how-to block ads
|
Share Topic
 |
 |
|
|
| | Re: caps? said by neufuse:So even fiber carriers have caps in canada? This is starting to get ridicilous. A point-to-point link between two points costs practically the same to operate at 1Mbps, 10Mbps, 100Mbps or even 1Gbps. As long as subscribers do not use their internet, you can have a million subscribers at 10Gbps hooked to your 1Gbps aggregation network to your single 100Mbps Cogentco internet feed. When subscribers start stressing the network though, it has to be overbuilt to handle substantial usage spikes and this easily doubles operating costs. Fibers are not magically immune to network load and those traffic spikes get nastier as end-user speeds get faster.
Unlimited is only an illusion, the bandwidth between two points on a network is finite and ISPs over-subscribe their links to keep their costs under control. Any ISP claiming affordable unlimited internet access would crumble under infrastructure and transit expenditures if a large chunk (maybe 20%) of its subscriber base decided to put the claim under an extended (months) burn-in test. The premise of "Unlimited" is that the ISP is charging a large enough premium to easily cover the typical transit cost of its "Unlimited" subscribers and that those subscribers will still use the internet intermittently. In general, abnormally high usage accounts are effectively subsidized by lower-usage accounts.
Caps are also somewhat flawed in that they require subscribers to buy a speed+cap package even though the subscriber may only be interested in only one of the two components, forcing people who are only interested in getting what little they do done faster to also buy the larger cap or those who want to download more but don't care about the speed to buy into the faster speed to get the cap. IMO, speed and caps should be sold as two separate components of broadband access... or perhaps speed should be dropped altogether as a differentiator since the cost to provide access on fiber/ADSL2 is the per-port cost regardless of artificially imposed rate limits.
Personally, I like the idea of usage-based billing since it would mean people only pay a monthly fee to cover fixed operating costs (possibly including a small base cap) and then contribute a supplemental fee proportional to their contribution to the network's load. Under a metered scheme, an ISP could simply connect everyone to the maximum speed possible (say 100M symmetrical) for $35/month and charge $0.10/GB afterward.
People are hypocrites: they want competition but do not want to pay for it. Not surprising that so few are willing to take the plunge and risk competing against the telco/cableco's regional monopolies. | |
|  | | Re: caps? All this makes sense... However, Moore's law dictates that we should see a doubling every 2 years. We are seeing the oposite trend in this industry.
Generally, the high degree of competition forces companies to move forward or die. The entrenched nature of the Telco/Cable co Duopoly (which is what this article is about), are not keeping pace with developments.
Enter a player who actually wants to compete for business and Canadians get a response like this from the encumbant. | |
| | | | Re: caps? said by backness:All this makes sense... However, Moore's law dictates that we should see a doubling every 2 years. We are seeing the oposite trend in this industry. Moore's pseudo-law is about transistor count per chip. Modal dispersion in fiber has been limiting practical modulation speeds to about 10Gbps on kilometric links for decades, Moore's law cannot do a thing about that, this is why optical 100GbE uses 10WDM... but then, you need wide-spectrum fiber which is a relatively new development so there is fairly little of it out there.
Another brick wall comes from the fact that the lowest-loss fibers used for long distances (10-100km) are finely engineered for a handful (1-3) of industry-standard wavelengths, those fibers would be unlikely to ever support more than 40Gbps. Replacing millions of kilometers worth of this stuff will cost a fortune so carriers are in no hurry to do so.
Fiber breakthroughs enable a tenfold improvement about once per decade, they do not provide much incremental improvement in-between. At the other end of the spectrum, you see the same trends in enterprise and home networks: about 10 years of 10Base-T/TX or worse, a solid 12+ years of 100Base-TX with GbE about to become mainstream, probably 10+ years of GbE to come... networking in general is on a ~10 years tenfold upgrade cycle. | |
| | | | | Re: caps? Unfortunately, its not called moore's theory, It's called a law because it has been proven to be valid. Also it relates to many other types of technological innovation as well.
I think you have missed the issue here. Moore's law is valid in areas where the is a competitive marketplace (ie. capacity, transistors, ect.).
The situation that we face in Canada is the opposite. We should call it the Bell law. You get your services chopped in half each year until the price doubles. This can only happen to a captive audience who is forced to bend over and take it and certainly not in a competitive marketplace such as the manufacture of circuit boards or other electronics. | |
| | | | | | Re: caps? said by backness:I think you have missed the issue here. Moore's law is valid in areas where the is a competitive marketplace (ie. capacity, transistors, ect.). Moore's law is only a factoid, not an actual law. The "Cult of Moore's Law" is something that simply came into being as more people observed the tendency.
Moore's "law" pretty much only applies to goods driven by transistor counts, most non-silicon-bound electronic goods are not subject to this 'law':
• average TV size has only doubled over the last 15 or so years
• 42" LCDs have been stagnating around the $1500 price point roughly four years
• cars today are not faster, not cheaper nor drastically more fuel-efficient than similar cars from 20+ years ago
• household appliances have remained largely unchanged over the last 50+ years
• houses are shrinking and their price is increasing
• food is not more nutritious nor cheaper either
• etc.
Aside from high-tech silicon-constrained gizmos, most goods and services are driven by other practical concerns and could not care less about Moore. In the case of long-haul optical links, photonics is the limiting factor and photonics is on a slower technological curve than silicon.
As for Bell and other carriers, these guys have regional monopolies and this does enable them to leverage technological progress to increase their profit margins instead of improving services. | |
| | | | | | | Re: caps? Firstly, Moore's law is called a law becuase it has been proven. It would be called Moore's theory if it had not been proven.
Secondly, the size of the device has nothing to do with Moore's law.
Thirdly, 42"LCDs have not entered the mass market and therefore will not be able to apply the economies of scale principle which is a major component of the Law. I can keep going if you like....
Here is some reading for you: (Although I incorectly Applied Moore's law, it should be Buttler's law of photonics instead)
Network capacity According to Gerry/Gerald Butters,[21][22] the former head of Lucent's Optical Networking Group at Bell Labs, there is another version, called Butter's Law of Photonics,[23] a formulation which deliberately parallels Moore's law. Butter's law[24] says that the amount of data coming out of an optical fiber is doubling every nine months. Thus, the cost of transmitting a bit over an optical network decreases by half every nine months. The availability of wavelength-division multiplexing (sometimes called "WDM") increased the capacity that could be placed on a single fiber by as much as a factor of 100. Optical networking and DWDM is rapidly bringing down the cost of networking, and further progress seems assured. As a result, the wholesale price of data traffic collapsed in the dot-com bubble. Nielsen's Law says that the bandwidth available to users increases by 50% annually.[25] | |
| | | | | | | | Re: caps? said by backness:Thirdly, 42"LCDs have not entered the mass market and therefore will not be able to apply the economies of scale principle which is a major component of the Law. I can keep going if you like.... My point was that "Moore's Law" does not apply beyond silicon-constrained devices. As soon as you take a step beyond computerized gadgetry and computers, mature market laws apply.
The digital semiconductor market is still at least five years away from maturity. Moore cultists panicked when manufacturers ran into snags at 90nm, they will likely have a few more chances to do so again along the way to 12nm... the end of the practical process shrink road is getting close. Once we get there, Moore's Law will be history. | |
|
| |
|
