of those are cheap because the cost of buffering 1 second of data
with high speed RAM for multiplexing is enormous.


A whole second - how long are your frames? Even then at 100Gbps 1s = ~12..5GB - lets say 16GB including some overheads is about 40.

Of course you clock it in serially and store it in parallel - the late Dr. David Dalzell worked out how to do that [using ECL] 30 years ago, although I easily guessed his method.

Better to have direct symmetric gbit Internet without buffering
than buggering up everyone with multiplexing of any kind for FTTP.


The idea of laying 3456 fibre cables to cabinets that where I am serve at
most a few hundred premises is clearly ridiculous.



That is because you are an internet troll and don't know what you want.
Its about $1 per 50 meter per individual fiber cost. So if you got 10,000
people in blocks of flats to connect up, you need 3 of these cables
instead of hundreds of cables because it saves a lot of space, time
and money bringing fiber to an area because nothing else can compete
with it.


10,000 flats in one place? - even the Barbican ain't that big.

Round here there are FTTP cabinets about 500m from each other, which means that each one has a few hundred premises within an approx 250m radius.

So kilo fibre cables are neither necessary nor even desirable and anyway you still have to get from where you split them (cabinet) to user premises.

BTW yesterday me and a mate were watching the footie on BT TV at full HD - fast moving picture obviously. Unbeknown to me my partner was watching youtube pop videos on my lap top via Wi-Fi and we were surfing on a tablet and an S8 and S9. All on then end of an ~50Mbps link.

Did the picture hiccup - NO.

Where does this requirement for domestic gigabit broad band actually come from? What would subscribers use it for - looks like a solution looking for problem to me!

I wonder how sales at Hyperoptic are going?




In any event it is the cost and difficulty of getting the fibres to the
premises that is the obstacle. Even if you put them in 10cm deep
micro-trenches they will still get severed every time any other utility
has to dig up.


Probably openroach and others have this habit of knocking out
competitor cables. Others use a map and take out insurance
to prove they ain't stupid.


Gas, water, sewers, electric, VM, BT, council traffic light wiring - we have had a problem like that round my way for ages - possibly as long as a whole week...