Hard News: Campbell comes back
54 Responses
First ←Older Page 1 2 3 Newer→ Last
-
The other thing that happens with fat pipes (I have cable here in the UK) is you get what they refer to as 'traffic shaping'. So I nominally have a 110MB/s pipe (upgraded from 0.5MB/s several times for free). But I rarely get that, largely because I don't need it but also because I use the net when everyone else in the area does, and there are teenagers hereabouts. So to give everyone connected to the local hub a workable net experience the ISP limit your bandwidth at certain times. It's poverty of the commons innit? So it needs a regulator.
In addition if I constantly stream videos while playing online through the cable modem via my PS3/XBox?Wii then I may well find that the ISP has throttled back my pipe to allow others on the local box to get a look in. The ISP's boards are full of people asking if they have been shaped and bitching about it. Of course it is also a revenue gaining scam since I can always pay more for a fatter pipe which funds them installing a bigger local box etc.
Having enjoyed the benefits of cable for several years (it is VERY stable) I would be most reluctant to got to ADSL over copper. For one thing going all cable (including the phone) means I could finally have nothng to do with BT, nothing at all.
Now if we can just see off Phorm, all will be sweetness and light ;-)
-
Bugger, I sure didn't type 110MB. Make that 10MB/s.
-
Rich said:
Which begs the question, since one can get 24Mbit with current DSL copper technology, and VDSL2 is even faster, why do we need fibre to the home? Is this a subsidy for a small minority that want HDTV movies in the time in takes to make a cuppa?
Did you actually read anything more about VDSL2 than just the headline speed? Such as, maybe, VDSL2 deteriorates quickly from a theoretical maximum of 250 Mbit/s at 'source' to 100 Mbit/s at 0.5 km and 50 Mbit/s at 1 km, but degrades at a much slower rate from there, and still outperforms VDSL. Starting from 1.6 km its performance is equal to ADSL2+? Most NZers are a lot more than 1km from the exchange. Phonelines aren't run in a straight line, they're full of bends and joins and other things that both lengthen the run and lower the quality of the connection. Those speeds are very much reliant on a "clean" connection, and there aren't very many homes in NZ that actually have one. The copper's been spliced and joined and repaired and generally fucked with so much that in a lot of cases it struggles to give people even ADSL speeds never mind VDSL2.
Why does everyone assume that faster speeds are simply for downloading movies? What about graphics artists who work from home? Some of these small companies us DVDs to ship work to clients, because it's too damn slow to try and use what passes for "broadband" here. Some even send DVDs to the US, for crying out loud!
Also, copper has real limits. We keep pushing them out, but every increase in speed has shorter operating distances. Fibre runs at high speeds for the length of the connection, not diminishing with distance. Run FTTH and we could get "at source" VDSL2 speeds to everyone regardless of how far they are from the exchange. Even if they're high country farmers.Not to mention that while copper seems to be very versatile, I'm not convinced that fibre can be upgraded as easily. It certainly seems that any upgrade would need new subscriber equipment in every home in an area. I'm happy to be proved wrong on that, though?
And it doesn't require an upgrade of subscriber equipment for every new variant of DSL that comes out? You have to buy a new modem, usually, to go from ADSL to ADSL2 to ADSL2+ to VDSL, and so on. How is fibre any different? Plus, fibre's expansion capacity is virtually unlimited and over much longer distances. You can run 100Mb/s for certain over several km using fibre, but there's no way in hell to do that using copper. Also, if the roll-out is done with some foresight (yes, yes, I know, foresight in NZ telecommunications, I must be off my rocker), people could be attached with equipment capable of doing 1Gb/s right from the get-go. The cost is very, very similar to 100Mb/s (go have a look at the price of fibre-based network cards), and gives a lot of future-proofing.
Also, fibre termination doesn't change. Like you don't need to get new phone jacks in order to use new variants of DSL, you can continue to use existing fibre terminations if there are upgrades to the transmission equipment. In the same way that you can connect a 10Mb/s NIC to a Gig-E switch, and it'll work, you can plug an old fibre NIC into a higher-speed network. They just negotiate a common speed and run with it. I believe my home exchange has been upgraded to ADSL2+, but we're still using an ADSL router. Works fine. No upgrade required, but if I want higher speeds I'm going to have to fork out. No different to fibre.
-
One of the best use cases for cheap broadband for business is backups. Offsite backup to someone else's storage is safer, cheaper and easier IF the bandwidth is there.
-
Rich said:
But wouldn't it be cheaper to make a clean copper connection the basic standard than to go to fibre?
*choke* You mean re-lay the copper to every home in the country? Buy tens-of-thousands of kilometres of high-quality copper wire and run it down every street, back-road and goat-track in the country? Can I please have some of what you're smoking?
Seriously, if you're going to do that you really should just future-proof and lay fibre. Fibre is dramatically cheaper than copper (in those kind of quantities, maybe 1/6th the cost of the equivalent quantity of copper), and the biggest argument against running FTTH is the cost of actually getting it to the end-user. If the plan is just to rip everything and start again anyway, then fibre would be the choice of all the people who make such decisions.
-
BTW, folks, kudos for more substantive discussion and analysis of National's broadband policy announcement than the Herald, TVNZ, National Radio or Three can be bothered with so far.
As far as I can tell, David Cunliffe thinks its crap, but doesn't actually have Labour's response ready to go yet. Gee, I couldn't have guessed the first half of that for myself. :)
-
Which begs the question, since one can get 24Mbit with current DSL copper technology, and VDSL2 is even faster, why do we need fibre to the home? Is this a subsidy for a small minority that want HDTV movies in the time in takes to make a cuppa?
I suspect that once our connections kick up to a sufficient speed and cap (or capless), you'll start to be able to subscribe to (say) American cable TV via the internet. You'll stream BBC straight over the net.
You won't go to the store to get your DVD, you'll get the movie by clicking your mouse. Etc.
Took my computer about an hour to download a relatively low resolution version of my Anaheim Ducks losing to Dallas Stars in game 6 yesterday. About 120 minutes of video. Up that to full HD in full screen, it would have taken 3 hours or so (and speeds at my house seem to have doubled recently). When it takes 5 minutes, that'll be fast enough.
-
I'm sorry to question peoples expertise and quibble the need to spend them Beeeeleon dollars now!
But I *did* read the VDSL2 article and I knew a bit about it anyway (I have worked in comms based IT for over 20 years, you know).
My point is that I haven't seen tangible evidence that FTTH is going to be more cost effective than copper + cabinetization. (You need cabinets with fibre as well - there needs to be a switch/concentrator somewhere in the system). Making (say) 10Mbit the standard for copper wouldn't involve rewiring the network - a high proportion of circuits can already run that sort of speed if Russell's comments on the Point Chev cabinetization rollout are anything to go by? I'm prepared to be proved wrong on the upgradeability issue but would like a reference.
On the usage side, I'm assuming this is all about FTTH, not business services. My work already has fibre, so I'm assuming CBD locations are already cabled for business customers. I'm sure it would be lovely for graphic designers working from home to have 100Mbits, but how many of those are there? (Maybe it'd be cheaper to just give them all a million dollars and a pencil).
For the home applications that will use all this lovely bandwidth, I'm told it's a mistake to ask about them. Just build it, and they will come. Reminds me of something: For carrying on an undertaking of great advantage; but nobody to know what it is.
-
I use video conferencing systems occassionally, as well as occassional dilly-dallying with a variety of virtual worlds - both of which are rather useful or fun. I sometimes get rather excited to think where these technologies could end up in say 10 or 20 years, assuming technological progression continues to occur. A merging of virtual and real worlds into one system would be a huge benefit to society and the economy.
We would of course need an insane amount of bandwidth to run such a delightful future.
-
Rich, mate, no one could have imagined what 800 baud modems would do for business, like how newspaper reporters could suddenly deliver photos and stories complete from on site to the editors, allowing more time in the field for the same outlay.
No one could imagine the wonders that 14.8, 28.8, or 33.6 would do either. The explosion of the web, online booking for services, online help for products, replacement manuals, and societal participation such that literally every topic imaginable was soon available as text + gif, all indexed by near-instant search engines of ever-increasing quality.
Like a library, only like every library ever, only with a librarian who has the book you want in hand as you walk in and wants to talk with you for hours about the fine details of the topic.
Businesses track inventory, make purchases, contact customers, advertise, seek interested partners, take reservations and make sales, and more, all for a fraction of the cost it used to take.
Faster speeds, early 64k, later 56k, 128k, 256k, 512k, 1M, 2M, ... every step spawns entire new industries, makes old tasks so cheap that those who used to sell those services disappear (or legislate and sue), allows most business to do more for less.
Not to mention the ever-increasing difficulty that repressive governments have in hiding inconvenient facts from the public.
What we'd all end up doing with 10M is like asking what we'll do with the findings of the Large Hadron Collider once it's up and running. Until it's there, we can't really imagine. Thousands of imaginative people will try things (because starting small on the net is amazingly cheap), and a handful will take off, and change everyone's life in the ways that they will find a new desire for. -
tussock: You said it far better than I ever could.
Rich: You seem to be quite confident of the quality of the copper that feeds most of NZ. Lots of people aren't, and for good reason. Also, you're still thinking small. You see 10Mb/s as the baseline, but where to from there? 10Mb/s isn't even particularly fast now, and by the time it's available to a large majority (80%+) of the population we'll be standing in the same place we are now, relative to the speeds available to the rest of the world.
Don't forget, we're trying to get into the top half of the OECD, and stay there! Which means we have to be running faster than the people who're a) already there, and b) not there but want to be. And we have to keep running faster, even if we make it there, because everyone else is running too. Saying that 10Mb/s is a perfectly acceptable speed, for now and for evermore, loses sight of the fact that we'd barely make it into the top half of the OECD now if we had 10Mb/s to every home. In a couple of years' time, we'd be back in the bottom half and we'd have lost that couple of years in which we could've been building an infrastructure that would let us keep delivering ever-faster speeds to consumers in the same way that our "competitors" are.You want some examples of things that we could do with FTTH, that are difficult now. OK, here's a couple. 1) Video-on-demand. That's where things are heading, and it's been observed more than once in the MSM in the last year that as things stand in NZ, it's not going to happen. Even 10Mb/s is pushing it, especially for high-def content. 2) Proper telecommuting. I'm not just talking about being able to fire up a Citrix session and use Word/Excel/Outlook, I'm talking about full-on video conferences, with multiple home-based participants, using quality cameras and without jerking or voice-sync issues. Can't do that currently, either, because our up-stream speeds are shit. They can't improve much with ADSL, either, because of that whole asymmetric thing.
So there's two examples, both here-and-now, and both really stretching the capabilities of even VDSL2 never mind ADSL or ADSL2+. Fibre has better latency, meaning that voice-based applications work better, and it scales beyond anything possible with copper. You seem to be fixated on us being "good enough" for "right now", but the Government (and National) have said that they want us in the top half of the OECD. That means being better than we are now, and if we want to stay there it means staying better than no fewer than 15 other countries who will also be trying to improve the services available.
-
Don't forget, we're trying to get into the top half of the OECD, and stay there!
What's this 'we'?
Personally I couldn't care where we fall compared to other countries. As long as it does what it needs to do with improving technology. I just don't think it'll be doing that unless we get faster.
-
The "we" of a direction set by our glorious elected representatives. AKA, the Polly Tubbies.
With all the bitching and griping and moaning about how terribly NZ does in the OECD, National cannot ignore this opportunity to actually improve our standing in some way. Labour's already said they want our ranking to shift, and National's agreed. Ergo, "we" have been told by the residents of the Beehive that this is something we want.
Personally, I, like you, don't much care about the OECD stuff I just want it to work at speeds and prices comparable to other countries.
-
For telecommuting, I think the main issue is not so much a lack of quality video-conferencing as a desire for employers to have their employees onsite 9-5 so they can be sure they aren't slacking off.
I'm amazed that most of my friends who work for the public service aren't allowed to telecommute. You'd have thought that the government would have made it a general rule that anyone who can work from home should be allowed to.
Given that the economy looks like it's turning down though, maybe anything that puts $1.5bln of cash into the system might be a good thing though. The 21st century version of throwing banknotes off buildings?
-
Yes, employers are proving to be remarkably obtuse about the whole thing. Proper telecommuting setups are also a good business continuity plan in the event that H5N1 actually fully mutates into a human-human transmissible virus and the world gets struck by a flu pandemic. And if that doesn't eventuate, at some point there will be a pandemic, with consequent impacts on social interaction. Adequate facilities for staff to work from home will mitigate greatly the adverse economic fallout.
-
Just a thought - it should be technically possible to have ADSL that favours upstream over downstream - in which case you could have two phone lines, one for up and one for down. No?
-
Rich wrote:
Just a thought - it should be technically possible to have ADSL that favours upstream over downstream
Sure, it's "technically possible" to do that. But to the best of my knowledge there's no variant of DSL that actually does that. So you'd have to write the spec, test it, get it approved by IEEE, get someone to make the kit for the telco end, and get consumer equipment manufacturers to produce stuff that could use it. It's not just a matter of twiddling some settings, it's a whole new specification with all the attendant development work. Figure three-to-five years if you start tomorrow before it's actually possible to get it deployed into homes.
in which case you could have two phone lines, one for up and one for down
That'd work if you had equipment that could bond the links. Since a reverse-asymmetry variant of DSL would require completely new consumer kit, it would actually be fairly easy to cope with this concept in hardware. Then you'd just have to set things up on the ISP end to allow two links to be bonded, but link bonding is old hat now.
-
Right right answer, wrong explanation
The ADSL system relies on having all the "loud" high bandwidth signals at the DSLAM end and the "quiet" signals at the subscriber end. Running reverse DSL causes too much crosstalk and reduces capacity/distance.
I found a reference to the software actually being available - it's obviusly just the above problem thats stopped anyone deploying it. That, and the fact that home network connections aren't really intended for publishing, as opposed to consumption.
-
Thats network equipment (DSLAM) software not consumer software, BTW. You can't try and hack your modem to get mega-upstream bandwidth.
-
Rich, that's interesting. Makes sense. The whole near-end cross-talk vs far-end cross-talk problem hadn't occurred to me, but it's a logical (and very significant) impediment to having "standard" and reverse-asymmetric DSL connections bonded to give a fast bi-di link. True symmetric DSL is very, very speed limited, with the "high speed" variant giving a hot 5.69Mb/s to a distance of 2.7km. Pretty lousy, really.
If we want to look beyond low speeds (10Mb/s is fast-ish now, but won't remain so for too much longer) we need to go fibre.
When Telecom was in the process of removing the downstream speed cap, there was a discussion on the NZ Network Operators' Group mailing list about the limitations that upstream speed imposes on downstream. Namely, a 128kb/s upstream will limit the usable downstream to about 4Mb/s for some types of TCP traffic because of the need to send acknowledgement packets. That problem persists at higher data rates, and there's only so much capacity available in the upstream before the downstream must be reduced to minimise cross-talk (as per that link you posted). Fibre has no such problem, even though consumer FTTH connections in the US are frequently installed with asymmetric upstream/downstream speeds to discourage running servers at home. That's purely an ISP policy decision, not a technical limitation as it is with DSL. -
You mean a TCP connection needs 1/40 of the downstream data rate to send its ACK packets? Sounds about right. Though for things like streaming telly, an option is to use UDP send-and-pray and let any noise appear as dropouts, just like on a Sky dish.
-
In any case, DSL's not quite dead yet, if John Papandriopoulos and John Cioffi have their way.
-
Rich said:
You mean a TCP connection needs 1/40 of the downstream data rate to send its ACK packets?
I didn't do the maths myself, I simply relied upon the reputations of the people making the claims. But when you put it that way, yes, that does sound about right.
There are a number of options for streaming media, including things like multicast. Their deployment within NZ's backbone has been underway for quite a while, increasing gradually. They're not quite so reliant on the upstream connection speed of the clients, usually, but they still require downstream speed, especially as their definition and thus size increases.
Regards the improvements of bandwidth vs computers in general, 10Gig-E is a mainstream standard, 40Gig-E has been standardised, and 100Gig-E is in progress. Five years ago, Gig-E was the mainstream and 10Gig-E was the very bleeding edge. Progress happens, and pretty fast. The nicest thing is that you just have to upgrade the ends, not the middle, provided you've got fibre. Look at the SCC, which just keeps on getting faster without having to be re-laid. Can't do that with copper.
-
To utterly digress:
My Ph.D. dissertation is currently bound by a confidentiality agreement due to intellectual property issues
Does that happen often? I've always thought of the concept of a university was that research (aside from that being done as a specifically commercial venture) was for the advancement of mankind's knowledge and had to be published.
-
My Ph.D. dissertation is currently bound by a confidentiality agreement due to intellectual property issues
Sounds to me like the thought is about to be put into action?
Post your response…
This topic is closed.