It does seem to be a rule of data processing that we use all we can get of it. As BT starts to test its 21st century networking (21CN) with partners like Entanet, it's a safe bet that this rule will mean the dream of universal high speed data will turn sour as people try to squeeze more data through a limited pipe. The world of …
That's what you get for choosing cheap and unlimited
Well a lot of that is what you get for choosing ISPs that choose to compete entirely on price and saying "unlimited".
If you choose an isp that doesn't try to be the absolute cheapest and who has clearly defined limits on their bandwidth you'll get a much better service.
I use aaisp.net who aren't the cheapest (although they are not exactly expensive either) and I'm limited to 3GB download at peak times but unlimited off peak and have no problems at all with packet loss and latency. If I get over 1% packet loss on a connection that occured anywhere in my connection or ISPs links I'd be very suprised.
Any ISP that offers "unlimited" data to it's customers at a price of less than hundreds of pounds per month is simply lying to their customers as it's simply impossible to provide that for everyone at the price. They hope that people won't use their bandwidth and that they wont notice that the if they do they wont get any good quality of service.
Choose an ISP that doesn't promise anything they can't deliver, and pay a price that means they can afford to care a tiny bit and you'll get a better connection. The options are there it's just that people are blinded by the cheaper prices and no obvious limits.
As for upload speeds, I would like a slightly faster upload even if I had to trade off a slower download. Instead of 7.5G/400K I'd like 6G/1G or something. It should be technically possible for BT to achieve that but they don't at the moment.
It's not the data saviour . . .
. . . it's the M25
Same traffic rules appears to apply.
You provide wider roads and the amount of traffic will rapidly increase to cause congestion.
So the providers throw in things like 'variable speed limits' to try and give everyone a chance of getting on the motorway. Everyone can get on the motorway but now it's all slowed down.
The providers then say 'we didn't promise high speeds but we did say we make it possible to get on to the motorway.
But when you're stuck in a crawl on the fast lane you can't even get off and head for the back roads as they've all been dug up to make room for the motorway.
Broadband these day never promises speed --only connection.
hah, mine was better. I had 56K.
oh, theres my coat. cabs waiting, see ya.
oh, before I go.
Sometimes, I wish I had no soul, then I could work in marketing and sell all these lies, upto 4x10^6Mb/s with unlimited downloads.
I do feel bad for the ISPs, but they went along with the lies, so really, I don't feel that bad. They painted them selves into a corner. They promoted a business model which was a fabrication.
I've overpaid for crappy service, I've been kicked off ISPs for "unfair use" (that was funny, it was kazaa, BT, and a 33.6K modem)
Quite simply, I'm waiting for the middle class moral outrage while the ISPs play victim.
meh, its one of the reason I left the country, doesn't affect me.
(that'd be the international NIMBY)
Why is no one addressing this question?
I've yet to see any commentator examine why broadband down the phone line is so much crapper in Blighty than elsewhere.
I'm in France, where ADSL2+ is basic fayre. Despite being nearly 2Km from the exchange, I get a solid 10Mbps down and 1Mbps up, utterly uncapped, and with a static IP address to boot. Also in the 30 euro per month package is unmetered phone calls to 50 countries, which I can use from anywhere on the net with a SIP client, plus a couple of hundred free telly channels and the option to pay for VOD movies. It all works just dandy, and all the big ISPs, even Orange, will flog you something similar.
I'll bet things are even rosier in the Nordic countries. So, what's wrong in Britain?
You forgot the most important reason to limit upload capacity, as the wholesale net bandwidth ISPs buy is symetric they use the difference between the DSL upload and the wholesale for other customers who rent webspace / dedi / co-lo off them, if ISPs went SDSL and gave you a symetric connection they would not make any money because of BT Wholesales price structure which was designed by OFCOM (/TEL) to make sure you never get the upload, because they want a broadcast like net, where only the "great and the good" can speak and where descent can be quashed with copyright notices and raids on server farms, it's like the government banning protest near Parliament, except this is much more subtle.
Of course it isnt the saviour!
When other countries are doing 50mb+ no sweat, of course 21CN isnt the saviour.. .heck 21CN is nothing but an updated copy of what Australia's been running on for years.
We need a full rollout of FTTP or VDSL if the UK is going to be serious about a decent broadband network in the futrure
your analysis is just the good old fat cats'conspiration stuff everybody loves so much. Unfortunately, it is also pure economy.
Do you know, for instance, that SDSL cards are much more expensive than ADSL ones ?
So, bringing SDSL to retail customers would just not fit with isp's current business models. Or it would make subscription more expensive.
But given the fact that what the average customer notices is :
1. the price
2. download bandwidth (because he's been used to isp communicating on download speed)
...right now, bringing SDSL to mass market would just be a suicide for the isp.
I normally like Guy's stuff but there are at least three things wrong with this article.
1) It's not news
What's been going on was entirely predictable. The economics of networking, and broadband in particular, are such that you can't have truly unlimited broadband at an affordable price. No ISP can offer it, not now, not ever. Everything else follows from accepting that basic fact of economic reality.
2) Carphone Warehouse aren't a real ISP
Why quote CPW as an example of how bad things have got? Everyone with a clue could have predicted that the "free" services would be at the bottom of the charts for service quality, pretty much from day one. SUrely you saw it coming too, Guy? If you want an example of how bad things have got, pick a quality non-LLU ISP and talk about them, that might actually be *news*.
3) 21CN and LLU
CPW are an LLU operator, the only bit of BT kit in their broadand picture is the cable between customer and exchange. From there on, it's kit owned and/or operated not by BT but by the LLU operator. No 21CN in the LLU picture.
The internet needs a new direction
The only reason telephone systems work the way they do (fixed line or mobile) is because the user is guaranteed enough bandwidth and low enough latencies to provide an adequate service.
The internet on the other hand gives no such guarantees. You pay for an 8 Mb line and the small print says that this could quite possibly drop to nothing depending on other users!
Forget QOS, afterall why should VOIP get better service than peer to peer? Maybe I signed up for peer to peer!
What we need is guaranteed minimum service levels for the pipe, then the user can decide what to use that for with total flexibilty. Service plans would then be based on levels of those guaranteed minimums.
VOIP doesn't need masses of bandwidth, nor does it need amazingly low latencies but it does need a certain level of guaranteed, ordered packet delivery.
The lot of it.... 21cn aint gonna sort anything, hell, isdn2 cant even run on it, so the programs taken a pause in roll out.
bt just does my head in - and i work for them... sad.
AI Break Down of Quantum Communication .... The Idiot's Guide to AI NEUKlearer Bombe
"Broadband these day never promises speed --only connection."
Words of Undoubted Infinite Wisdom, Keith. Thanks for sharing them with this connection. And as with all "products", quantity is always tempered and mitigated by the arrival of levels of quality so that a market can be servered in Tiers rather than overwhelmed with a simple base element..... allowing variations on a theme to be successfully milked for all its worth, without collapse due to rapid uptake.
And furthermore, all that is Really Connected Virtually, is Information and it is easily lost, although most probably diverted and/or stolen, unless you have a Universal RoadMap and/or a Particular and Peculiar Plan for IT.
Random Bits and Bytes/QuBits and Quytes are, in isolation, both Practically and Virtually Useless and Worthless, and especially so if they are Uniquely IntelAIgently Designed to Create a NEUKlearer Picture.
Such is the Nature of these HyperRadioProActive pieces of Information, that in Stored Isolation without all other Key Complementary and Complimentary Elements/Pieces of the Jigsaw, are they not without Burning Consequence in unnecessary Delay/Decay, given the Inherent Toxicity of Individual Pieces without Master Key Positioning/Uniquely, QuITe Specifically Designed Use in ITs Master Key Program ....New World Order Program ..... Universal Big Bang Theory of Everything Picture.
However, unlike its earlier nuclear radioactive Big Brother model, which becomes increasingly more dangerous and more practically useless as the model progresses/regresses to release its Explosive Maximum Power and Irradiation with Destruction, the Total Information Awareness Universal Big Bang Theory of Everything Picture Model, the NEUKlearer HyperRadioproActive Transparency Model, Irradiates ITs Destructive EMP Effect NOT on all pieces of the Information being delivered and received for Completion of the Picture/Model ..... for then is IT IntelAigently Designed to be Mutually Beneficial and Harmless, ..... but on those who would divert and retain Elements/pieces in an Effort to Maintain the earlier nuclear secret Position/Picture.... with nuclear being most definitely nothing of the sort, it being a very dark secret control parameter for destructive powers that be.
The Absurdity and Dumb Ignorance in any Maintenance of such a Debilitating Self-Destructive Position against AI NEUKlearer Model is surely in the Knowledge, that unlike in the earlier and failing nuclear model where Ore/Source is scarce and difficult and dangerous to handle, in the Transparent NEUKlearer Model is the Ore/Source Universally and freely and harmlessly available with the only danger being in and with those who would think to prevent ITs Sharing of Knowledge.
And all of that is also a Challenge to BT 21CN to Provide a Mentor Service which Delivers Leadership without Taps to Monitor and Redirect Output ie Provide a SMARTer Intelligence Service because .... with all Communications/ISPs/Networks routed over their Infrastructure, the withholding of Information in any particular Network because of its Sensitivity/Perceived Sensitivity is a recipe for Easy Abuse with any Tom, Dick or Harry with any sort of public role enabled to hide behind its cover.
The Age of Privacy and Secrecy and Grand Scheme Players in Obscure Private Orders is Gone. Information and the Questioning of Information to discover the Intelligent Quotient/Material Relevance of ITs Source so that the Best Intelligence Information rather than the Incumbent Intelligence Leads is the Nature of the Virtual Reality Today ..for a Beta Transparent Future.
And that goes for you no matter whether you live in Afghanistan or Zimbabwe, for that is the way that IT is.
Ok.... Rant Over. IT just needed to be said. And you're either for IT or you're with the terrorists, by George.
VoIP does need...
"a certain level of guaranteed, ordered packet delivery."
Indeed it does. Now, who remembers the difference between packet switching and circuit switching, and who remembers why packet switching appears to be cheaper per Mbit/second?
The answer: packet switching is cheaper largely because it offers no guarantees of delivery, let alone ordered delivery or even end to end flow control; the same resources (kit, capacity) can be used to apparently offer more Mbit/s but *with no guarantees*. With packet switching if something goes wrong (eg a lost packet), someone else will sort it out (eg error recovery code in applications). Packet switching just moves cost from one place (the network) to another (usually the applications). Sometimes that's a win overall, sometimes it's not.
Circuit switching offers guaranteed bandwidth and guaranteed delivery and the like, but the guarantees cost, and in today's markets low price usually trumps top quality.
Everything in between CS and PS (MPLS, IPQoS, etc) is just IP-technology marketeers trying to dodge the fundamental issue, which is that trying to fit a quart into a pint-wide fibre may appear to work some of the time, but cannot possibly work all of the time, not even if Paul Daniels is in charge. Just ask any broadband user who has bought a service with worse than 1:1 contention without actually having it explained properly (which is probably 99.99% of us).
Anyway, who needs consumer/SoHo VoIP (at least for cheap calls) when you've got plenty of outfits offering (eg) UK calls of unlimited length for 5p per call (even if they do have VoIP hidden in the middle somewhere)? If they can do it for that price, without punters having to bother with all this VoIP hassle and pathetic sub-mobile call quality, howcome BT (and Virgin) calls are so expensive (and CPW and pretty much all the big names)?
As for the way this stuff affects the delivery of IPtv... ha ha ha ha ha ha ha ha ha, what is the emperor wearing today?
Am I missing something?
But when was 21CN (or any NGN for that matter) supposed to have an impact on data speeds? 21CN is effectively a change to the carrier on the network, and the replacement of legacy voice switches with softswitch's.
For the carrier its supposed to bring a cost saving, due to more capable softswitchs and the IP backbone, both of which improve capability and reduces operating cost as they are a dam site easier to manage then the legacy kit that they are replacing (cheaper to buy, smaller, cooler etc).
What 21CN does allow is abstraction from the physical layer of the last leg, so it doesn't matter whether your coming in on copper, or fiber, or Ethernet, you still have access to the same core services using the same kit (ok the MSAN may need a different card, but its better than a new switch and trans coder). This has two huge benefits (to the customer), 1) all those people that currently have fiber to the door, will be able to get a broadband service and 2) it means that if (/when) fiber to the door finally gets rolled out to the rest of the uk - there is little or no (comparatively) extra cost to the core of doing so.
So what will change...
IMHO, this article should have focused on the thing that really matters, Money! What NGN's really give carrier's is a much more flexible billing and enforcement system - QOS is a means to an end in this. For anyone who has sat through a vendor presentation from the likes of cisco, nortel, ericsson, nokia, etc there is one word that keeps coming up - PROFILE. This is the differentiation of services per subscriber when everyone is connecting over the same network.
Back to money, providing a data service (ISP) makes pretty much no money nowadays (hence why all the small isp's are going under /selling out to BT). The big change - as far a I can see this NGN's will be the selling of services on top - you want VoIP, that will be £1 extra a month, P2P thats an extra £3. 21CN gives the carrier the ability to be able to do this by managing an end user through profiling and its done right at the point where the customer comes in to the network - not as currently - at the point where the carrier peers with other networks, the transmission savings are huge!
I'd say it wont be long before capped /unlimited becomes bronze, silver, gold and then your broadband will be free, if you want to surf, with a fee payable per service that you want to run on top (e.g. VoIP, p2p, video) - just like if you want caller display or voice mail at the moment on your POTS connection.
Just my 2p
Crap analogy. Extra lanes on the motorway means the "middle-lane hoggers" simply move from lane two on a three lane motorway to lane three on a four lane motorway meaning there is still only one lane to pass these morons creating a bottlenext. Doesn't happen to broadband lines.
If yer so smart, you come up with a good analogy!!! have you not heard of the motorways where you can also be fined for 'going too slow' in the fast lane, as well as going too fast in the outside lane??
meanwhile here is a good one....
An analogy based on all you can eat pizza, bless...
What happens when there isn't enough pizza again...
What happens when there isn't enough pizza again
What happens is that those few customers who have realised that "all you can eat" isn't guaranteed, and who have a genuine need for "all you can eat", with budgets to match, will have made alternative (more expensive) arrangements well in advance. And they'll have had plans to test the kitchens under load, and all that.
Meanwhile, those who have limited budgets or limited Clue and thus haven't made alternative arrangements will wait for their pizza while the kitchen tries to match its limited capacity against the (hopefully short term) peak in demand. Some people will be unhappy, but hopefully they will be getting what they are contractually entitled to - possibly not very much at all, depending on the small print.
Similarly, those who were happy to save money by using IP bandwidth rather than circuit switched end-to-end-guaranteed bandwidth can sit back and enjoy their cost savings as they wait for their data/pizza to arrive. And wait. And wait.
There really is no such thing as a free lunch. Bandwidth is bandwidth, if you want it cheap you get packet switched and contended, if you want it guaranteed you get to pay for the privilege.
Wabbit02 is right - this has relatively little to do with 21CN and everything to do with access speeds. I'm surprised that nobody's mentioned Ofcom's current consultation "Future Broadband - policy approach to next generation access" - http://www.ofcom.org.uk/consult/condocs/nga/future_broadband_nga.pdf
Regrettably, the document focuses on why we should not deploy fibre yet. The argument is that we don't have enough evidence of demand to justify a reasonable return on investment. So splendid though Guy's article is, (and I for one support every word!) it is really a collection of assertions about demand - it isn't the sort of "evidence" that will convince an investor - or the regulator, for that matter. Add to that any number of other reasons why BT isn't in any particular hurry to deploy fibre, and you can see why we're still in that sterile argument over that obsolescent technology called xDSL.
Stephen Timms and other Ministers are now talking about the need for a national fibre access infrastructure - but will not spend taxpayers' money by intervening in the market. So the challenge is to convince them, and Ofcom, that the demand exists and the time to come up with a national policy that will give us "real broadband" where and when we want it, is right now. Got any genuine, reliable facts and figures, anyone?
A simple question
"....will have made alternative (more expensive) arrangements well in advance. And they'll have had plans to test the kitchens under load, and all that."
And what could possibly be their excuse for not stepping up to the plate and taking over as Ace Supplier, Anonymous Coward?
The good old days
Long, long, long before the advent of ADSL. When you lot were on 33K and 56K modems, I was on a cablemodem trial in the Nynex Manchester region for 18months.
Websites and content were so simple that they arrived on my computer screen in the blink of an eye and webcams updated at 10fps.
Them was the good old days when the internet was simple: when men were men; women were women; and small, green furry creatures from Alpha Centauri were...
- Mounties always get their man: Heartbleed 'hacker', 19, CUFFED
- Batten down the hatches, Ubuntu 14.04 LTS due in TWO DAYS
- Samsung Galaxy S5 fingerprint scanner hacked in just 4 DAYS
- Feast your PUNY eyes on highest resolution phone display EVER
- AMD demos 'Berlin' Opteron, world's first heterogeneous system architecture server chip