Home
 » ISP News, Key Developments » 
Sponsored Links

The UK Gov’s Technical Definition of Gigabit Capable Broadband

Friday, Aug 13th, 2021 (8:31 am) - Score 8,800
Gigabit-Broadband-Speed-Sign-in-Black

One aspect of the UK Government’s new £5bn Project Gigabit programme that we haven’t touched on much is the “technical definition” for how they actually define such broadband ISP products, which is partly because this has been in a state of flux as the project evolved. But it now looks much closer to final.

Just to recap. Project Gigabit seeks to extend “gigabit-capable broadband” speeds to reach at least 85% of UK premises by the end of 2025 (currently 42%), and they also aim to get “as close to 100% as possible“- depending upon how the industry responds. Some 80% of coverage is expected to be achieved by commercial rollouts (reaching c.60% by the end of 2021), which leaves the new project to focus on the final 20% of rural and semi-rural areas (5-6 million premises).

The project is technology neutral, which means it can be delivered via any gigabit-capable network technology, such as Fibre-to-the-Premises (FTTP – the preferred method), Hybrid Fibre Coax (DOCSIS 3.1+) and wireless connectivity (i.e. Fixed Wireless Access or 5G etc.). But all of these must still meet specific technical criteria in order to be eligible for gap funding under the procurement programme.

The Building Digital UK team this week published their latest Open Market Review (OMR) for the recently announced Phase 2 of Project Gigabit, which thus seems like a fitting moment to see how the technical definition has changed since the first project phase was announced back in March 2021.

NOTE: OMRs are designed to identify where gigabit-capable broadband currently exists or is planned to be built within the next 3 years (i.e. defining where public money is needed to address market failure).

Technical Definition of Gigabit-Capable (August 2021)

Infrastructure that can support gigabit-capable downstream services directly or via third-party providers without restriction, as set out below:

a) wholesale infrastructure capable of supporting gigabit-capable downstream wholesale and retail services directly or via third-party providers without restriction, as set out below.

b) connections that are gigabit-capable (capable of delivering 1000Mbps or more download speeds) at the time of delivery of the connection without the need for future hardware upgrades or modification [this principally applies to CPE but also backhaul/other network upgrades] i.e. gigabit capability to be available from day one and if the consumer takes a slower speed it must be soft upgradeable without undue delay. At least one wholesale product with a download speed capability of at least 1Gbps and an upload speed of at least 200Mbps with busy hour performance as laid out below.

c) products with a clear and comprehensible explanation of the minimum, normally available [see note at bottom] and maximum advertised download and upload speeds [as defined by Ofcom].

d) low data latency in accordance with Ofcom regime or codes of practice, recent industry norms and Industry Standards for the requirements of real-time services (or otherwise, in the absence of prevailing standards 10 ms and below for ninety-five per cent (95%) of the time).

e) supports real-time services (e.g. voice/video calling, telematics, telemedicine etc.) with performance indicators (e.g. jitter, packet loss etc.) in line with recent industry standards, norms or in the absence of industry standards, i.e. 2ms for jitter, zero point one percent (0.1%) for packet loss for ninety-five per cent (95%) of the time.

f) actual data speeds and performance during the busiest hours of the day (not more than 4 out of every 24), that do not degrade to less than 33% of the headline download speed and an upload speed equivalent to 20% of the minimum download speed and providers’ service specifications (note: for performance where lower values are better, such as latency, jitter and packet loss, then a factor of 100% above would apply instead).

g) actual data speeds and performance that do not degrade outside of the busiest hours below 95% of the higher of the download and upload speeds set out above. In the event that such minimum download speeds are regulated by a new minimum set by the Regulator, such new minimum download speeds shall apply for the duration of this Agreement. On becoming aware of the new minimum download speeds, the Supplier shall promptly provide confirmation of such to the Authority for its written approval.

h) actual data speeds and performance that do not degrade as take-up of services approaches 100% of the addressable market (including any part arising from switch-off of legacy networks), to be demonstrated by firm commercial and technical (including capacity upgrade) plans to be submitted at the contract bid stage and to be based upon forecasts of up to 7 years.

i) where service offerings and performance vary by locality e.g. as a result of subscribers’ distances from infrastructure, gigabit capability to be maintained for all potential customers.

j) order fulfilment and rectification within typical industry timescales, supported by demonstrably efficient wholesale service management processes;

k) maintenance of customer service levels and network availability in line with industry norms, ideally supported by Service Level Agreements;

l) service provision that does not unfairly discriminate against particular types of services, providers, subscribers or third parties (e.g. via traffic shaping or Quality of Service measures).

m) offering of wholesale access products on open and non-discriminatory terms in line with the principle of technological neutrality, to enable the interconnection to the subsidised network of any technology which other communications providers and/or retail providers may reasonably consider appropriate in accordance with the wholesale access requirements.

NOTE: Normally available and minimum download speeds may include the usual framing and packet overheads of the technologies used, provided that they amount to no more than a few percent of the total traffic i.e. data speed is defined as (user data traffic + overheads) / time.

The main changes seem to be that, firstly, the wholesale product requirements have now been much more strictly defined by performance (i.e. they must support at least 1Gbps download and 200Mbps upload). Secondly, the old definition also made mention of a requirement for “products with 100 Mbps download speed as a minimum,” but this has now been removed in order to focus on true gigabit speed services.

The old definition also called for “upload speeds in line with industry norms for corresponding download speeds (e.g. typically 20 Mbps and above for 100 Mbps download services),” but they’ve removed that in favour of a more prescriptive technical requirement (e.g. “upload speed equivalent to 20% of the minimum download speed“). We also get some specific jitter and packet loss requirements, which was only loosely referenced before.

On the whole, the current definition looks much more solid than what they started with, although it’s worth noting that BDUK said they “will review the criteria for gigabit-capable networks within three months of the launch of the Dynamic Purchasing System, based upon the consultation with industry, and the UK regulator, around the criteria [above].” So further changes cannot be ruled out.

The Government’s definition has to straddle a difficult line. On the one hand, they only want to fund truly gigabit-capable services, but on the other they have to avoid making the requirements so strict that they end up effectively mandating that every home receive the business equivalent of an expensive leased line (i.e. they have to make allowances for the fact that, in order to be affordable, consumer broadband connections need to share their capacity between many users and so speeds may fluctuate). Broadly, they seem to have found a fair balance, but that’s just our opinion.

Share with Twitter
Share with Linkedin
Share with Facebook
Share with Reddit
Share with Pinterest
Mark-Jackson
By Mark Jackson
Mark is a professional technology writer, IT consultant and computer engineer from Dorset (England), he also founded ISPreview in 1999 and enjoys analysing the latest telecoms and broadband developments. Find me on X (Twitter), Mastodon, Facebook and .
Search ISP News
Search ISP Listings
Search ISP Reviews
Comments
55 Responses
  1. Avatar photo Phil says:

    Interesting – is this “gigabit-capable broadband” only for defining funding or the Government’s gigabit status/roll out? if the later would this now exclude Virgin Media – even though the 1Gb service using DOCSIS 3.1 download but the only the poor 52Mbps upload falling way short the 20% (200Mbps) for this new definition?
    Originally VM was excluded from the stats as it was defined as FTTP (before come tech neutral).

    1. Mark-Jackson Mark Jackson says:

      Once DOCSIS 3.1 is fully applied, then it can deliver faster upload and download speeds than those currently sold via Virgin Media’s retail service. A 200Mbps upload speed is perfectly possible, and I suspect if they have a punt for funding then it would be under a new FTTP model, most likely via a new wholesale branch (i.e. not the same as their retail products).

    2. Avatar photo CarlT says:

      200 Mbit upload isn’t going to be possible without some pretty extensive network and also some in-home installation upgrades.

      The existing plant is largely too noisy and frequency constrained, and 3.1 would have to coexist with 4-6 legacy channels.

      Can’t wring enough capacity out of it to deliver 200 to a single customer. 100 is probably about the limit.

    3. Avatar photo Winston Smith says:

      The VM Thatcham trial managed 2.2Gps down and 214Mps up over DOCSIS 3.1, I believe.

    4. Avatar photo CarlT says:

      Yes. Trials can manage a lot of things production can’t, not least because in production you have to share the capacity.

      500 Mbit is available via 3.1. However that requires extensive network upgrade a company planning to overbuild with FTTP aren’t going to do.

    5. Avatar photo Roger_Gooner says:

      VM is predominantly an urban ISP, so building out its FTTP network to deliver an upload speed of 200Mbps in rural and semi-rural areas could be costly. My understanding is that vouchers under the GBVS will be £3,500 for businesses and £1,500 for homes, and time will tell if that’s enough to persuade VM to make the necessary investment.

    6. Avatar photo kuro68k says:

      Even 200 up is a joke. Japan has 10,000 symmetrical in all major cities as a baseline now.

  2. Avatar photo Winston Smith says:

    So currently BT FTTP doesn’t meet the download requirement and VMO2 doesn’t meet the upload requirement or the latency (assuming they average mean ping time for some local server).

    That just leaves a handful of altnets.

    1. Avatar photo Winston Smith says:

      LEO satellite broadband will also struggle to ever meet the latency requirement even if they manage to reach the requisite speeds.

      Do Ofcom have a precise definition of latency?

    2. Avatar photo CarlT says:

      Have another look. They include overheads so your standard ‘900’ products are okay, and the 200 upstream requirement is, coincidentally, what Openreach have on the business variant of their gigabit product.

      Purely coincidental, of course.

    3. Avatar photo NE555 says:

      Openreach (and BT wholesale) *do* sell a 1000/220 FTTP product.

      It’s ludicrously expensive and hardly anyone resells it, not even BT retail and BT business as far as I can tell.

      You can get it from Cerberus for £495 installation and £160 per month (+VAT) on a 12-month contract.

      The requirements for wholesale and for 200M upload seem to rule out Virgin, which will make the Gigabit target much harder to deliver (but more meaningful)

    4. Avatar photo CarlT says:

      Yes, the 220 up is intentionally priced very highly to deter purchase. Openreach have a concern if they priced it too low businesses would abandon profitable, dedicated circuits.

      I’m not sure it’s such a great idea making it cheaper to purchase 2 * 1000/110 though. Requires 2 fibres, 2 CBT connections, etc.

    5. Avatar photo GNewton says:

      @CarlT: Bonding multiple lines together, whether fibre or VDSL, makes perfect sense for many users, especially when you need appropriate upload speeds. This can be a realistic solution for e.g. small businesses.

    6. Avatar photo CarlT says:

      I’m aware, GNewton, it’s what I do for a living. I have 3 circuits at home so eat the dog food too.

  3. Avatar photo Peter says:

    I understand Openreach to sell 200mbps upload, but this is very expensive when you find a supplier

    1. Avatar photo kuro68k says:

      OpenReach holding everyone back. They should be offering 1000 symmetrical.

    2. Avatar photo Peter says:

      Probably because there isn’t a large enough network to compete with them for it to be worth while. Small AltNets here and there isn’t much competition for a large national network.

  4. Avatar photo Tim says:

    Wait, what? No one can provide 1000Mbps unless the CPE has 2.5Gbps Ethernet. 1Gbps Ethernet maxes out at ~940Mbps.

    1. Avatar photo Samuel says:

      “data speed is defined as (user data traffic + overheads) / time.”

      The reason Ethernet maxes out at around 940Mbps is because of overheads, which is included in the calculation.

  5. Avatar photo R Walker says:

    Don’t be so awkward by putting the truth in the way

    1. Avatar photo Kenneth says:

      If anyone really wanted faster than 1 gbit, adding a £25 2.5gbe ethernet card would be no big deal. If you check nearly all new motherboard over £120 support it as standard. Two of my pcs already have it. Its no big deal to install, and anyone that really wanted that speed would see it as a minor obstacle to get over 1Gbit.

  6. Avatar photo Meadmodj says:

    If this is the definition for the BDUK New Procurement then how are the national figures and the Government 85% target to be calculated?. Will all providers offering residential “Giga” headline speed products be asked to sign up to meet paragraphs e, f, and g or specifically confirm where they do ?

    If some of the existing and currently proposed rollout plans (technology neutral) can not meet this my view is that they should be excluded from the figures.

  7. Avatar photo Jack says:

    So Zzoomm is on point with it’s 2000Mbps / 200Mbps upload (symmetric available at £10 extra to give you 2000Mbps upload)

    Shame they aren’t widely available but are expanding

    1. Avatar photo bob says:

      @Jack
      “So Zzoomm is on point with it’s 2000Mbps / 200Mbps upload (symmetric available at £10 extra to give you 2000Mbps upload)”

      200Mbps would only be 10% of the download so would it qualify?

    2. Avatar photo Meadmodj says:

      and the other criteria. ZZOOM are using XPSPON but they would still need to clarify they can sustain them as their network is taken up. As all providers in my view will need to do for each area if the Gov/Ofcom/BDUK are going to use them in their stats.

  8. Avatar photo Matt says:

    I’m on 900mbit FTTP with Zen/Openreach. My speedtests usually have a 17ms ping. Does that mean it wouldn’t meet the 10ms specified in section D or am I reading that incorrectly? What are they using to measure the latency against exactly?

    1. Avatar photo CarlT says:

      They’ll be using the delay between home and the equipment terminating the connection most likely.

      The speed of light through fibre makes anything else problematic. If you’re in Scotland and connect via an ISP in London you’re not going to make that round trip in 10 ms.

    2. Avatar photo 125us says:

      Your ping is RTD. You have to halve it for one way latency. Without knowing what you’re pinging and where it is the number is meaningless though.

    3. Avatar photo Sam P says:

      If you’ve ruled out the router and plugged directly into your PC’s ethernet you should try cleaning the end of your fibre optic cable going into the openreach ONT.

    4. Avatar photo CarlT says:

      Dirty fibre will cause packet loss more than higher latency.

      Chances are just that OP is in Scotland and/or reaching London via Zen’s Manchester POP.

  9. Avatar photo 5G Infinity says:

    For clarity, OR’s 900Mbps/120Mbps delivered by BT and ISPs using BT Wholesale xlates to a 1Gbps download, but 120Mbps up does not xlate to 200Mbps including overhead, more like 150Mbps.

    Current ping on OR is 4ms, if it’s higher than 7ms there could be a problem at the xchg.

    Mark – will OR need to tweak it’s upload nationally to meet this criteria?

    1. Avatar photo CarlT says:

      There’s already a 220 up Openreach FTTP product so they’re fine.

      Yes, it’s insanely expensive compared with the 1000/115 at £80 a month versus less than £32 but it’s there.

      https://www.openreach.co.uk/orpg/home/products/pricing/loadProductPriceDetails.do?data=M80QNeH46o4g6JKGD604vTypQOKfNn%2Beo6vmoVhAOBZZ6rNZujnCs99NbIKJZPD9hXYmiijxH6wrCQm97GZMyQ%3D%3D

      These criteria are highly likely to have been influenced by Openreach’s product set. They are by a mile the biggest FTTP network and will probably deliver a bunch of the government procurements.

    2. Avatar photo GNewton says:

      @CarlT: The price tables don’t make sense to me. E.g. it lists

      Standard Connection:
      Up to 40Mbit/s / 10Mbit/s,
      Up to 55Mbit/s / 10Mbit/s,
      Up to 80Mbit/s / 20Mbit/s,
      Up to 110Mbit/s / 15Mbit/s,
      Up to 115Mbit/s / 20Mbit/s,
      Up to 160Mbit/s / 30Mbit/s,
      Up to 220Mbit/s / 20Mbit/s,
      Up to 220Mbit/s / 30Mbit/s,
      Up to 330Mbit/s / 30Mbit/s,
      Up to 330Mbit/s / 50Mbit/s,
      Up to 550Mbit/s /75Mbit/s,
      Up to 1000Mbit/s /115Mbit/s
      for £99.39

      Standard Connection:
      Up to 500Mbit/s /165Mbit/s,
      Up to 1000Mbit/s /220Mbit/s
      for £500.00

      Wouldn’t you expect different prices for each speed range?

    3. Avatar photo John says:

      Those are the installation/connection prices.
      Why do you expect them to be different for each speed tier?

      The annual rental is listed on the linked price list and each speed tier is a different price.
      That’s what I would expect.

      The reason the bottom 2 tiers have higher connection fees is because they are business tiers.

      The additional connection charge is banked and can be used for any additional upgrades that might be required on the PON to support those higher tiers (like XGS-PON, though unlikely to be needed for the vast majority of PON’s).

    4. Avatar photo CarlT says:

      I’d actually expect all of those to be the same. The higher installation charge further deters customers from purchasing.

  10. Avatar photo Buggerlugz says:

    Its all well and good the government harping on about gigabit availability, but unfortunately the government isn’t going to make gigabit broadband anywhere near affordable to the customer.

    Looking through the governments rose tinted specs, everyone will have gigabit capable internet by the end2025, sadly no one will be able to afford it though.

    1. Avatar photo NE555 says:

      What’s your definition of “no one”? The current cost of £50-£60 per month may not be affordable for some but it is for many, and the price is only going downwards from there. Typical households spend twice that on gas and electricity, three times that on council tax, and five times that on running their car.

      I’d be more sympathetic to the argument that no one *needs* a gigabit connection at home. 330M is more than enough for most offices, let alone a single home. However it’s really about future-proofing: once there’s fibre in the ground it can run at gigabit, 10 gig or faster as the future requires.

      Getting rid of rate-adaptive copper also gets rid of the “up to” speeds. In future if you buy an 80M service, you’ll get 80M – not something that varies depending on where you live or even the weather outside.

    2. Avatar photo CarlT says:

      People need paying, and that comes out of broadband bills. Networks need paying for, that comes out of broadband bills. Construction is relatively expensive and high quality here.

      Cheap / good / fast. As has been the case for years pick 2, and don’t whinge if you don’t get all three as the third is a bonus.

      Places where it’s way cheaper usually have hugely more variable speeds which would inevitably result in people posting on here complaining.

      As with everything else buy what you can afford.

    3. Avatar photo Adrian says:

      I was paying £40 per month for the old ADSL product almost 20 years ago with the blue saucer USB modem for 1meg.

      The problem with the UK has been a race to the bottom pricing and companies need to make a return now on upgrades

      Companies can’t sink billions into networks and have customers on sub £20 packages.

      If people want gigabit then they will have to pay for it and given the current price I think for 1000 times the speed from 20 years ago for a similar price is a bargain.

    4. Avatar photo Seagull Lover says:

      Cityfibre based 1 Gig services are certainly far cheaper than Openreach ones. For example Zen offer 1000 down/1000 up Mbps for £40/m over Cityfibre network and TalkTalk will also offer something similar once they release their Future Fibre 900 service this Autumn. Of course the Cityfibre FTTP coverage is less than Openreach FTTP coverage but nevertheless its increasing rapidly. They’re easily the biggest Altnet.

  11. Avatar photo Superfast20 says:

    It doesn’t really matter what the definition is, because unless ISPs and network providers are forced to abide by it, it’s a waste of time and effort discussing it. We had definitions of 24mbps and then 30mbps for superfast. ISPs use super-fast to describe its products for anything that is FTTC, regardless of speeds. Openreach are just the same, the checker on their website defines superfast as anything that it connected to an FTTC enabled cabinet. You may as well label the whole network as gigabit capable now, as it is technically- just subject to an upgrade. Same as lines now advertised as being capable of receiving superfast, but require FTTP on demand or community fibre project in order to actually receive such speeds.

    1. Avatar photo CarlT says:

      While a nice rant all the technology that can fulfil the criteria is not rate adaptive so the previous issues where, despite the median customer reaching superfast a minority don’t due to line length, are not an issue.

    2. Avatar photo Superfast20 says:

      @carlt – imagine if they applied broadband principles to everything else?

      100 people order and pay for a car but only 75% get one? I’m sure you would happily tell them to keep your money and accept a push bike instead?

      100 people buy and pay for a pint, but only 80% get a pint , the others only getting water. I’m sure you would be happy with paying £5 for a pint of water whilst 80% enjoying their beer.

    3. Avatar photo CJ says:

      I guess you’re not aware of Ofcom’s Broadband Speeds Code of Practice? The ISP doesn’t hold you to the contract if they can’t deliver at least the minimum speed guaranteed in their pre-sales estimate. Which is better than how the car industry deals with diesel cars that don’t meet the claimed emission standards, they’re the ones who keep your money regardless.

      Like broadband speeds, the amount of beer you can get for a certain price also varies by location. When I visit London, I don’t get a full pint of beer for my £5 like I do in most other places. If getting a full pint of beer for £5 is important, you don’t choose to live in London.

    4. Avatar photo CarlT says:

      @Superfast20 None of it is relevant to this article and the criteria within it so not that bothered.

    5. Avatar photo GNewton says:

      “but require FTTP on demand”

      Actually, FTTPoD is pretty much a dead product now.

  12. Avatar photo spotify95 says:

    Guess these definitions are clearly there to rule out Virgin Media products. Their upload speeds are truly horrific: 52Mbps for a 1Gbps download is just poor. Their other packages aren’t good either, 10Mbps on a 100Mbps package (and our legacy products we used to have were 70/5 and 30/2). The latter is so poor that 3G could outperform the upload speed, let alone 4G.
    Hopefully true gigabit internet will soon be available thanks to FTTP rollout. I’d also prefer something like what Hyperoptic offer: symmetrical services on all but the base product. That would allow 1Gb/1Gb speeds, which would be definitely worth paying for.

  13. Avatar photo Adam Jarvis says:

    It’s only taken 20 years to come with a technical definition for gigabit capable broadband, and we’re not there yet.

    Given mobile is being used as part of the 10Mbps USO, how about a technical definition to define a 4G/5G mast in terms of its back haul, along with the number of concurrent connections/sustained transfer speeds from various set waypoints from each mast. Maintenance programmes should also be part of this standard, to make sure foliage is cut back in summer. This should be mandatory for any new mast, and should be applied retrospectively when any upgrade takes place.

    4G/5G is not some ubiquitous imaginary blanket that float above in the sky that the mobile operator’s marketing men want us to believe, it’s a very real connection to a nearby mast (or lack of mast, rurally) with tree/foliage growing and blocking its path in summer.

  14. Avatar photo zzing says:

    Everything seems to be a bit arbitrarily defined without any basis in reality nor any stretch goal to make ISPs uncomfortable and actually deploy technology that matters.

    Latency doesn’t seem to take the worst case scenario (Scilly Isles to Shetland = 1183km using the slowest 1625nm light travelling at 203.940 m/µs) comes to 5.8ms. Give a generous 25% waste for numpty infrastructure overhead and 7.25ms would be a suitable maximum.

    Asymmetry at 20% upload is also arbitrary – who and how was 20% defined as even remotely acceptable? Even the useless GPON is 1:2, but the more useful XGS-PON is 1:1 – why is there any asymmetry defined at all?

    And contention effects allowing for 66%, 0.1% packet loss, 2% jitter at 95% what is an ‘industry norm’? This is woefully low, even for broadband – it should be 99% and packet loss and jitter a magnitude too high. Leased lines can be 4-5 9’s and a further magnitude lower.

    Anyway, my view is the PHY should be defined as any technology capable of providing any service. Whether an ISP decides to provide a leased line service or a broadband service, the PHY should be the same. A leased line simply reserves more of a PON compared to broadband. Similarly if I choose to pay ECCs to lay a dedicated line, why can’t I just slap a broadband service on it, instead of this turkey that is FTTPoD?

    1. Avatar photo CarlT says:

      Regarding latency fibre doesn’t run in straight lines between everything it follows population centres same as the motorway network.

      Include the ‘numpty infrastructure’ on top and you’ve 50%+ delay on top of the theoretical.

      The latency will likely be round trip, too, not one-way, due to that many networks can be asymmetrical.

      I imagine that’s the maximum permitted access network round trip given the bids are by network builders, not ISPs, and delays upstream and downstream can be different depending on the access network architecture.

    2. Avatar photo CarlT says:

      ‘Similarly if I choose to pay ECCs to lay a dedicated line, why can’t I just slap a broadband service on it, instead of this turkey that is FTTPoD?’

      The other end of the telco’s fibre doesn’t go to the broadband network.

      Worth noting that, depending on the product ordered, leased lines / DIA use broadband as a bearer.

    3. Avatar photo CarlT says:

      ‘Anyway, my view is the PHY should be defined as any technology capable of providing any service. Whether an ISP decides to provide a leased line service or a broadband service, the PHY should be the same.’

      I don’t know of anywhere this is the case. I’m aware of places you can rent dark fibre but it’s all passive, and goes to an ODF then your ISP who decide what they’re lighting it with.

      One physical layer for everything would require point to point fibre and increased costs at the telco as they need a 10 GBase-BX port for every customer, and the customers all 10 Gb CPE. The switch ports and CPE would all have to be upgraded if the telco wanted to sell anything above 10 Gb.

      Dark or grey fibre, point to point or using WDM, seems the way to go. Forcing use of a single transmission layer for all is impractical.

      GPON is hardly hopeless. The ONTs are cheap, no burst laser upstream, and they deliver a more than adequate service for 99% if used appropriately. XGSPON pricing has only relatively recently come down to reasonable levels.

      PON saves a lot of ports for telcos. Force use of point to point and you force infrastructure providers to fundamentally change how they build their networks. Fibre is cheap, the space, hardware and power for 30,000 10GBase-BX switch ports and ODFs for each of a thousand headends isn’t, this ignoring the extensive ductwork required to accommodate the fibre.

      Trying to serve everything via PON would be excruciating. So much bespoke work would be necessary. A 10 Gb order on an XGSPON segment would have to be rejected, require own dedicated fibre or own NGPON wavelengths.

      Much as we would all love 10 Gb, and for infrastructure builders to be pushing the limits, they don’t have to bid on these projects and if requirements are too onerous won’t.

  15. Avatar photo zzing says:

    Forgive my cynicism Carl, but after seeing a copper network that’s around 40 years old, digs are not something to take lightly, and I don’t expect fibres to change once in the ground based on the fact we’ve had modem > ISDN > ADSL > ADSL2 > VDSL2 > GFast that has been extracted from copper.

    Prior to fibre, leased lines were T1/E1, ISDN30, EFM is still used albeit disappearing, but all using the same point-to-point copper infrastructure. The point is the exchanges are already there sized appropriately for point-to-point. Yes it’s considerably more expensive, but it does mean a lot of options are available for future topologies.

    The issue I have basically is what is being stuck in the ground now is likely, due to inertia against digs, be with us for the next 40 years. Of course there will be technological advances, just like DSL came about way after the copper was in the ground, and I do hope that PONs will be able to handle it, as we’re pretty stuffed without it.

    Over time, you’ve convinced me of the value of PONs in general. GPON itself with it’s 2.5/1.25 gbps bandwidth just hasn’t got the capacity to manage a 32-way split with adequate headroom and thus way too much contention to manage a proper gigabit set of connections – hence why I said XGS is a better option to use here, because realistically, how often is that ONT actually going to get changed?

    As for the rest, this doesn’t seem this is putting any pressure on ISPs to meet the service requirements. Latency on VM and TTB is beyond terrible. On BTW it’s OK, but it’s degraded since ADSL2+ days (I may have 60 more mbps download on FTTC, but latency has doubled, even though my modem is on fastpath).

    1. Avatar photo CarlT says:

      Point to point is always best, however PON should have some mileage given WDM.

      If a GPON split runs out of steam XGSPON or PON split are an option for capacity relief. GPON isn’t suitable for high symmetrical gigabit uptake though and telcos know it – at least one of the telcos in the states removes the gigabit option from a PON when they have an existing customer on it.

    2. Avatar photo NGA for all says:

      We never have a proper debate on costs. BSG/AM in 2009 said FTTC would be £5bn, it was closer £3bn with half of that subsidy. FTTP GPON -£30bn,- will be closer to £12bn -re-using some middle mile paths. What is the incremental cost to PTP. Too much subsidised FTTC Cure (re-arranged copper) will now block FTTP – this was only done for the subsidy.

      The construction work to add the hinged fittings to poles, attach
      connection blocks and pulling the cable through does not look excessive. A Contractor team of 2 with a picker seem to be completing 4-5 DPs a day for hinges and connection blocks.

      Demand for >100MBps for those on 30-80Mbps will be a much bigger issue for completing this next stage. The continual gaming of costs means UK has paid a big price so far. What would the PtP costs be?

Comments are closed

Cheap BIG ISPs for 100Mbps+
Community Fibre UK ISP Logo
150Mbps
Gift: None
Virgin Media UK ISP Logo
Virgin Media £24.00
132Mbps
Gift: None
Plusnet UK ISP Logo
Plusnet £27.99
145Mbps
Gift: None
Zen Internet UK ISP Logo
Zen Internet £28.00 - 35.00
100Mbps
Gift: None
Sky Broadband UK ISP Logo
100Mbps
Gift: None
Large Availability | View All
Cheapest ISPs for 100Mbps+
Gigaclear UK ISP Logo
Gigaclear £15.00
150Mbps
Gift: None
YouFibre UK ISP Logo
YouFibre £19.99
150Mbps
Gift: None
Community Fibre UK ISP Logo
150Mbps
Gift: None
BeFibre UK ISP Logo
BeFibre £21.00
150Mbps
Gift: £25 Love2Shop Card
Hey! Broadband UK ISP Logo
150Mbps
Gift: None
Large Availability | View All
The Top 15 Category Tags
  1. FTTP (5443)
  2. BT (3497)
  3. Politics (2513)
  4. Openreach (2285)
  5. Business (2242)
  6. Building Digital UK (2226)
  7. FTTC (2040)
  8. Mobile Broadband (1954)
  9. Statistics (1770)
  10. 4G (1648)
  11. Virgin Media (1603)
  12. Ofcom Regulation (1446)
  13. Wireless Internet (1384)
  14. Fibre Optic (1384)
  15. FTTH (1380)
Promotion
Sponsored

Copyright © 1999 to Present - ISPreview.co.uk - All Rights Reserved - Terms , Privacy and Cookie Policy , Links , Website Rules , Contact
Mastodon