Replacing fibre with the next gen fibre

So now it seems our fibre is 'so last century' and in the future may need replacing with OAM compatible fibre which could up the data throughput by 100 times.

It has always been the case - anyone that thinks fibre is truly "future proof" doesn't understand technology. However, current fibre can suit the majority of needs for a good number of years, it will however need to be replaced at some point though with whatever the "new" thing is (some think it will be wireless technologies but I would still put my money on some sort of new "cable" going in the ground).

Current generation of fibre in the ground has several speed iterations left, the optics that generate the light may need changing but lots more speed can be squeezed out of it

So no panic needed unless you are someone selling the latest and greatest and want higher sales figures

In reply to a post by burble:

So now it seems our fibre is 'so last century' and in the future may need replacing with OAM compatible fibre which could up the data throughput by 100 times.

OAM = Operations, Administration and Management: this applies to the electronics at each end, not to the fibre itself.

Existing single-mode fibre (e.g. G.652.D) can carry 10Gbps over tens of kilometres. With DWDM you can send many 10Gbps channels over the same strand. And the technology to light the fibre is continually improving.

Not quite sure what prompted your comment, but this was in the guardian today

Also from a year ago https://www.gla.ac.uk/news/headline_555908_en.html

Their picture also makes it clear that this is aimed at the core network rather than the final drop to homes and business which is how a lot of the non scientific press is covering it.

In reply to a post by MrSaffron:

Also from a year ago https://www.gla.ac.uk/news/headline_555908_en.html

Their picture also makes it clear that this is aimed at the core network rather than the final drop to homes and business which is how a lot of the non scientific press is covering it.

Yes I'm sure it is being considered mainly for core network, but who knows where we are going, it's not long ago 56k dialup was state of the art, then when 512k broadband came along it was 'wow', now my 4MB connection is considered by many to be totally inadequate. In fact I've just remembered when we got the first IBM PC in the office, the choice had to be made between 10,20, or 30MB hard disc.

In reply to a post by burble:

In fact I've just remembered when we got the first IBM PC in the office, the choice had to be made between 10,20, or 30MB hard disc.

I used to sell micro-computers before (and after) the IBM PC was invented, before hard discs for such existed. When they did come on the scene, a 5MB hard disc added approximately £1000 + Vat to the selling price.

A 128GB SD card to go in my dashcam now costs £39.99 from Argos!!!

Mind-boggling for those who were there.

My first hard drive was I think 10 - 20MB which took up 3 5.25inch bays and required its own ISA Controller Card.

There was no optical drives and my floppy drive was a single sided 5.25 inch drive, so that was all my bays taken up.

Also back then there was no internet, all we had was BBS to call and play, message on.

Ah those were the days

Look at the Hard drives now, 8TB each, SSD @ +2TB, RAM 32 - 128GB upwards, its just mind blowing how much has changed over the last 10 to 20 years.

Paul

Edited by PaulKirby (Thu 25-Oct-18 08:13:45)

I agree, however it is really terrifying how inefficient things appear to have become. 512K of memory and a 20Mb hard disk used to be able to achieve an incredible amount.

One of the biggest drawbacks as I see it is that now everybody is expected to have massive storage/ massive memory and a high speed internet connection for downloading huge updates. Luckily I do live in a high speed BB area, however I dread to think what one of these updates does to a slow rural connection. That is quite apart from the help systems moving on line, including those to do with connection problems.

My first business computer was a six foot rack, paper tape based and no disk at all. 32K ( yes K) magnetic core memory. We had a huge upgrade to 64K to allow multi tasking.

In reply to a post by burble:

Yes I'm sure it is being considered mainly for core network, but who knows where we are going, it's not long ago 56k dialup was state of the art, then when 512k broadband came along it was 'wow', now my 4MB connection is considered by many to be totally inadequate. In fact I've just remembered when we got the first IBM PC in the office, the choice had to be made between 10,20, or 30MB hard disc.

Remember that you can do symmetric 10Gbps down a single ordinary OS2 fibre today at distances up to 100km with an SFP+ delievered to your door tomorrow. They are pricey you will need to SFP+'s and they are £1000+VAT each mind you.

WIth two fibres you can go all the way to 100Gbps at distances up to 40km, with a transceiver delivered to your door within the week. These are even more expensive at £4k+VAT each and again you would need two.

You can go even faster over even longer distances but tight now that is custom carrier grade stuff reserved for long distance fibre links.

Given symmetrical 10Gbps can be done down the fibre everyone is laying right now with off the shelf components with higher speeds down the same fibre being very likely nobody is going to be using anything but OS2 single mode cable for homes and SME's drops for a very long time, probably not in my lifetime.

Yes its good for 100Gbps using off the shelf components today. Stick some CWDM splitters at each end and get the right QSFP28 transceivers and you can do that fully symmetrical down a single fibre. It is actually good for at least 1Tbps probably more. That is a lot of extra generations.

All the next generation fibre technologies are for backbone only and that is likely to remain true for the lifetime of anyone alive today.

In reply to a post by jabuzzard:

All the next generation fibre technologies are for backbone only and that is likely to remain true for the lifetime of anyone alive today.

The life expectancy of males born in the UK in 2017 was 79.2 years. Which is rather a long time for technological development .

Including the medical advances that could well increase that to well over 100.

I'm somewhat amused by the "elder" which appears beneath your username, given the context of your post..!

In reply to a post by jabuzzard:

In reply to a post by burble:

Yes I'm sure it is being considered mainly for core network, but who knows where we are going, it's not long ago 56k dialup was state of the art, then when 512k broadband came along it was 'wow', now my 4MB connection is considered by many to be totally inadequate. In fact I've just remembered when we got the first IBM PC in the office, the choice had to be made between 10,20, or 30MB hard disc.

Remember that you can do symmetric 10Gbps down a single ordinary OS2 fibre today at distances up to 100km with an SFP+ delievered to your door tomorrow. They are pricey you will need to SFP+'s and they are £1000+VAT each mind you.

WIth two fibres you can go all the way to 100Gbps at distances up to 40km, with a transceiver delivered to your door within the week. These are even more expensive at £4k+VAT each and again you would need two.

You can go even faster over even longer distances but tight now that is custom carrier grade stuff reserved for long distance fibre links.

Given symmetrical 10Gbps can be done down the fibre everyone is laying right now with off the shelf components with higher speeds down the same fibre being very likely nobody is going to be using anything but OS2 single mode cable for homes and SME's drops for a very long time, probably not in my lifetime.

You can source ZR1000 SFP+ transceivers for about 1/3 the price you've quoted from FibreStore. Actually £344.53 + VAT @ 19% ex Germany.

https://www.fs.com/uk/c/10g-sfp-plus-63?100-120km=481

I've only bought plain old 1000BaseLX (10km) 1GB SFPs from them, but they were less than a tenner each including VAT, and they have been perfect so far.

Edited by Pheasant (Thu 25-Oct-18 16:47:44)

I think with current technology able to get over 4Tb/s over a single fibre pair we should be okay for a bit

In reply to a post by Ignitionnet:

I think with current technology able to get over 4Tb/s over a single fibre pair we should be okay for a bit

I think there was a place in Germany or some place that got up to 42Tbits in their tests, but if I recall it was over a few fibres.

So I think the cables we have now will be fine for a while.

Paul

In reply to a post by Pheasant:

You can source ZR1000 SFP+ transceivers for about 1/3 the price you've quoted from FibreStore. Actually £344.53 + VAT @ 19% ex Germany.

https://www.fs.com/uk/c/10g-sfp-plus-63?100-120km=481

I've only bought plain old 1000BaseLX (10km) 1GB SFPs from them, but they were less than a tenner each including VAT, and they have been perfect so far.

You can, but I was quoting for BiDi optics aka using a single fibre (from fs.com in fact) and you are pointing to normal duplex fibre optics. The relevance been that the final drops being installed by the majority of FTTP providers are simplex so solutions that avoid replacing the final drop (the most expensive bit of any FTTP install) are of particular relevance.

In reply to a post by RobertoS:

In reply to a post by jabuzzard:

All the next generation fibre technologies are for backbone only and that is likely to remain true for the lifetime of anyone alive today.

The life expectancy of males born in the UK in 2017 was 79.2 years. Which is rather a long time for technological development .

Including the medical advances that could well increase that to well over 100.

So OS2 fibre is good for at least 1Tbps probably more, I cannot see in the next 100 years needing that sort of bandwidth to your home. This is the same single mode fibre that you might have installed 30 years ago for 10BaseFL Ethernet.

I think people just don't quite grasp how amazingly future proof single mode fibre optic cable is.

20 years ago most people would have likely had a maximum of 56Kbps. Now we have an average of I think around 33Mbps after 20 years. If we have the same multiplication over the next 40 years then we will already need more than 1Tbps. Maybe things will slow down but generally the pattern holds in IT.

It is quite possible we will be shifting inanimate objects over the net in less than 50 years. House (contents) moves by wire not expensive removal vehicles.

As we will have quantum home computers, why not quantum data packets?

I was lucky because 20 years ago I was part of my employer's telecommuting trial and had an ISDN30e line with a whopping 128 Kbps. I also discovered that my ISP supported ISDN on their dial-up 0800 number so could get rid of the 56K modem and free up my phone line.

In reply to a post by ian72:

20 years ago most people would have likely had a maximum of 56Kbps. Now we have an average of I think around 33Mbps after 20 years. If we have the same multiplication over the next 40 years then we will already need more than 1Tbps. Maybe things will slow down but generally the pattern holds in IT.

Yeah that's just not going to happen. The question is what could you require 1Tbps for in your home? A 4k 50fps HDR stream is only ~50Mbps..

Past performance is not an indication of future performance. There is two maybe three process shrinks left, perhaps with a final switch to a more exotic semiconductor as a final roll of the dice, then it's all over as immutable laws of physics get in the way of further improvements.

Sure you might have quantum computing, but that won't be in your home or a mobile device as liquid helium cooling prohibits that. Improvements have radically slowed in recent years and will continue on that path.

Home computers with spinning discs drives will never happen either, as the home air-conditioning isn't good enough.

I still remember back in the 80s showing a lad in the office how fast a 5.25" floppy was compared to the cassette tape on his home computer.

In reply to a post by jabuzzard:

In reply to a post by Pheasant:

You can source ZR1000 SFP+ transceivers for about 1/3 the price you've quoted from FibreStore. Actually £344.53 + VAT @ 19% ex Germany.

https://www.fs.com/uk/c/10g-sfp-plus-63?100-120km=481

I've only bought plain old 1000BaseLX (10km) 1GB SFPs from them, but they were less than a tenner each including VAT, and they have been perfect so far.

You can, but I was quoting for BiDi optics aka using a single fibre (from fs.com in fact) and you are pointing to normal duplex fibre optics. The relevance been that the final drops being installed by the majority of FTTP providers are simplex so solutions that avoid replacing the final drop (the most expensive bit of any FTTP install) are of particular relevance.

Just a minor point you've the terminology the wrong way around. Fibre pairs are simplex, single fibre solutions duplex using FDM.

In reply to a post by jabuzzard:

Past performance is not an indication of future performance. There is two maybe three process shrinks left, perhaps with a final switch to a more exotic semiconductor as a final roll of the dice, then it's all over as immutable laws of physics get in the way of further improvements.

Sure you might have quantum computing, but that won't be in your home or a mobile device as liquid helium cooling prohibits that. Improvements have radically slowed in recent years and will continue on that path.

And industry experts have been saying all of that for years. It was quite a while back they said the laws of physics wouldn't allow the shrinking process to continue for processors - yet still it does, they keep finding new ways to keep it going.

The predictions that IT tech will stop in its growth rate have been going on for decades, at some point they might stop but I wouldn't count the industry out from finding yet more ways to keep people upgrading.

ADSL Max (8Mbps) was definitely all anyone would need as it could stream SD TV with absolutely no issue (and even HD streams) and what would anyone need more for.

Where did someone say ADSL Max (the Max part is a marketing rather than a standard addition) was all anyone would need?

Yes a 7 Mbps ADSL connection will support SD and HD streaming, but how many got the maximum sync?

Various people on this site at various times have said "why would any upgrade to x as y is all they need". It is a recurring theme.

For some people it is the case that they may be able to manage with speeds of 6 or 7 Mbps so those comments may be perfectly reasonable.

Extrapolating to say that everyone only needs that sort of speed is stretching things, and the distance limits of ADSL have been well known for years.

I can't find any posts but whenever a new technology comes up someone invariably says "the [current technology] does everything anyone needs, why would anyone upgrade to [new technology]". Very similar to the comment that was posted that I responded to in this thread.

In reply to a post by Ignitionnet:

Just a minor point you've the terminology the wrong way around. Fibre pairs are simplex, single fibre solutions duplex using FDM.

Try going to fs.com and searching for simplex and then get back to me when you are ready to accept I am correct

Sure but this is different, you need room temperature and above super conductors. These are just as far away today as they where 20 years ago. In fact nuclear fusion power stations are nearer. Basically it needs a major break through in physics.The improvements in semi conductors since the 1970's have ALL and I mean ALL been engineering based. There has be no new physics period. The joy ride is nearly over.

In reply to a post by ian72:

In reply to a post by jabuzzard:

Past performance is not an indication of future performance. There is two maybe three process shrinks left, perhaps with a final switch to a more exotic semiconductor as a final roll of the dice, then it's all over as immutable laws of physics get in the way of further improvements.

Sure you might have quantum computing, but that won't be in your home or a mobile device as liquid helium cooling prohibits that. Improvements have radically slowed in recent years and will continue on that path.

And industry experts have been saying all of that for years. It was quite a while back they said the laws of physics wouldn't allow the shrinking process to continue for processors - yet still it does, they keep finding new ways to keep it going.

The predictions that IT tech will stop in its growth rate have been going on for decades, at some point they might stop but I wouldn't count the industry out from finding yet more ways to keep people upgrading.

ADSL Max (8Mbps) was definitely all anyone would need as it could stream SD TV with absolutely no issue (and even HD streams) and what would anyone need more for.

The last time I heard this was when we got to a certain feature size we would start getting quantum tunnelling going on. For those that don't know what this is basically an electron can decide it does not need to continue going down this piece of wire and will "tunnel" out the wire and go somewhere else instead. Welcome to the weird and wonderful world of quantum mechanics.

This must have been the early 1990's because it was when I was doing my physics degree and doing just this on the course. I remember discussing it with some friends and doing some back of the envelope calculations that sure the odd electron might tunnel out but if you have 100 electrons that make up a binary one and two tunnel away it's not really a problem.

On the other hand if you have 100 electrons and 90 of then decide to tunnel away you do have a problem. It's not hard (well if you have a degree or better in physics) to sit down and work out when that becomes an insurmountable problem. Basically it depends on the width of the "well", the height of the well and the energy of the electron.

We are really close to that point right now, and that is not even counting the incredible cost of mass producing devices at that scale. Just ask Intel who their 10nm process is working out for them.

There are a whole bunch of other factors too that limit continued improvements in transistor density and we are close to all those theoretical limits too.

So yes the end of the incredible joy ride that has been IT is close at hand without some form of new physics coming into play. However while there are lots of these in "labs" none have made it to practical use for a whole range of reasons. They are all further away than nuclear fusion power stations at the moment, despite having more money spent on them.

In reply to a post by jabuzzard:

Sure but this is different, you need room temperature and above super conductors. These are just as far away today as they where 20 years ago. In fact nuclear fusion power stations are nearer. Basically it needs a major break through in physics.The improvements in semi conductors since the 1970's have ALL and I mean ALL been engineering based. There has be no new physics period. The joy ride is nearly over.

You�ve used my simplified example of disc air conditioning to ignore my initial point about quantum computing.

As for that requiring liquid helium, you think that is impossible in the home in the lifetime of some people alive today? I had treatment years ago in a normal room temperature doctor�s surgery with a hand-held liquid nitrogen spray. A home computer inside a liquid-helium envelope is a perfectly reasonable expectation, even if such temperatures are still required.

You say there will be no new physics period! What on earth do you think quantum theory and mechanics is? You think that is the end of the line?

Scientists thought that when the atom was �discovered�. If your beliefs are correct then we are near to the end of research into particle physics. I await your Nobel Prize award for the proof of that.

By the way, you might like to follow that up with a paper explaining what dark matter is, seeing as nobody seems to know yet. Not that knowing will change anything, just like Edison didn�t change anything . There won�t be any new physics period ... ROFL.

Edited by RobertoS (Mon 29-Oct-18 11:56:30)

Helium is not a plentiful gas, so using it for home computing is a difficult cost justification

Nitrogen is a different matter and liquid cooled PC is already a reality, just not using nitrogen usually

True, but helium will be a by-product of the fusion power used to cool it .

All I'm basically arguing is against:-

In reply to a post by jabuzzard:

Yes its good for 100Gbps using off the shelf components today. Stick some CWDM splitters at each end and get the right QSFP28 transceivers and you can do that fully symmetrical down a single fibre. It is actually good for at least 1Tbps probably more. That is a lot of extra generations.

All the next generation fibre technologies are for backbone only and that is likely to remain true for the lifetime of anyone alive today.

Specifically, "... likely to remain true for the lifetime of anyone alive today". That is over 100 years. Everything will be different. Think back to 1918 and the advances in science and technology since then. We could well need 1Tbps to the home.

Edited by RobertoS (Mon 29-Oct-18 12:58:37)

If anything what needs to happen in IT and tech world is to look less at pushing speeds which usually increase power consumption and need esoteric cooling methods, but delivering the current speeds of processing possible but in lower power consumption and with less wasted heat.

This is what will drive things forward

In reply to a post by RobertoS:

We could well need 1Tbps to the home.

And if and when we do:

1. The technology to deliver it will be developed;
2. That technology will almost certainly work over the existing 9-micron single-mode fibre which is being installed today.

You can already run 100Gbps on such fibre, and more if you put sufficiently expensive equipment at each end.

The claim made at the start of this thread was that today's "last century" fibre would soon become obsolete and need to be replaced by some unspecified "next gen" fibre. I see no evidence for that claim.