One aspect of the British government’s newly launched £5 billion gigabit plan is how they actually define the “technical definition” of such broadband ISP products. This is partly because this is always in a state of constant change, because Project development. But now it looks closer to the final result.
Just review it. The Gigabit project aims to extend the speed of "Gigabit-capable broadband" to at least 85% of locations in the UK by the end of 2025 (currently 42%), and they also aim to be "as close to 100% as possible"-depending on How the industry responds. It is expected that approximately 80% of the coverage will be achieved through commercial promotion (approximately 60% by the end of 2021), which allows the new project to focus on the last 20% of rural and semi-rural areas (5-6 million premises).
The project is technology-neutral, which means that it can be delivered via any network technology that supports gigabit, such as fiber to the home (FTTP-the preferred method), hybrid fiber coaxial cable (DOCSIS 3.1+) and wireless connection (ie fixed Wireless access or 5G, etc.). But all of these must still meet specific technical standards to qualify for gap funding under the procurement plan.
This week, the Building Digital UK team released their latest Open Market Review (OMR) for the second phase of their recently announced gigabit project, so this seems to be the right time to understand how the technology definition has changed since the announcement of the first project phase. March 2021.
Gigabit Technology Definition (August 2021)
The infrastructure that can support gigabit downstream services directly or through third-party suppliers without restrictions, as shown below:
a) The wholesale infrastructure can directly or through third-party suppliers unrestrictedly support downstream wholesale and retail services with gigabit capabilities, as described below.
b) A connection with Gigabit capability (capable of providing download speeds of 1000Mbps or higher) when the connection is delivered, without future hardware upgrades or modifications [this is mainly applicable to CPE, but also applies to backhaul/other network upgrades] ie Gigabit Capacity is available from day one, and if consumers use a slower speed, it must be soft-upgradable without unreasonable delay. At least one wholesale product with a download speed of at least 1Gbps and an upload speed of at least 200Mbps. The performance during peak hours is shown below.
c) Products with clear and easy-to-understand explanations for the lowest, usually available [see note at the bottom] and maximum advertising download and upload speed [defined by Ofcom].
d) Low data latency required for real-time services required by Ofcom system or code of practice, the latest industry regulations and industry standards (or in the absence of current standards, 10 milliseconds and below is 95% (95%) of the time).
e) Support real-time services (such as voice/video calls, telematics, telemedicine, etc.), and its performance indicators (such as jitter, packet loss, etc.) comply with the latest industry standards, specifications or in the absence of industry standards, that is, 2ms For jitter, the packet loss rate is 0.1% (0.1%) in ninety-five percent (95%) of the time.
f) The actual data speed and performance during the busiest time of the day (not more than 4 hours per 24 hours) will not be degraded to less than 33% of the title download speed, and the upload speed is equivalent to 20% of the minimum download speed and download speed. Provider’s service specification (note: for lower values for better performance, such as delay, jitter, and packet loss, a factor of more than 100% will be applied).
g) The actual data speed and performance will not degrade below 95% of the higher download and upload speed outside of the peak hours. If such minimum download speeds are regulated by new minimum speeds set by regulatory agencies, such new minimum download speeds shall apply during the validity period of this agreement. Upon learning of the new minimum download speed, the supplier should immediately provide confirmation to the authorities for written approval.
h) The actual data speed and performance will not decrease as the use of the service approaches 100% of the addressable market (including any part resulting from the shutdown of the traditional network), as evidenced by the company’s business and technology (including capacity upgrades)) The plan is submitted during the contract bidding stage and is based on a seven-year forecast.
i) In cases where service supply and performance vary from region to region, for example, due to the different distances between users and infrastructure, it is necessary to maintain gigabit capabilities for all potential customers.
j) Complete orders and rectifications within a typical industry time frame, and are supported by an obvious and efficient wholesale service management process;
k) Maintain customer service levels and network availability in accordance with industry regulations, preferably supported by service level agreements;
l) Will not unfairly discriminate against specific types of services, providers, subscribers, or third-party service provision (for example, through traffic shaping or service quality measures).
m) Provide wholesale access products on open and non-discriminatory terms in accordance with the principle of technology neutrality, so that other communication providers and/or retail providers may reasonably consider any technology and subsidized network interconnection wholesale access requirements.
Note: Generally available and minimum download speeds may include the overhead of common frames and packets of the technology used, provided that they do not exceed a few percent of the total traffic, that is, the data speed is defined as (user data traffic + overhead)/time.
The main changes seem to be that, first of all, the requirements for wholesale products are now more strictly defined by performance (that is, they must support at least 1Gbps download and 200Mbps upload). Secondly, the old definition also mentioned the requirement of "products with a minimum download speed of 100 Mbps," but it has now been deleted to focus on true gigabit speed services.
The old definition also required “upload speeds to comply with industry specifications for corresponding download speeds (for example, 100 Mbps download services are usually 20 Mbps and above)”, but they have deleted it to support more standardized technical requirements (for example, “upload The speed is equivalent to 20% of the minimum download speed"). We also got some specific jitter and packet loss requirements, which were only loosely quoted before.
In general, the current definition looks more reliable than the one they started, although it is worth noting that BDUK stated that they “will review the standards for gigabit-capable networks within three months of the launch of the dynamic procurement system, based on The industry and UK regulators negotiate the standard [above]." So further changes cannot be ruled out.
The definition of government must cross a difficult line. On the one hand, they only want to fund services that are truly gigabit capable, but on the other hand, they must avoid making the requirements so strict that they end up effectively forcing every family to obtain an expensive leased line Business (that is, they must take into account the fact that in order to be affordable, consumer broadband connections need to share its capacity among many users, so speeds may fluctuate). In general, they seem to have found a fair balance, but this is just our opinion.
The interesting thing is-this kind of "Gigabit broadband" is only used to define funding or is the government's Gigabit status/launch? If the latter would now exclude Virgin Media-even if the 1Gb service uses DOCSIS 3.1 downloads, but the only bad 52Mbps upload is lower than this newly defined 20% (200Mbps)? Initially VM was excluded from statistics because it was defined as FTTP (before technology neutrality).
After fully applying DOCSIS 3.1, it can provide faster upload and download speeds than currently sold through Virgin Media's retail service. An upload speed of 200Mbps is entirely possible, and I suspect that if they have capital investment, then it will adopt the new FTTP model, most likely through a new wholesale branch (that is, different from their retail products).
Without some fairly extensive networks and some home installation upgrades, 200 Mbit upload is impossible.
Existing equipment is largely too noisy and frequency limited, and 3.1 must coexist with 4-6 traditional channels.
Cannot squeeze enough capacity to deliver 200 to a single customer. 100 is probably the limit.
I believe that the VM Thatcham test reduces 2.2Gps and improves 214Mps compared to DOCSIS 3.1.
Yes. Experimentation can manage many things that production cannot do, especially because you have to share capacity in production.
500 Mbit is available through 3.1. However, this requires a lot of network upgrades, and a company that plans to overuse FTTP would not do this.
VM is mainly an urban ISP, so building its FTTP network to provide 200Mbps upload speed in rural and semi-rural areas may be expensive. My understanding is that the voucher under GBVS will provide businesses with £3,500 and families with £1,500. Time will tell whether this is enough to convince VMs to make the necessary investments.
Even if it goes up by 200, it's a joke. Japan now uses 10,000 symmetrical all major cities as a benchmark.
So currently BT FTTP does not meet the download requirements, and VMO2 does not meet the upload requirements or delays (assuming they average the average ping time of some local servers).
Only some alternative networks remain.
Even if LEO satellite broadband manages to achieve the required speed, they will struggle to meet the delay requirements.
Does Ofcom have an accurate definition of delay?
Look again. They include management fees, so your standard "900" product is ok, and the 200 upstream requirement happens to be Openreach's requirements for the business variant of its gigabit product.
Openreach (and BT wholesalers) *do* sell 1000/220 FTTP products.
It is so expensive that almost no one resells it. As far as I know, even BT retail and BT business are not good.
You can get it from Cerberus for an installation fee of £495 and a 12-month contract of £160 (+ VAT).
The requirements for wholesale and 200M uploads seem to exclude Virgin, which will make delivery of gigabit targets more difficult (but more meaningful)
Yes, 220 up is deliberately priced very high to prevent purchase. Openreach is worried that if the price is too low, companies will give up profitable dedicated circuits.
However, I am not sure if it is a good idea to buy 2 * 1000/110 cheaper. Requires 2 optical fibers, 2 CBT connections, etc.
@CarlT: Connecting multiple lines together, whether it is optical fiber or VDSL, is very meaningful to many users, especially when you need an appropriate upload speed. This may be a realistic solution for, for example, small businesses.
I know, GNewton, this is my way of making a living. I have 3 circuits in my house, so I also eat dog food.
I know Openreach sells 200mbps uploads, but when you find a supplier, it is very expensive
OpenReach hinders everyone. They should provide 1000 symmetry.
It may be because there is not a large enough network to compete with them, so it is worth it. Small AltNets don't have much competition here with large national networks.
Wait, what? Unless CPE has 2.5Gbps Ethernet, no one can provide 1000Mbps. 1Gbps Ethernet can reach up to ~940Mbps.
"Data speed is defined as (user data flow + overhead)/time."
The reason why Ethernet reaches its maximum value around 940Mbps is overhead, which is included in the calculation.
Don't be so embarrassed by keeping the truth out of the door
If someone really wants faster speeds than 1 gbit, then adding a £25 2.5gbe Ethernet card is no big deal. If you check almost all new motherboards over £120, it is supported as a standard configuration. My two computers already have it. Installation is no big deal, anyone who really wants this speed will see it as a small obstacle beyond 1Gbit.
If this is the definition of BDUK’s new procurement, how to calculate the national figure and the government’s 85% target? Will all suppliers that provide residential "Gigabit" title speed products be required to register to meet the requirements of paragraphs e, f, and g, or specifically confirm where they are doing it?
If some of the existing and currently proposed rollout plans (technology neutral) cannot satisfy this, my opinion is that they should be excluded from the numbers.
So Zzoomm is 2000Mbps / 200Mbps upload (symmetrical can pay an extra £10 to provide you with 2000Mbps upload)
Unfortunately they are not widely used but are expanding
@Jack "So Zzoomm is a 2000Mbps / 200Mbps upload (for symmetry, you can pay an extra £10 to provide you with 2000Mbps upload)"
200Mbps only accounts for 10% of downloads, so is it eligible?
And other standards. ZZOOM is using XPSPON, but they still need to clarify that they can maintain them while their network is occupied. If Gov/Ofcom/BDUK wants to use them in their statistics, I think all providers need to do it for each region.
I am using Zen/Openreach's 900mbit FTTP. My speed test usually has a ping of 17 milliseconds. Does this mean it does not meet the 10 milliseconds specified in Part D, or am I reading it wrong? What do they use to accurately measure delay?
They are likely to use the delay between home and device to terminate the connection.
The speed of light passing through the fiber can make any other problem a problem. If you are in Scotland and connected through an ISP in London, you will not complete the round trip within 10 milliseconds.
Your ping is RTD. You must halve it to reduce the delay in one way. I don't know what you are pinging and where it is, although it doesn't make sense.
If you have excluded the router and plugged directly into the Ethernet of the PC, you should try to clean the end of the fiber optic cable that enters the openreach ONT.
Dirty fiber will cause packet loss more than higher latency.
It is very likely that OP will arrive in London in Scotland and/or via Zen's Manchester POP.
For the sake of clarity, the 900Mbps/120Mbps of OR provided by BT and the ISP using BT Wholesale is converted to 1Gbps download speed, but the 120Mbps rise does not translate to 200Mbps, including overhead, which is more like 150Mbps.
The current ping on OR is 4ms, if it is higher than 7ms, there may be a problem with xchg.
Mark-or does it need to adjust the content it uploads across the country to meet this standard?
There is already a 220 up Openreach FTTP product, so they are good.
Yes, it's ridiculously expensive compared to the 1000/115 at 80 pounds a month and the 1000/115 at less than 32 pounds, but it's there.
https://www.openreach.co.uk/orpg/home/products/pricing/loadProductPriceDetails.do?data=M80QNeH46o4g6JKGD604vTypQOKfNn%2Beo6vmoVhAOBZZ6rNZujnCs99NbIKJZPD9hXYmijCwr%3G6D6
These standards are likely to be affected by the Openreach product set. They are the largest FTTP network and may provide a lot of government procurement.
@CarlT: The price list doesn't make sense to me. For example 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, priced at £99.39
Standard connection: up to 500Mbit/s/165Mbit/s, up to 1000Mbit/s/220Mbit/s, £500.00
Don't you want different prices for each speed range?
These are installation/connection prices. Why do you expect them to be different for each speed grade?
The annual rent is listed on the linked price list, and the price is different for each speed class. This is what I expected.
The reason for the higher connection costs of the bottom two layers is because they are the business layer.
The additional connection fee has been deposited in the bank and can be used for any additional upgrades that the PON may require to support those higher layers (such as XGS-PON, although most PONs are unlikely to be required).
I actually hope all of these are the same. Higher installation costs further hinder customers' purchases.
The government has been talking about the availability of gigabit, which is all good, but unfortunately, the government will not make gigabit broadband available anywhere customers can afford it.
Looking at the government’s rose-colored specifications, by the end of 2025, everyone will have a gigabit Internet, but unfortunately no one can afford it.
What is your definition of "no one"? The current monthly fee of £50-60 may not be affordable for some people, but it is affordable for many, and the price will only drop from there. The typical household spends twice as much on gas and electricity, three times as much on municipal taxes, and five times as much on driving.
I even agree with the argument that no one *needs* a gigabit connection at home. 330M is sufficient for most offices, let alone a family. However, this is indeed future-oriented: once there is optical fiber underground, it can run at gigabit, 10 gigabit or faster speeds in accordance with future requirements.
Get rid of the speed of adaptive copper also get rid of the "up to" speed. In the future, if you buy an 80M service, you will get 80M-it will not be different depending on where you live or even the weather outside.
People need to pay, and this comes from broadband bills. There is a charge for the network, which comes from the broadband bill. The buildings here are relatively expensive and of high quality.
Cheap/good/fast. As has been the case for many years, choose 2, if you don't get all three, please don't complain, because the third one is a reward.
Cheaper places usually have more variable speeds, which will inevitably lead to people posting complaints here.
Like everything else, buy what you can afford.
About 20 years ago, I paid 40 pounds a month for my old ADSL product and used a 1meg blue disc-shaped USB modem.
The problem in the UK has always been a race to lower prices, and companies now need to upgrade to reap the rewards
Companies can't put billions of dollars into the network, nor can they let customers buy packages for less than £20.
If people want Gigabit, then they will have to pay for it. Given the current price, I think it's a good deal to buy it at a similar price at 1,000 times the speed it was 20 years ago.
Cityfiber-based 1 Gig services are definitely much cheaper than Openreach services. For example, Zen offers 1000 down/1000 Mbps Mbps over the Cityfiber network at a price of £40/meter, and TalkTalk will provide similar services after launching their Future Fiber 900 service this fall. Of course, the coverage of urban optical fiber FTTP is smaller than that of Openreach FTTP, but it is growing rapidly. They can easily become the biggest Altnet.
It doesn't matter what the definition is, because unless ISPs and network providers are forced to comply with it, discussing it is a waste of time and effort. We have the definition of 24mbps, and then the ultra-high speed of 30mbps. Regardless of the speed, ISPs use any FTTC that describes their products very quickly. Openreach is the same. The checker on their website defines ultra-high speed as anything connected to an FTTC-enabled cabinet. You can now also mark the entire network as a gigabit network because it is technically-it just needs to be upgraded. It is the same as the lines now advertised as being able to receive ultra-high speeds, but requires on-demand FTTP or community fiber projects to actually receive such speeds.
Although all the technologies that can meet the standard are not rate-adaptive, the previous problem is that although the median customer reaches ultra-fast, a small number of people are not due to the length of the line. This is not a problem.
@carlt-imagine if they apply broadband principles to everything else?
100 people order and pay for a car, but only 75% get one? I'm sure you will be happy to tell them to keep your money and accept a trolley?
100 people buy and pay for a pint, but only 80% get a pint and the others only get water. I believe you will be happy to spend £5 on a pint of dehydrated, while 80% of people enjoy their beer.
I guess you don’t know Ofcom’s broadband speed practice guidelines? If the ISP cannot provide at least the minimum speed guaranteed in its pre-sale estimate, the ISP will not let you comply with the contract. This is better than the auto industry's way of dealing with diesel vehicles that don't meet the claimed emission standards, they will keep your money anyway.
As with broadband speed, the amount of beer you can get at a specific price varies from location to location. When I visit London, I don’t get a pint of beer for 5 pounds like in most other places. If getting a pint of beer for £5 is important, then you would not choose to live in London.
@Superfast20 None of them are related to this article and the standards in it, so don't worry.
"But on-demand FTTP is required"
In fact, FTTPoD is now almost a dead product.
Guess these definitions are clearly meant to exclude Virgin Media products. Their upload speed is really terrible: 52Mbps at 1Gbps download speed is really bad. Their other packages are not good, 10Mbps on the 100Mbps package (the traditional products we had in the past were 70/5 and 30/2). The latter is too bad, the upload speed of 3G is not as good as 4G. Hopefully, thanks to the launch of FTTP, real Gigabit Internet will be available soon. I also prefer the services provided by Hyperoptic: symmetrical services for all products except the basic products. This will allow 1Gb/1Gb speeds, which is definitely worth paying for.
It took only 20 years to develop a technical definition for gigabit-capable broadband, and we have not yet achieved this.
Given that mobile is used as part of the 10Mbps USO, how to define the technical definition of the 4G/5G mast in terms of backhaul, and the number of concurrent connections/continuous transmission speed from various set waypoints of each mast. Maintenance plans should also be part of this standard to ensure that leaves are reduced during the summer. This should be mandatory for any new mast and should be applied retrospectively when any upgrade occurs.
4G/5G is not the kind of ubiquitous imaginary blanket floating in the sky that the marketers of mobile operators want us to believe. It is a very real connection with nearby masts (or lack of masts, rural areas), trees/leaves Growth and blocking its path in summer.
Everything seems to be defined arbitrarily, without any realistic basis, and no extended goal to make ISPs feel uncomfortable and actually deploy important technologies.
The delay does not seem to be the worst case (Silly Islands to Shetland Islands = 1183 kilometers, using the slowest 1625 nanometer light propagating at 203.940 m/µs) reaching 5.8 milliseconds. To provide 25% waste for a large amount of infrastructure overhead, 7.25 milliseconds would be a suitable maximum.
The asymmetry of 20% upload is also arbitrary-who and how defines 20% even remotely acceptable? Even the useless GPON is 1:2, but the more useful XGS-PON is 1:1-why define any asymmetry?
What is the "industry standard" that the competition effect allows 66%, 0.1% packet loss rate, 2% jitter, and 95%? Even for broadband, this is pitifully low-it should be 99%, and the packet loss rate and jitter amplitude are too high. The leased lines can be in the order of 4-5 9s and lower.
In any case, my opinion is that PHY should be defined as any technology that can provide any service. Whether the ISP decides to provide dedicated line service or broadband service, PHY should be the same. Compared with broadband, leased lines just retain more PONs. Similarly, if I choose to pay for ECC to lay the leased line, why can't I directly take the broadband service on it instead of this FTTPoD turkey?
Regarding the delay, the fiber does not run in a straight line between everything that it follows the population center, just like a highway network.
Including the'numpty infrastructure' at the top, you have 50% or more latency on a theoretical basis.
Since many networks may be asymmetric, the delay may also be round-trip instead of one-way.
I think this is the maximum allowable round trip to the access network, because the bidding is provided by the network builder rather than the ISP, and the upstream and downstream delays may vary depending on the access network architecture.
"Similarly, if I choose to pay for ECC to lay the leased line, why can't I directly use the broadband service instead of this FTTPoD turkey?"
The other end of the telecommunications company’s fiber is not connected to the broadband network.
It is worth noting that, depending on the product ordered, the dedicated line/DIA uses broadband as the bearer.
"In any case, my view is that PHY should be defined as any technology that can provide any service. Regardless of whether the ISP decides to provide leased line service or broadband service, PHY should be the same.
I don't know where this is the case. I know that you can rent dark fiber optics, but it is passive, and then go to ODF, and then your ISP decides what lighting they use.
The physical layer of everything needs point-to-point fiber and increases the cost of telecommunications companies, because they need to provide each customer with a 10 GBase-BX port, and the customers are all 10 Gb CPE. If a telecommunications company wants to sell anything above 10 Gb, both switch ports and CPE must be upgraded.
Dark or gray fiber, point-to-point or use WDM, seems to be the way to go. It is impractical to force the use of a single transport layer.
There is little hope for GPON. ONTs are cheap and there is no burst laser upstream. If used properly, they can provide 99% of sufficient services. The price of XGSPON has only recently fallen to a reasonable level.
PON saves a lot of ports for telecommunications companies. Mandating the use of peer-to-peer and forcing infrastructure providers to fundamentally change the way they build networks. Fiber is cheap, but the space, hardware, and power used for 1,000 front-end 30,000 10GBase-BX switch ports and ODF are not cheap, which neglects the large number of pipe systems required to accommodate fiber.
It would be very painful to try to provide services for all content through PON. So much customization work will be necessary. The 10 Gb command on the XGSPON segment must be rejected, requiring its own dedicated optical fiber or its own NGPON wavelength.
Just like we all like 10 Gb, for infrastructure builders who want to push the limits, they don't have to bid for these projects. If the requirements are too onerous, they won't.
Please forgive my cynical Carl, but after seeing a copper cable network that has been in use for about 40 years, the excavation is not an easy task, and I don’t want the optical fiber to change underground because we already have modems> ISDN> ADSL extracted from copper wire> ADSL2> VDSL2> GFast.
Before the advent of optical fiber, leased lines were T1/E1, ISDN30, and EFM was still in use. Although it was disappearing, they all used the same point-to-point copper infrastructure. The key is that the exchange already exists in a size suitable for point-to-point. Yes, it is much more expensive, but it does mean that there are many options available for future topologies.
My question is basically what is trapped underground now, due to the inertia of digging, it may be with us in the next 40 years. Of course there will be technological advances, just like DSL appeared after copper wires were buried in the ground. I really hope that PON can handle it, because without it we are very full.
Over time, you have convinced me of the overall value of PON. GPON with 2.5/1.25 gbps bandwidth itself does not have the ability to manage 32-way splits with sufficient margin, so there is too much competition to manage the correct Gigabit connection set-so I say XGS is a better choice to use here, Because in fact, how often does the ONT actually change?
As for the rest, this does not seem to put any pressure on ISPs to meet service requirements. The latency on VM and TTB is terrible. By the way, it's ok, but it has been downgraded since ADSL2+ days (I may have downloaded 60 mbps on FTTC, but the delay has doubled even though my modem is on the fast path).
Point-to-point is always best, but PON should have some mileage for a given WDM.
If GPON splitting loses momentum, XGSPON or PON splitting is an option for capacity mitigation. However, GPON is not suitable for highly symmetrical gigabit access, and telecommunications companies also know this-at least one telecom company in the state removed the gigabit option from the PON when it had existing customers.
We have never had a proper debate on cost. BSG/AM stated in 2009 that the FTTC will be 5 billion pounds, and if there is a half subsidy, it will be close to 3 billion pounds. FTTP GPON-£30 billion-will be close to £12 billion-reusing some mid-mile paths. What is the incremental cost of PTP. Oversubsidized FTTC Cure (rearranged copper) will now prevent FTTP-this is just for subsidies.
The construction work of adding hinge fittings to the poles, connecting connecting blocks and pulling cables does not seem excessive. A two-person contractor team and a picker seem to complete 4-5 hinges and connecting blocks of DP every day.
For users of 30-80Mbps, the demand for >100MBps will be a bigger problem to complete the next stage. The constant game of cost means that the UK has paid a huge price so far. How much does PtP cost?
Your email address will not be published. Required places have been marked *
This website uses Akismet to reduce spam. Learn how to handle your comment data.