Hard News: Shooting for the Moon
115 Responses
First ←Older Page 1 2 3 4 5 Newer→ Last
-
Clearly, some people haven't got a clue about what it costs to put in underground cable everywhere. By the time you get planning permission from fractionated regional authorities and have dealt with the unexpected and unmapped sewer/electricity main/geological formations/legal challenges/Maori claiments looking to clip the ticket etc etc etc you've got cost over-runs everywhere. Then someone in charge of the money says "why are we spending all this Y money when we have only got tiny x number of customers in that part of town"? And guess what - no one can give the money people a sensible answer. Chicken and egg arguments cut no ice with the money people and wishful thinking never gets anything done once the accountants get involved so the whole thing gets canned.
Which is exactly what will happen in Auckland to the National Party's inititiative if they win the treasury benches.
-
I'd love for Wimax to come along and solve all our problems but is that still going to happen? I remember a couple of dire articles a few months back like this that had me rather worried. Still not sure if that is angst or an actual problem.
Oh and there is plenty of broadband internet whining in the UK at present - based around BT's 21st century network. Apparently it is delayed and not good enough compared to Europe. So apparently the grass is always greener.
Regarding Telstra - their talk of 17.5% pretax profit was probably written in reference to their great fear of declining traditional revenue. Their financial models probably require next gen/new wave networks to achieve this rather ambitious return in order to make up for their projected revenue loss.
-
Russell said:
I understand Vodafone's annoyance at the sudden movement of the goalposts that is the cabinetisation project.
It's not even very sudden. At the time that TCNZ announced it and everyone started whinging about how it was "shifting the goalposts", commentators such as Juha were pointing out that cabinetisation had been rumoured for at least 18 months. The shock for those of us who follow such things was more that people said it was a sudden, anti-competitive move. Oh, and that Telecom was actually going to do it, which may have been the real source of woe for their competitors given TCNZ's long history of not investing in infrastructure.
-
Tom: it'll be a good test of National's resolve to make the process work, then, won't it?
Ben: wireless will never beat cable, unless someone finds a way to make it a non-shared medium and deals with several other fundamental issues. Check out the actual throughput figures for wireless vs wired connections. With wireless, even 802.11n, you're doing well to get 75% of the theoretical maximum under perfect conditions. With wired, it's entirely possible to get 90% or more of the theoretical maximum. I've seen captures of people getting 930Mb/s over a 1Gb/s ethernet connection, which will never be possible with wireless.
-
Captain Anal here...
- Amsterdam has a wonderful ducting system in place: they're called canals :)
- Careful with the "Mbps" (millions of bits) "MB" (millions of bytes - a factor of 8 larger...) http://en.wikipedia.org/wiki/Mbit/s#megabit_per_second
- Streaming HDTV needs 28 Mbps to eliminate all artifacts BUT (as Motorola discovered when they ignored GSM) people will tolerate much lower quality
- The handshake / error correction tips HDTV up to those sort of demands: put in a bit of decent buffering (say, 30 seconds), combine it with a variable quality mechanism (e.g. the credits, blank bits and so on get compressed much more) and you can get away with 10Mbps easy...
- The NPV of empty ducts is huge. From an accounting point of view, these things are a significant asset.
- Don't dismiss things like blowing fibre down gas / water pipes. It's been done all over the place and makes a huge difference to the cost. When I lived in Portsmouth in the 90's Council partnered with the CableCo. to put reusable ducts down the middle of the street. It could then be reused for all services except waste. I expect its on their books as an asset!
-
slarty, I can barely watch YouTube as a straight stream at home now, and that's stupidly low-quality video. I end up buffering the entire clip before I start playing anything. We can't even do shit video reliably at the moment, over 2Mbps DSL.
On the existing "ducts" thing, anyone know if it's actually been suggested in the discussions about how to get this thing done? All the main population centres have reticulated water, and using the pipes would make it a fairly quick way to get the fibre laid. Would also pretty much entirely negate RMA concerns since the pipes are already there. Not sure how the taniwha would feel about it, though.
-
I'd love for Wimax to come along and solve all our problems but is that still going to happen?
I've just checked my facts regarding WiMax and Woosh's roll out date. I now have it on good authority that the roll out is set for late 2009. As far as I can tell they will be looking at 802.16m (Gigabit Wimax) and they are not alone globally speaking, there is some background on the technology here
As for using existing ducting sewers etc. that's good for fibre/cable to the street but that last "mile"/ few meters really is the problem, just think about it connecting to a backbone fibre every 20 meters or so would create all kinds of expense (like, do you run fibre to all existing properties or just those who want it?)
I like this scenario.
Fibre to the cabinet (already under way)
Fibre to the street/locality
WiMax to the homes/business
What do you guys think of that? -
I don't think that the NZ government should be making the technical call on what to use.
Rather, how about setting requirements for a 5 year and 10 year schedule, e.g:
5yr: 75% with 20Mbit, 95% with 2M, everyone (including rural) with 512k
10yr: 75% with 100M, 95% with 20M, everyone with 2MThen put the implementation for each district out to tender using whatever technolology mix the tenderer wants, WiMax from lampposts or whatever.
There would need to be some access rules, and arrangements for e.g. Telecom copper to be acquired at a fair price, but not one that reflects a monopoly position.
-
Rather, how about setting requirements for a 5 year and 10 year schedule, e.g:
5yr: 75% with 20Mbit, 95% with 2M, everyone (including rural) with 512k
10yr: 75% with 100M, 95% with 20M, everyone with 2MBeen done. I haven't got time to go looking now, but from memory, it's relatively similar to your suggested schedule.
-
Yeah, take your point about YouTube - the way you get this stuff to work is control both ends of the pipe... so your video distribution box is in a cabinet at the end of the street. Sounds pricey but may actually be cheaper than getting good quality copper / fibre over a long-ish distance.
Main thing that stuffs up video over IP is inherent in the nature of the protocol - unlike old fashioned stuff like SNA there's no QoS built into the packet, so you do it all at the choke points... and there's no management of those on teh internets...
Hence get your control point as close to the consumption as possible. It's probably not the last mile or the pacific wholesale that's stuffing up your video, but the contention within the NZ backbone (download PingPlotter and have a look. It's interesting!)
-
Network infrastructure isn't something that should generate spectacular returns. It's meant to be a safe investment, not an especially profitable one.
That implies low risk. Installing a fibre network when you have no idea on takeup and acceptable pricing isn't low risk, in my book.
-
Is wireless going to be able to keep scaling? Will the actual throughput rates (as opposed to the theoretical maximums) keep pace with demand for capacity by end users? Can the standards actually be agreed upon and available in consumer equipment in time for them to be of use? The kit to terminate fibre connections is here now, but 802.16m is still only a draft spec.
I'm always leery of claims that wireless can be made to perform at high speeds. Thus far every wireless standard that's meant to be "fast" has spectacularly failed to achieve any such thing. Look at 802.11n, which claims 200Mbs in both directions but typically gets 40Mbs on the downlink. That's faster than 11g, sure, but it's way short of the headline rate. Big promises require big evidence, and every time the promises fail to eventuate it requires even bigger evidence to support the next big promise.
-
It's probably not the last mile or the pacific wholesale that's stuffing up your video, but the contention within the NZ backbone
More likely it's the fact that my flat is on Go Large, and we're thus competing with the remaining people on the plan for what is, I'm sure, a shrinking pool of bandwidth. The ComCom case against TCNZ over Go Large will be interesting, when it finally gets to court in twenty-something-teen. Such a useless system :/
Rich: it's still low risk because the asset stays available and viable. It's not going to rot away, it's not going to become obsolete in 18 months. That makes it low risk. Retail services are high risk because what's being sold is so ephemeral. Network infrastructure, regardless of the potential uncertainty of use at the time of installation, is not.
-
So my next question is: what price would you pay for 50/20Mbit/s service? With, say 100GB of data.
As someone else has noted, that data cap would start to become a problem. Currently I have a 20 GB data cap for low quality video. Once the speed goes up to that, I'm going to be downloading 5GB files.
I'd be happy to pay up to $100/month which is about what I pay now for phone line and internet, as long as the connection could take care of my phone as well.
-
I'd be happy to pay up to $100/month which is about what I pay now for phone line and internet, as long as the connection could take care of my phone as well.
That's the kind of responses I'm getting too but we're going to need bigger international pipes to get rid of capping but that is already on the cards with Kordia and PIPE
-
More likely it's the fact that my flat is on Go Large, and we're thus competing with the remaining people on the plan for what is, I'm sure, a shrinking pool of bandwidth. The ComCom case against TCNZ over Go Large will be interesting, when it finally gets to court in twenty-something-teen. Such a useless system :/
Aaah. You see, information really is everything.
Add up your estimated cost differential, send them a bill, send them a warning then file an action in the small claims court. It'll be faster (it's open and shut breach of contract)...
-
Steve, it's not just international pipes that lead to capping, it's the anti-peering nonsense. While traffic destined for domestic servers has to trombone through the US or West Island we're going to be stuck with restrictions on how we use the 'net.
A huge bitch I have right now is that, despite my work being attached to reportedly with the fastest 'net connection in the country, I cannot watch "broadband" video streams from TVNZ and TV3 smoothly. I had to use TVNZ's international dialup stream to watch Sir Ed's funeral in anything approximating real-time, which is complete bollocks.Oh, and on the pricing thing, my $150-175 suggestion was an expectation that it would be a complete replacement for current landline communications at home. No "plus phoneline" crap.
-
Matthew,
Yes, the anti peering nonsense indeed. I had someone try to explain the reasoning to me once, all I could say was WHAT? It came down to either I was stupid or they were, you can guess what my analysis of that was. I also wonder why it appears that PA is hosted in San Francisco, is it?
I often have to "cheat" to get video streaming to work on my setup, snatching it from the cache and saving the file. For some reason I haven't even bothered to try and figure out this won't work for TVNZ on Demand, it seems to only get the stream up until the first "ad. break" I'll have a play with VLC to stream after a good buffering period one day but that kind of defeats the idea of "On Demand" eh?
I may sometimes come across as an apologist for Woosh but it's just old fashioned loyalty and the fact that I like wireless, cables and fibre just seem so 20th century somehow.
What's your take on WiMax? -
Matthew: If networks are always low risk, what about Iridium, or Telecom's CDMA white elephant. Both of those work fine, but failed to meet business expectations.
-
Neither Iridium nor Telecoms CDMA could really be called networks. Rather protocols and associated hardware. Or am I just being picky ;-)
-
I also wonder why it appears that PA is hosted in San Francisco, is it?
It is, along with a number of other sites built and hosted by CactusLab.
There's a simple reason for it: it's much cheaper (and simpler) to host your New Zealand website outside New Zealand.
-
Matthew: If networks are always low risk, what about Iridium, or Telecom's CDMA white elephant. Both of those work fine, but failed to meet business expectations.
Iridium has so much wrong with it, including horrifying compression, 10 kilobit data transmission rates, call rates of up to $US14 a minute and, these days, no consumer devices available at retail.
I don't think it's fair to compare the proven, long-lived and scalable technology of fibre-optic networks with that.
-
it's much cheaper to host your New Zealand website outside New Zealand.
AAAAAAAAHHHHHHHHHHG!!!!
Even teh internets are leaving New Zealand
But I still won't vote for Shon Key ;-) -
Rich, they're hardly comparable. Iridium's network wasn't exactly something that could be rolled out in usable sections, was it? As for CDMA, Telecom made a stupid decision to go with a technology that was known to be on the way out. The decision had to have been made at least in part because it meant customers faced that much more difficulty and expense in shifting to BellSouth/Vodafone, because there's no sound technical explanation.
Steve, my take is that it's yet another wireless spec that failed to live up to the promises. So far none of the wireless protocols has delivered the headline speeds claimed when rolled into the real world, and that includes WiMax. They all fall short, often dramatically short, which is a real problem when trying to use them as fixed-line replacements. Right now 40Mbps is good enough, but it's not going to be good enough for long and the lead time on developing and deploying new protocols is often the better part of a decade. The 802.16m working group has also said that they're going to look to deploy into the military space first, and get that stable and functional before they try for consumers. That means at least another three years, probably more like five, before consumer equipment is available, and that's from when the spec is finalised which isn't expected to be until late next year.
-
Neither Iridium nor Telecoms CDMA could really be called networks. Rather protocols and associated hardware.
That's a good point. It's like saying that ethernet is a network. I hadn't thought of it like that, but you're right.
Post your response…
This topic is closed.