What happens after the Fibre rollout?

I apologize if this is a rather silly question which I feel it might be, but what happens when the majority of the UK have access to a gigabit option?

Where would Openreach and other ISPs shift their focus? and I realize this probably requires quite a lot of speculation, but if the goals shifted then to even faster speeds like 10 gbps and the like, would that require another overhaul of the network, running different cables and the like or is it something that in the future could be more easily upgraded?

I suppose I'm asking where the bottleneck is.

Edited by Xuse (Thu 02-Dec-21 04:52:23)

Just talking about the so called “last mile” part of the network…

At some point in the future the active equipment (electronics & optics) on either end (at the exhange etc and in the premises) will be swapped or replaced. The optical fibre itself and any other components that form the passive part of the network will stay the same and not require any physical upgrading. The raw data carrying capacity of single mode fibre that forms the basis of FTTP is in the terabits so quite a long way from needing to be changed.

There is already thought and planning that has gone into this both (1) at the level of standards development where for example next tier technology uses a complimentary set of wavelengths of light such that the existing/legacy protocol can coexist with the new tier (eg GPON with XGS-PON or 25G PON or 50 GPON) and (2) at the network operator level where for example Openreach install a WDM (wavelength division multiplexer) at the exchange to allow a second feed into the network without altering or disconnecting any existing services and customers.

The “core” part of the network is continually being upgraded as well, but is obviously far less visible. There are multiple providers not just BT that operate and build their own independent core networks that can be leveraged by ISPs.

The process here is similar - the raw fibre cabling may get added to but it not usually ripped out and replaced. The “ends” just get fatter and faster so for example network speeds here operate in the 100’s of Gigabits per second.

To give an example there was a press article on Tuesday announcing Neos Networks providing access to 550 BT exchanges with their 100G network here. They will connect a further 150 exchanges next year.

Companies like Google now build their own transcontinental undersea fibre networks - no longer reliant on traditional telco operators. It goes on and on.

In reply to a post by Xuse:

if the goals shifted then to even faster speeds like 10 gbps and the like, would that require another overhaul of the network, running different cables and the like

In short, no. The fibres deployed (single mode fibre) have vast capacity. This is a one-time upgrade.

But to deliver (say) 10G or 25G services, you need new electronics at the exchange end and the customer end. The standards work in parallel on the same strand of fibre using different wavelengths of light: so on a single passive optical network you can be serving a bunch of existing customers with GPON *and* other customers using the newer technologies at the same time.

As already said, fibre has huge capacity with WDM etc... the record is in excess of 178,000Gb/s

https://www.ucl.ac.uk/news/2020/aug/ucl-engineers-se...

This is a single mode fiber and (very) expensive end sets, but the FTTP deployed fibre in the
last mile will be able to go to >10Gbps easily, at a cost. I doubt very much demand will require
it in the next 10 years.

Going above a Gbps in a home is probably a waste of time for 99.9% of domestic users.

I'm hoping to get FTTP in the next few months, with 500Mbps I'll be able to download a
5GByte movie (in 4K with 5.1 sound) in 80 seconds, it'll take me 90 minutes to watch.

Deployed Wifi delivers significantly less than 1 Gbps. Most 802.11ac APs are 866Mbps peak,
and less for most connections (i.e. not standing next to access point). Wifi 6 offers higher
throughput, but is intended for offices and applications with multiple users.

Wired equipment faster that GbE has dropped in price a lot in the last ten years but 10GbE/10GbT
switches are still £100s (down from >$1000/port a decade ago). I watch this space carefully but doubt I'll
be able to justify to Mrs HeadQuarters (or myself) the extra cost of re-equiping our home network.

On AliExpress I've seen 2.5GbT wired cards for less than £15 but even the cheapest switches are £200.

https://www.aliexpress.com/item/1005001929449005.html

Don't get me wrong I'm nuts about networking. I've 10GbE fibre kit in my workshop and have worked on
100Gbps equipment but defo don't need to deploy it domestically.

What would I (or the pro domestic user, SoHo or small commercial) need the capacity for?

ORs backbone will have 100GbE kit deployed, I saw it at Martlesham in the lab in 2008. $500k a port then. They could probably get away with never laying another fibre in the backbone with the per fibre
speed improvements.

My worries (and thoughts) are for the last 5% of mostly rural users on < 2Mbps ADSL who'll not be
connected in the next few years, some maybe ever. I guess there's always StarLink but at a high
price.

Thanks everyone for sharing how it works and helping me to understand it a little better. It's good to know that the option is there for them to upgrade should there be a need for it. It would have been quite rubbish if they had to redo all the cabling etc in 10 years or whatever.

I know it's unlikely for us to need faster than gigabit internet, but at the same time we don't know what the future holds there might be something in our future that does require it though admittedly, I am at a loss as to what it could be. Some kind of new video tech that requires exceedingly high bitrates and resolution ?

In reply to a post by fleddy:

Wired equipment faster that GbE has dropped in price a lot in the last ten years but 10GbE/10GbT
switches are still £100s (down from >$1000/port a decade ago). I watch this space carefully but doubt I'll
be able to justify to Mrs HeadQuarters (or myself) the extra cost of re-equiping our home network.

On AliExpress I've seen 2.5GbT wired cards for less than £15 but even the cheapest switches are £200.

https://www.aliexpress.com/item/1005001929449005.html

Don't get me wrong I'm nuts about networking. I've 10GbE fibre kit in my workshop and have worked on
100Gbps equipment but defo don't need to deploy it domestically.

10G fibre SFP's are pretty cheap (£20), even if you go for BiDi's (about £30) so just a single fibre core needed. Go shopping for them on fs.com. A silent passively cooled MikroTik CRS309-1G-8S+IN 8 x SFP+ port 10G switch is £159+VAT from euroDK. The four port version is £87 before VAT

What would I (or the pro domestic user, SoHo or small commercial) need the capacity for?

Moving big files about on and off NAS, high-res game production, animation or video editing at 8K or beyond in future. Plenty of folks now a re based from home or freelance from home

ORs backbone will have 100GbE kit deployed, I saw it at Martlesham in the lab in 2008. $500k a port then. They could probably get away with never laying another fibre in the backbone with the per fibre
speed improvements.

Layer 2 kit from ISP that connects at the exchange from and ISP onto the FTTP OLT will be using 10 Gbps Cablelinks.

Backbone networks now are routinely running 100G, 400G and shifting to 800G

As far as I know, OR doesn't have any backbone network to speak of - apart from the PSTN, which is non-IP and being switched off by 2025 anyway. IPStream and Datastream have been canned.

If you're an ISP and want to pick up FTTC or FTTP traffic, then either you bring your own network to OR headend exchanges, or you buy capacity from a third party backbone network such as BT Wholesale or Talktalk.

In reply to a post by Xuse:

Thanks everyone for sharing how it works and helping me to understand it a little better. It's good to know that the option is there for them to upgrade should there be a need for it. It would have been quite rubbish if they had to redo all the cabling etc in 10 years or whatever.

I think this is something that a lot of people don't get. The single mode fibre that is being deployed today is for ever. Anything better would require the known laws of physics to be broken in regimes that they have been tested to death in. About the only way to get better is point to point neutrino communication but that would require a monumental breakthrough in neutrino detection and generation to be practical, but would give speed of light line of sight communication between any two points, where the line of sight means through the middle of the earth to Australia. Back in the real world that is never going to happen.

I am sure someone will now mention hollow core fibre. Thing is this does not increase bandwidth just reduces the latency as the light is now travelling in air instead of glass. But if you are 20km from the exchange it would reduce latency by 0.067ms on a ping which unless you are a bottom feeding high frequency trader is meaningless. Even for a hard core gamer 0.067ms is not going to make any difference. All that is presuming that they actually work out how to mass produce it in the first place, and come up with practical methods of splicing it.

Actually technically hollow core fibre could give faster speeds as it reduces the dispersion in the light. However given ordinary fibre can go to many Tbps replacing the fibre from the exchange to your house for more bandwidth is never going to happen unless it suffers damage.

What I say is given it will take well over a decade to get everyone hooked up to a fibre, think what broadband speed you had 15 years ago. Would you be happy with that today? So while you don't need the speed now you can't predict the future and given the major cost in the fibre roll out is putting the fibre in not the equipment at either end, and that it will never ever need upgrading again best to get rolling out the fibre now because if we wait till we actually need the speed it will be a decade or more before we can all have it.