Internet download speeds compared - copper and cable TV platforms are banjaxed

Impetus

Netindex has released download speed comparisons (Dec 3) showing average download speeds in countries across the world. Ireland ranks # 40 – even slower than the US which ranks # 32.

The fiberized countries are at the top of the list – Hong Kong, Singapore, South Korea, Sweden, Andorra, Macau, Japan etc.

Even Romania has Gigabit fiber (http://www.rcs-rds.ro/internet-digi-net/fiberlink?t=internet-fix&pachet=digi_net_fiberlink_1000 )

Customers in countries that rely on copper (even VDSL2+) and cable get a raw deal, as multi-media content increasingly dominates. The number and picture resolution of TV picture quality deteriorates on cable systems that also try and offer internet and other services due to bandwidth limitations of co-ax and the need to use high compression ratios as a result. Fibre allows competition over the same platform. Internet speed relating to page fill times also suffer on copper/co-ax systems due to high contention ratios – which also leads to other problems like video buffering.

Netflix has 4K (Ultra HD) TV on the back-burner. 4K TV sets are on sale in most countries and deliver stunning pictures. DTT, satellite and cable haven’t a hope of delivering a wide choice of 4K channels, even with the latest CODECs. Customers and the media industry will be relying on a free, fast and open internet platform to deliver this content.

Thanks to competition in the computer industry, most people have access to very fast computers in 2013. As more and more services reside on a duopoly dominated internet (in Ireland’s case eircom and UPC), consumers and business users are stuck in the chain is as strong as its weakest link syndrome – ie the speed / contention ratio imposed by the internet provider.

http://www.netindex.com/download/allcountries/

Cuddlesworth

4K@30fps streamed is around 20Mbps at h264 and around 10Mbps at H265, which will be in mainstream use around the same time people actually start buying 4K Tv's.

Cable in Ireland in theory can do up to 400Mbps and upcoming is a gig with the 3.1 standard. VDSL can hit 120Mbps I think.

So where are you going with this? Network speed is starting to hit the same issue that other computing services are, any further performance gains are not noticeable to the normal user.

Impetus

In Ireland cable runs from house to house to house, so one cable has to typically support several dozen properties, or more. So even if they can push EURODocsis 3,1 to 1GB, and if this will work in practice over the legacy coax, the 1GB of speed is subject to division (contention) perhaps offering 35 Mbits/sec (real sustainable speed) to each household in practice (depending on cable network layout). This is a critical issue for IP video distribution systems.

As for VDSL2+, even if every house was say 200 m from the nearest box, (and if those boxes do not have to be removed due to planning permission bureaucracy), the real world distance between each house and the closest box might be 700m and the average house might experience a download speed of say 40 Mbits/sec – far from the theoretical maximum you cite.

Ultra HD has an uncompressed speed of 144 Gbits/sec – in the context of picture quality and how high does one go with compression – even with H.285/HEVC.

Each increment of stretching the speed of legacy resources requires substantial capital investment and only lasts for a limited period before showing its real world limitations. Surely the most intelligent approach is to focus the spending on industry-wide shared FTTP, and free up duct space by removing the copper (and only having to bear the running costs of one platform)? Fiber wins on maintenance costs too!

MrO

Impetus wrote: »

In Ireland cable runs from house to house to house, so one cable has to typically support several dozen properties, or more. So even if they can push EURODocsis 3,1 to 1GB, and if this will work in practice over the legacy coax, the 1GB of speed is subject to division (contention) perhaps offering 35 Mbits/sec (real sustainable speed) to each household in practice (depending on cable network layout). This is a critical issue for IP video distribution systems.

As for VDSL2+, even if every house was say 200 m from the nearest box, (and if those boxes do not have to be removed due to planning permission bureaucracy), the real world distance between each house and the closest box might be 700m and the average house might experience a download speed of say 40 Mbits/sec – far from the theoretical maximum you cite.

Ultra HD has an uncompressed speed of 144 Gbits/sec – in the context of picture quality and how high does one go with compression – even with H.285/HEVC.

Each increment of stretching the speed of legacy resources requires substantial capital investment and only lasts for a limited period before showing its real world limitations. Surely the most intelligent approach is to focus the spending on industry-wide shared FTTP, and free up duct space by removing the copper (and only having to bear the running costs of one platform)? Fiber wins on maintenance costs too!

This is spot on. DOCSIS and VDSL are milking that copper and coax infrastructure. It's hugely dependent on sharing in the final network section, distance based limitations and the quality of the oustide plant.

It's a stop gap. We couldn't envisage the amount of bandwidth required today back in the 90's - what will we require in another 10 years? Areas that have invested in quality access infrastucture (FTTH) drive innovation and attract investment.

Cuddlesworth

Impetus wrote: »

In Ireland cable runs from house to house to house, so one cable has to typically support several dozen properties, or more. So even if they can push EURODocsis 3,1 to 1GB, and if this will work in practice over the legacy coax, the 1GB of speed is subject to division (contention) perhaps offering 35 Mbits/sec (real sustainable speed) to each household in practice (depending on cable network layout). This is a critical issue for IP video distribution systems.

As for VDSL2+, even if every house was say 200 m from the nearest box, (and if those boxes do not have to be removed due to planning permission bureaucracy), the real world distance between each house and the closest box might be 700m and the average house might experience a download speed of say 40 Mbits/sec – far from the theoretical maximum you cite.

Ultra HD has an uncompressed speed of 144 Gbits/sec – in the context of picture quality and how high does one go with compression – even with H.285/HEVC.

Each increment of stretching the speed of legacy resources requires substantial capital investment and only lasts for a limited period before showing its real world limitations. Surely the most intelligent approach is to focus the spending on industry-wide shared FTTP, and free up duct space by removing the copper (and only having to bear the running costs of one platform)? Fiber wins on maintenance costs too!

You're basing your claims off a unsupported theory that curent infrastructure is unable to cope with future demands.

Your also assuming that the network will be at 100% utilisation 100% of the time. Which is a flawed assumption.

Current infrastructure for the most part is more then capable of meeting current demand. Future infrastructure(which 4k is) will undoubtedly be capable of meeting future demands as new technology comes forward. There is still a huge spectrum of analog frequencies waiting to be tapped into.

Also the point of H265 and other scalable codecs is that they can transition between resolutions and quality with the bandwidth available. So if the infrastructure becomes saturated, which is unlikely, those video streams will downgrade to a lower resolution. Most people don't notice netflix at the moment doing this and they won't in the future. Most people don't care about resolution past 480p. And the media companies have noticed.

Cuddlesworth

MrO wrote: »

This is spot on. DOCSIS and VDSL are milking that copper and coax infrastructure. It's hugely dependent on sharing in the final network section, distance based limitations and the quality of the oustide plant.

It's a stop gap. We couldn't envisage the amount of bandwidth required today back in the 90's - what will we require in another 10 years? Areas that have invested in quality access infrastucture (FTTH) drive innovation and attract investment.

All networks are based off sharing. Its been that way since the old PSTN networks, its like that now with broadband and mobile services. Its even like that with home networks. There is never enough capacity for 100% utilisation.

Your workplace is probably putting 48 people on a 1 gig link to the core. The core is probably dropping most of the company into a 1 gig link to infra switches in racks. And the whole organisation would usually be sharing 1-2Mbps between them to the outside world, if they are lucky.

Also, I can't see FTTH happening anywhere outside of really dense pre ducted apartment or housing developments in Ireland. Running any cable here is a nightmare. Half the houses I've looked at in the last three years are unable to order NTL due to neighbours not letting cable be run on the street.

crawler

Fibre and LTE will ultimately kill copper. The end.

MrO

Cuddlesworth wrote: »

You're basing your claims off a unsupported theory that curent infrastructure is unable to cope with future demands.

Your also assuming that the network will be at 100% utilisation 100% of the time. Which is a flawed assumption.

I can't speak for Impetus but I think it's interesting to talk about your points above (and the ones in reply to my post).

Current cable infrastructure is undoubtly able to support current demands, and where UPC high speed broadband is available in ireland they have (from what I can tell) a very high penetration rate. Future demands? Possibly, but with significant investment - the move to DOCSIS 3.1 is not cheap and cable companies will have some hard decisions to make.

Current copper (twisted pair) infrastructure though? Where VDSL 2+ has been deployed we can expect headline speeds of (with vectoring) of 100Mbit/s - best case. The reality is a significantly lower for reasons we all know about. The question of whether its good enough depends on how view things? is it good enough to support delivery of a packet TV service comparable to the cable offerings plus high speed internet. Is it good enough to support advanced services (e.g. medical monitoring at home) on top of this? Will it drive innovation of end user services in this country? I would say absolutely not - but time will tell.

So, the copper networks ability to cater for future demands? I think it's difficult to argue that we're not reaching (or have reached) the end of the road with copper. Even if you look at something like Gfast - youre talking about bringing fibre to the distribution point from the cabinet. You'd have to ask yourself - why not just go into the home with fibre. yes there is an expense - but once its done - its done.
To my knowledge there is no transmission medium faster that optical fibre - cant get more future proof than that.



In relation to your points about network utilisation/sharing, I'm certainly not advocating that networks are dimensioned to cater for 100% possible utilisation (and I doubt very much Impetus is assuming that either). Very few networks are dimensioned this way (there are exceptions) - it's just doesnt make any sense. However, networks should be dimensioned on peak utilisation and once a pre-defined threshold is exceeded, they should be upgraded so that no congestion is experienced by the end user (as opposed to the hard limits imposed by contention in the access network).

This unfortunately (for the most part) is not the way broadband has been delivered in Ireland to date.

Cuddlesworth

MrO wrote: »

I can't speak for Impetus but I think it's interesting to talk about your points above (and the ones in reply to my post).

Current cable infrastructure is undoubtly able to support current demands, and where UPC high speed broadband is available in ireland they have (from what I can tell) a very high penetration rate. Future demands? Possibly, but with significant investment - the move to DOCSIS 3.1 is not cheap and cable companies will have some hard decisions to make.

Cost will decrease over time. For a company like UPC, investment into 3.1 will be gradual, like it was with the previous standard. Its par for the course.

MrO wrote: »

Current copper (twisted pair) infrastructure though? Where VDSL 2+ has been deployed we can expect headline speeds of (with vectoring) of 100Mbit/s - best case. The reality is a significantly lower for reasons we all know about. The question of whether its good enough depends on how view things? is it good enough to support delivery of a packet TV service comparable to the cable offerings plus high speed internet. Is it good enough to support advanced services (e.g. medical monitoring at home) on top of this? Will it drive innovation of end user services in this country? I would say absolutely not - but time will tell.

Medical monitoring? I would say uptime is far more of a risk with that, then the small amount of bandwidth that it needs.

I'm actually not a big believer in copper. Not just from its current inception, but because of how it has been cabled. But as copper fades out, other options will arise. LTE and other wireless variants are not actually as bad as people think. Centralised distribution points off old telephone masts could easily cover the last mile at high bandwidth. Even technoligy within copper could provide larger speed boosts in the near term. There is enough of it out there to justify the R&D investment.

MrO wrote: »

So, the copper networks ability to cater for future demands? I think it's difficult to argue that we're not reaching (or have reached) the end of the road with copper. Even if you look at something like Gfast - youre talking about bringing fibre to the distribution point from the cabinet. You'd have to ask yourself - why not just go into the home with fibre. yes there is an expense - but once its done - its done.
To my knowledge there is no transmission medium faster that optical fibre - cant get more future proof than that.

So fiber has some serious problems. Its a 1 to 1 connection first of all. Its very expensive to buy per meter then copper or coaxial. Its very fragile and very prone to failure in any type of dynamic environment(such as hanging off poles). The equipment around fibre is more expensive and not as reliable, both in the cabinet and at the home. It requires far more advanced training to install, troubleshoot and fix. And the fix costs themselves would be of a magnitude more then copper or coaxial, due to the added complexity.

Keep in mind I'm not saying Fiber to the home is not the future. There is no better median for it. But we live in a country with two providers. Eircom and UPC. Eircom are broke. UPC have such a advantage there is no point in bothering. So I just can't see FTTH happening outside of very specific and high density locations where the infrastructure was put in place before hand.

MrO wrote: »

In relation to your points about network utilisation/sharing, I'm certainly not advocating that networks are dimensioned to cater for 100% possible utilisation (and I doubt very much Impetus is assuming that either). Very few networks are dimensioned this way (there are exceptions) - it's just doesnt make any sense. However, networks should be dimensioned on peak utilisation and once a pre-defined threshold is exceeded, they should be upgraded so that no congestion is experienced by the end user (as opposed to the hard limits imposed by contention in the access network).

This unfortunately (for the most part) is not the way broadband has been delivered in Ireland to date.

Networks are defined on a median average utilisation during peak hours, aimed at around 70-80% to provide a decent cost to performance ratio. My experience with ISP's is limited, but I wouldn't be surprised if they also provided the same service.

Impetus

Cuddlesworth wrote: »

Also, I can't see FTTH happening anywhere outside of really dense pre ducted apartment or housing developments in Ireland. Running any cable here is a nightmare. Half the houses I've looked at in the last three years are unable to order NTL due to neighbours not letting cable be run on the street.

FTTP/H installation costs are highest in urban areas due to ducting and related digging up the place. In and to rural areas, fiber can be rolled around powerline cables or use existing telephone poles. Towns and villages are no different to urban areas of big cities. One off houses/business premises can be served by fiber via powerline/phone poles.

Install a single, open fiber system across the country, and open it up to existing telcos, cable, and anybody new who wants a chunk of bandwidth that will enable "content providers" of all kinds to roll out their service nationwide - without re-inventing the wheel of the "last km".

It is the same as the national road network. We don't have a separate road network for Mercedes Benz cars, and another one for Volvo trucks. And another for buses. All brands and types of vehicle share the same tarmac. The same goes for electricity - independent generators don't set up their own network of pylons all over the place.

It is a sign of a mature, well planned industry. (Something telecommunications seems to know nothing about in Ireland). The costs of a nationwide shared fiber network (both capital and running costs) would be far lower and offer a far greater quality and variety of services than the current, ill conceived, collection of obsolete legacy platforms.

And it will have to be done anyway at some stage, by each operator who plans to survive in the marketplace on their own - with more duplication of costs and dig-ups. And this will limit competition to near zero.

The customer and country will suffer as a result. And the customer outside urban areas will have to resort to cans and string to communicate, because eircom will have been through it's 99th buyout, sucked dry and totally unable to provide even a basic reliable voice phone service to the country.

I suspect this "vision of fiber" would not be interesting to NTL, but there is no obligation on the State to renew NTL's license if it doesn't wish to work in the national interest and compete on a level playing field, etc.

MrO

Cuddlesworth wrote: »

So fiber has some serious problems. Its a 1 to 1 connection first of all. Its very expensive to buy per meter then copper or coaxial. Its very fragile and very prone to failure in any type of dynamic environment(such as hanging off poles). The equipment around fibre is more expensive and not as reliable, both in the cabinet and at the home. It requires far more advanced training to install, troubleshoot and fix. And the fix costs themselves would be of a magnitude more then copper or coaxial, due to the added complexity.

While there are challenges I would dispute some of the points

It is not a 1:1 connection, it can be - but most FTTH deployments are PON based i.e. use splitters at 1:32 or 1:64 (even 1:128)

The cost per meter for fibre is pretty cheap these days and to my knowledge cheaper than copper. I'm open to correction on this - the big spend is on installation not materials.

Overhead fibre (the self supporting type) used for hanging on poles is actually stronger and more protected than the copper equivalents - look up a spec for overhead cable from the big cable manufacturers.

The electronics associated with optical transmission are no more or no less susceptible to faults or reliability issues. In fact - it's the same big manufacturers have the market share for both, not only that but the same devices will perform both functions with swap-able modules (e.g GPON/VDSL/Active Ethernet).

Fibre does require more training to splice and troubleshoot, however I believe that all things considered, the operational costs associated with copper are much higher than that of optical networks given the age and fragility of copper networks. I think Telcordia did a large scale study on this a few years back which is probably available somewhere.

colm_mcm

This is all well and good, but in the future will country people be in a situation where they can't survive with say 50mbps. Will they all move into the cities because they have no choice?
I know time marches on, and Internet has evolved, but now that media like music/video is on the internet, a lot of content is viewed via cloud etc, is there really much farther to go?
Ultra fast internet will be great, but not having it is hardly going to be a massive issue is it?
4k will also be great but so many people haven't even got HD yet.

Impetus

colm_mcm wrote: »

4k will also be great but so many people haven't even got HD yet.

Nobody in Ireland has HD yet, unless they have a satellite dish on Astra 1 (or an expensive video on demand download subscription) and an HD TV to watch content from France, Germany etc, or are watching a BluRay disk. Sky is not HD with its low bit rate and 720p scan. NTL is not HD - the company is contenting already contented broadcasts. RTE2 DTT is not HD (just upscaled SD). Mush TV!

Impetus

MrO wrote: »

The electronics associated with optical transmission are no more or no less susceptible to faults or reliability issues. In fact - it's the same big manufacturers have the market share for both, not only that but the same devices will perform both functions with swap-able modules (e.g GPON/VDSL/Active Ethernet).

The electronics associated with fiber optical are far more isolated from the electrical environment that they are more reliable - eg a lightening strike on a fibre optic cable has zero impact - to the cable itself or the associated electronics at either end. Equally flooding / damp does no damage to fiber, while it can totally mess up copper.

Fiber maintenance costs are a fraction of copper or coax as a result.

dxhound2005

Impetus wrote: »

Nobody in Ireland has HD yet, unless they have a satellite dish on Astra 1 (or an expensive video on demand download subscription) and an HD TV to watch content from France, Germany etc, or are watching a BluRay disk. Sky is not HD with its low bit rate and 720p scan. NTL is not HD - the company is contenting already contented broadcasts. RTE2 DTT is not HD (just upscaled SD). Mush TV!

This is the list of HD channels on Astra 1

http://www.flysat.com/astra19-hd.php

Which of those are the French and German ones you mean? And how are they different from say BBC 1 HD?

bealtine

crawler wrote: »

Fibre and LTE will ultimately kill copper. The end.

The only thing LTE will kill is your pocket...and perhaps the networks that supply LTE. As a shared resource between subscribers it's hard to see how LTE will kill off anything:)
Once the voice cross subsidization disappears then the real price of LTE will become apparent...

Fibre will ultimately kill copper though that's for sure and may the day come soon

Impetus

dxhound2005 wrote: »

This is the list of HD channels on Astra 1

http://www.flysat.com/astra19-hd.php

Which of those are the French and German ones you mean? And how are they different from say BBC 1 HD?

1080i high bitrate:

http://www.tf1.fr
http://www.france2.fr
http://www.m6.fr
http://www.arte.tv/fr
http://www.arte.tv/de
http://www.servustv.com
http://www.zdf.de

etc etc - data taken from a set top box I have.

Sky is 720p and heavily contented - I don't have that tabloid bouquet myself, but if one looks closely at the picture in a shop selling TVs in Ireland, the supposed "HD" is far fuzzier than the picture I observe at the same distance from the above example list of channels.

Impetus

bealtine wrote: »

The only thing LTE will kill is your pocket...and perhaps the networks that supply LTE. As a shared resource between subscribers it's hard to see how LTE will kill off anything:)
Once the voice cross subsidization disappears then the real price of LTE will become apparent...

Fibre will ultimately kill copper though that's for sure and may the day come soon

I have LTE in a continental country for some time. When it came out first, it was extremely fast. As more and more people signed up, LTE slowed down. It is still faster than HSPA+, and good for on the go data needs, but could never stand in place of fiber or even wireline data for home or business use surfing and watching video etc on PC or larger screens.

MrO

Impetus wrote: »

The electronics associated with fiber optical are far more isolated from the electrical environment that they are more reliable - eg a lightening strike on a fibre optic cable has zero impact - to the cable itself or the associated electronics at either end. Equally flooding / damp does no damage to fiber, while it can totally mess up copper.

Fiber maintenance costs are a fraction of copper or coax as a result.

Absolutely, there is no chance of impact as long as the cable has no metallic support element and even then it would be minimal as this would/should be isolated from the kit at the exchange/PoP site. I suppose I was thinking of the electronics in isolation i.e. comparing reliability between 'all-optical' equipment and that used to terminate copper or coax network connections .

Either way the point is operational costs are lower.

MrO

bealtine wrote: »

The only thing LTE will kill is your pocket...and perhaps the networks that supply LTE. As a shared resource between subscribers it's hard to see how LTE will kill off anything:)
Once the voice cross subsidization disappears then the real price of LTE will become apparent...

Fibre will ultimately kill copper though that's for sure and may the day come soon

Also, you'd have to imagine that for any reasonable LTE coverage - we'd need a substantial amount of access fibre to backhaul the traffic...

Alun

Impetus wrote: »

Sky is 720p and heavily contented - I don't have that tabloid bouquet myself, but if one looks closely at the picture in a shop selling TVs in Ireland, the supposed "HD" is far fuzzier than the picture I observe at the same distance from the above example list of channels.

All the German channels I receive, including the ones you mentioned, on Astra 19.2E are 1280x720p.

As far as Sky / Freesat / FTA goes, all the BBC / ITV / CH4 HD channels on Astra 28.2E are 1920x1080i. Although I understand the BBC are actually using an adpative system that switches between progressive and interlaced on the fly as needed to optimize bandwidth.

Irish DTT HD channels are all 1440x1080i, although the vast majority of the content is currently upscaled SD.

BTW Don't go by what you see when looking at TV's in shops in Ireland. They're often badly set up and fed with signals of dubious origin and quality.

Impetus

Alun wrote: »

All the German channels I receive, including the ones you mentioned, on Astra 19.2E are 1280x720p.

Rubbish! What is your source. My set top box is reporting the 1080i on the channels I listed, based on the signal it is receiving - ie in real time.

Alun wrote: »

As far as Sky / Freesat / FTA goes, all the BBC / ITV / CH4 HD channels on Astra 28.2E are 1920x1080i. Although I understand the BBC are actually using an adpative system that switches between progressive and interlaced on the fly as needed to optimize bandwidth.

Adaptive rate is legacy - this thread is about planning for the future infrastructure bandwidth requirements.

Unfortunately the British have this "make the assets sweat" bug in their genes - which leaves the country with antiquated infrastructure - eg trains and 100 year old metro systems, crappy motorways with downmarket burger stops, air traffic control systems that leave aircraft circle for ages, rather than computing just in time approach speeds, traffic lights that create chaos rather than priority based systems that give just in time priority to public transport etc. I'm saying this with sadness rather than criticality.

You won't find this on Continental Europe or even less so in those parts of Asia that have taken the bull by the horns. In the Anglo Saxon media they get grouped under the "developing" category. Perhaps the same media should put AngloSaxonland in a "no longer developing" category.

Alun wrote: »

Irish DTT HD channels are all 1440x1080i, although the vast majority of the content is currently upscaled SD.

Ditto

Alun wrote: »

BTW Don't go by what you see when looking at TV's in shops in Ireland. They're often badly set up and fed with signals of dubious origin and quality.

That is their problem. Nothing to stop them from putting up an Astra 1 facing dish and running http://www.arte.tv/fr with the sound switched down, to let people see the real capability of the set. Showing a set with a less than excellent picture is shooting themselves in the foot. [One of many reasons why Irish retailers are shutting down as people move to order online from Amazon etc]

Finally this thread has excessively in my view gone down the TV (worse still the nearing obsolescence TV's streaming based TV model) laneway, when it was intended to open up discussion on the cheapest, fastest, lowest contention rate broadband network with the longest shelf life.

Ireland has thrown going on for €100 billion into the trash can over the past decade or so, mainly due to bad planning and bad management of the money system (a joint fault of the ECB and Central Bank, who totally ignored the need for an intelligently designed proportional money rationing system for each postcode zone/industry/use).

It would be nice to draw a fresh line under this, and move on in planning terms, getting value for money out of every decision. And above all delivering a high quality solution. There is no single word for it in the English language. It German it is called "preiswert". ie acceptably good wine that doesn't cost €2,000 a bottle .... Starting with the overly complex, expensive to maintain proposed postcode system perhaps!

bk

Very interesting thread, however unfortunately there is also a lot of incorrect information here.

First of all, you can't just heap crappy unshielded twisted pair copper, with high quality shielded coax cable. There are very, very different beasts, with vastly different bandwidth capacity.

You also shouldn't compare the cable industry in the US with the cable industry in Europe and Ireland. The US cable industry has been investing very little capital in developing it's network due to greedy stock holders wanting to sweet the assets. The same isn't true here, UPC have been spending an absolute fortune (500 million on just 700,000 homes!) in building a very high quality, fibre deep network.

From what I've seen, the UPC network is very much equivalent to a FTTH network.

FTTH isn't magical either, the most commonly deployed version of FTTH is GPON, which has a total of 2.5Gb/s of bandwidth, shared with 32 homes (or 64 or 128, depending on the network design).

That means most FTTH networks have only 78mb/s guaranteed bandwidth per customer.

DOCSIS 3.1 has a total of 10Gb/s of bandwidth (not 1Gb/s as Impetus says, 1Gb/s is the speed they expect to sell to people). Now yes, this 10Gb/s is shared between multiple houses on the same cable loop, however it could be shared with 120 homes and still have the same amount of guaranteed bandwidth as FTTH GPON.

I believe UPC are already below 120 homes per loop, so they will have no problem matching the speeds of FTTH. Also cable companies constantly adjust their network, reducing the number of homes per loop, etc. to remove congestion and boost speeds.

So in my mind I honestly consider UPC cable to be FTTH.

BTW There will be no such thing as EuroDOCSIS 3.1, DOCSIS 3.1 gets rid of the separate 6MHz and 8MHz channels and uses a totally different approach. So there will be just one standard used world wide.

So no the cable network isn't banjaxed, it has a very long life ahead of it.

Now Eircom with it's crappy twisted pair cable is a different story. With vectorised VDSL2, they really are tapping out the maximum bandwidth possible on this cable at 100mb/s. They might be able to increase this to 200mb/s by using two lines and bonding, but that is pretty much it for copper phone lines.

FTTdp + G.Fast I don't really consider to be a true copper technology, it is really VERY fibre deep and I would consider it to be FTTH and just being a deployment detail. I expect they will interchangeably do FTTH and FTTdp + G.Fast depending on the difficulty to get into a particular house or building.

So yes, I do consider the copper telephone network to being close to being tapped out. And I think Eircom knows this. It is very clear that Eircom have designed their VDSL2 network with upgrades to FTTH in mind. I believe the thinking is to get VDSL2 out there as quickly as possible as a "good enough" for most people to stem the losses to UPC and LTE and to then slowly build out a FTTH network over time as needed.

The question I have, what technology is going to drive uptake of speeds over 100mb/s?

I honestly believe we are coming to a point of the law of diminishing returns.

50mb/s is good enough for most things, including multiple HD streams (about 10mb/s)

Even 4k Ultra HD is going to use streams of only 20 to 25mb/s using h.265, doable today with UPC 200mb/s.

And I'm not sure 4k UHD will take off in the same way as HD TV did. Take a look at this graph:

http://cdn.arstechnica.net/wp-content/uploads/2013/01/viewingdistrance.png

I'd need a 110" monster TV in order to even start benefiting from UHD.

So I'm not sure if there are any technologies on the horizon that will drive demand for 1Gb/s.

That is why Eircom going cheaply to "good enough" 100mb/s VDSL2, while slowly rolling out FTTH from there makes a lot of sense.

BTW I expect Irelands position in that index report to increase significantly over the next year, with the rollout of UPC 200mb/s plus more importantly Eircom VDSL2 (and perhaps ESB FTTH next year). I believe all the old ADSL2+ areas were pulling down the results of good speeds in UPC areas, VDSL2 will help improve our position greatly.

bealtine

bk wrote: »

The question I have, what technology is going to drive uptake of speeds over 100mb/s?

I honestly believe we are coming to a point of the law of diminishing returns.

50mb/s is good enough for most things, including multiple HD streams (about 10mb/s)

+1 for the analysis and agree wholeheartedly, however...

The "it's good enough" discussion is in of itself a whole new topic but my take on it is to remember it was Bill Gates that said 640k was good enough for anybody? Boy how wrong was that forecast my PC has 32Gb and it's a bit stretched at times. (all depends on your usecase)

We simply don't know what is in our future and what our bandwidth requirements will be while 50Mbs may indeed be "good enough" for now it may not be in the future. To my mind this is the bet eircom are going with for now.

Other countries have wrestled this issue and concluded that FTTH is the way forward (and the "good enough" argument applies to GPON too)


I believe that soon enough all copper access nets will be redundant be they cable or twisted pair even the basic cost of copper dictates this as fibre is cheaper to deploy and manufacture.

Nelbert

Impetus wrote: »

1080i high bitrate:

http://www.tf1.fr
http://www.france2.fr
http://www.m6.fr
http://www.arte.tv/fr
http://www.arte.tv/de
http://www.servustv.com
http://www.zdf.de

etc etc - data taken from a set top box I have.

Sky is 720p and heavily contented - I don't have that tabloid bouquet myself, but if one looks closely at the picture in a shop selling TVs in Ireland, the supposed "HD" is far fuzzier than the picture I observe at the same distance from the above example list of channels.

Sky is 1080i. "Contended" on a satellite platform? Number of subscribers doesn't affect bit rate!

Alun

Impetus wrote: »

Rubbish! What is your source. My set top box is reporting the 1080i on the channels I listed, based on the signal it is receiving - ie in real time.

I use a home made HTPC running MediaPortal, and can view the transmitted resolution and frame rate in real time on the screen for any satellite or DTT channel.

It also records as unadulterated transport streams (.ts files), and I can see the resolution, frame rate and whether it's progressive or interlaced from that too using MediaInfo.

That's my source.

See here too for ARD/ZDF at least ... http://www.digitalfernsehen.de/Warum-ARD-und-ZDF-fuer-HDTV-weiterhin-auf-720p-setzen.106588.0.html

crawler

bealtine wrote: »

The only thing LTE will kill is your pocket...and perhaps the networks that supply LTE. As a shared resource between subscribers it's hard to see how LTE will kill off anything:)
Once the voice cross subsidization disappears then the real price of LTE will become apparent...

Fibre will ultimately kill copper though that's for sure and may the day come soon

I said fibre AND LTE, not Fibre OR LTE

crawler

I agree with a lot of what Bealtine says....

Few things I would comment on...while VDSL is OK, unlike true broadcast networks of cable and FTTH (Usually using RF overlay as Verizon do) , VDSL has to give up substantial bandwidth for IPTV, anyone with eVision (If there is anyone!!) will know that they had to move from a high speed profile to a stable profile to get the service. Vectoring will improve this but only to some degree.

The real challenge comes with multiple devices streaming video at the same time - as screens become higher res (Retina display, 4K etc) then this really eats into the 100Mbps service.

Coupled with the growth of "screenagers" using multiple devices at the same time....and the onslaught of the Internet of Things (IoT) it really does get eaten.

Fibre is the best delivery method for fixed networks and (currently) LTE is for wireless/mobility. GPON is old hat and standard GPON is really just entry level at this point - the technology has evolved.

bk

bealtine wrote: »

The "it's good enough" discussion is in of itself a whole new topic but my take on it is to remember it was Bill Gates that said 640k was good enough for anybody? Boy how wrong was that forecast my PC has 32Gb and it's a bit stretched at times. (all depends on your usecase)

First of all, Bill Gates never actually said that ,it is an urban myth.

Second, I agree, I short of cringe when I hear myself saying this. I remember being in meetings with Eircom, Department of Coms, etc. 12 years ago as a member of IrelandOffLine, telling them all the applications that would require high speed broadband, IPTV, VoD, HD, game downloads, software updates, cloud services, etc. And yes I remember people from Eircom, etc. laughing and saying that no, there would be no demand for such services.

And now here we are and obviously everything I predicted has come to pass.

However despite being a computer scientist and a bit of a futurist who is on the cutting edge of technology, unlike 12 years ago, I just don't see any technologies coming in the short to medium term to will drive demand for 1Gb/s services.

Of course something absolutely revolutionary might come along, like direct mind interaction (think Total Recall), which requires massive bandwidth. But other then some revolutionary breakthrough like that, I just don't see any applications that will require big bandwidth.

Ultra HD is the only one and I'm dubious of how much demand there will be for it and even then, it's bandwidth demands aren't that great. Then can be meet by existing services.

I fear we may have come to a point of "good enough" technology and bandwidth to deliver it for the foreseeable future. Much in the same way that MP3's were found to be good enough and superior SuperCD and DVD Audio never took off or to a lesser extent, DVD was considered good enough and Bluray never really replaced it.

crawler wrote: »

Fibre is the best delivery method for fixed networks and (currently) LTE is for wireless/mobility. GPON is old hat and standard GPON is really just entry level at this point - the technology has evolved.

GPON might be old hat, but it is what most FTTH operators are deploying. It is what Eircom used in their trial FTTH deployments and it is what the ESB are rumoured to be using for their FTTH deployments.

Also replacements for GPON, like 10GPON are experimental and not standardised yet, which is why most companies continue to deploy GPON.

Don't get me wrong, I totally agree that the long term future is fibre, eventually, every single wired connection will be fibre. However I'm not sure it will happen as quickly as some people seem to think.

BTW Internet of Things devices typically use very little bandwidth, being a geek, I've a number of such devices, wifi controlled thermostat, alarm system, lights, etc. and they use very little data, minuscule amounts for command and control.

It is really only video (and maybe game downloads) that drives bandwidth usage.

SpaceTime

Bear in mind that the impact of FTTC and faster cable speeds takes a while to filter through to the stats too as they take a while for people to adopt them.

Ireland's stats will see a marked improvement in the next lot, particularly as more and more FTTC comes on stream and when the impact of UPC's 200mbit/s service is seen.

To be fair to UPC, any speed testing I've done on their network gets the advertised speed or above it.
There's a hell of a lot of fibre in their network.

UPC's HFC network is very very fast though. It's vastly better than what I've used on Comcast and other US networks.

eircom's FTTC rollout is a massive improvement but they're going to have to start moving towards FTTH.
From what I've heard the cabinets they've rolled out have sufficient fibre installed to allow for FTTH rollout when they feel the market's driving that way. They'll only spend money when they are forced to though so, I'd guess if UPC makes another few speed jumps, eircom will have to invest.

SpaceTime

bealtine wrote: »

The only thing LTE will kill is your pocket...and perhaps the networks that supply LTE. As a shared resource between subscribers it's hard to see how LTE will kill off anything:)
Once the voice cross subsidization disappears then the real price of LTE will become apparent...

Fibre will ultimately kill copper though that's for sure and may the day come soon

I wouldn't get too hung up on copper vs fibre. The thing that matters is capacity and speed.

The cable network in particular is becoming much, much higher capacity by bring fibre deep into the network but not necessarily to the end user directly. The coax last link gives you huge capacity and it can actually deliver speeds much closer to a raw fibre connection without the capital expenditure being too high.

Ireland's got the lowest density housing in Europe, so this stuff is going to be expensive and anything that gives us speed without huge capital expenditure is making ultra fast broadband possible here.

Fibre to home is proving too expensive even in countries like France where France Telecom / Orange has started to now invest in FTTC too as they underestimated the cost of rolling out FTTH to every home and France has a lot of low density suburban developments as well as some high density city cores.

Fibre-connected fixed-wireless-access networks are probably what will bring ultra fast broadband to rural areas here. It's unlikely you're going to see fibre to home in those areas to be perfectly honest.
We need to see some sensible allocation of radio bandwidth and proper infrastructure with fibre to the mast available for this sort of thing to really start to work.

There's nothing wrong with using LTE for that, provided that there's sufficient bandwidth allocated. Current mobile allocations wouldn't cut the mustard at all.