I've got a web site that discusses an alternative view of North American Numbering Plan expansion. The industry committee (no, consumers don't count) is leaning towards allowing 0/1 at the beginning of a prefix code, since it'll always be after an area code, and then stretching the area codes to 4 digits (using the second digit 9 as transition, or inserting 1 or 0 after the area code). I find this far from ideal.
North American Numbering Plan Forum presents my alternative. It ends up with 8-digit local dialing (4+4), which should be adequate for all but the largest metro areas (which will need two area codes). It also sorts the area codes into geographic, nongeographic/functional and an expanded freephone (800, etc.) space.
Implementation of such a change will take years, because there are necessary transition phases, the first of which is to move to mandatory 10-digit local dialing. Assume that that is inevitable; the only question is when. The plan I discuss has no flag days, plenty of "permissive" time for every stage of transition, and makes new numbers predictable.
The web site is a Delphi board, which means that it has its own discussion forum. I dont' know if it's a good idea to advertise it on/.:-) but hey it could use some activity!
You can't have variable (or, technically, nondeterministic) phone number length in the USA. It works in Germany because the network uses very different internal technology. German networks use "compelled signaling", wherein the terminating exchange asks for more digits until it is satisfied. American exchange use "en-bloc signaling", wherein the originating exchange collects the entire number and then passes it along. That is not going to change, and it requires the number length to be predictable by the originating exchange.
This happened because there are lots of rate centers (billing areas), and each rate center needs its own prefix (NXX) code. Now, with competition, each carrier needs its own NXX within each rate center that it's in. That eats up NXXs faster than Pac-Man.
There are technical solutions. NXXs can be shared among up to ten carriers by using hte next digit to identify carriers, so for instance 924-5 goes to VeriZontal and 924-6 goes to AT&T Local. Some states have ordered it, but some haven't, thus worsening things.
Even better, telephone numbers could be shared from a common pool, since they are already portable (can be moved to a different local carrier). A phone number is now like a DNS name, NOT an IP address. So a neutral pool registry could let carriers assing numbers from the common pool, and East Frog Dick (population 133) wouldn't need five prefix codes. BUT the FCC hasn't ordered that (the incumbent telcos are opposed to giving up their numbering advantages).
I agree that TCP/IP per se is a little bit tricky over high-speed wireless, but AX.25 is NOT the answer.
AX.25 is a variant of LAP-B, the X.25 layer 2 protocol. It is optimized for 300-1200 bps local links only, and if I felt like wasting the space, I could recite a litany of what's wrong with it. Indeed, AX.25 is even prone to congestion collapse, because it lacks even the most basic congestion control mechanisms. Been there, done that, back in the 1980s, on 2-meter amateur packet radio (both "raw" AX.25 and TCP/IP over AX.25, which btw usually used "unassured" mode, which worked better).
It is indeed possible to design a wireless subnetwork/datalink protocol that compensates for the problems of wireless. But AX.25 ain't it. I'm somewhat embarassed when I see commercial products still attempting to use it, although it's workable for the simple one-hop mobile dispatch application.
Most people don't understand how fragile TCP/IP or the Internet is. VERY fragile! The TCP congestion control algorithm is all that stands between it and total collapse. VoIP and other (video, for instance) streaming over UDP doesn't participate, so it keeps streaming away while TCP slows down and thus loses its share of bandwidth.
That's one reason why voice shouldn't be allowed to be a major share of Internet traffic: You can run voice over an "old-fashioned" five-nines phone network, but you can't run the Internet over it.
Yes, because there was no independent regulator. The PTT (Post, Telegraph and Telephone) was the ministry in charge, and they both ran the network and set the rules. They set their rates, terms and conditions. This is still the case in some countries, mostly smaller ones. And yes, some even regulate ISPs, or maintain monopolies over Internet service.
Ma Bell dealt with state commissions and the FCC. Most were docile, but at least gave a fig leaf for consumer protection. We gringos had it very, very good, phone-wise, compared to most of the world. Competition has caught on in a lot of countries (it's now basically mandatory in the E.U., for instance) and their service is catching up with ours. (And in the meantime, the entrenched RBOCs here are getting worse, as they fight competition by abusing their monopoly market power, not by improving.)
As a matter of federal policy since at least 1934 if not 1927 or earlier, basic telephone service has been treated, in the United States at least, as a basic human need. The method has been to subsidize local service via overpriced LD bills. That is economically inefficient and is being phased out, or at least phased down. But there are still taxes and cross-subsidies.
This was originally done for both altruistic and commercial means: Universal telephone service made the phone that much more valuable to business customers, who always paid above cost. The old network effect -- the value of each connection to a network rises with the number of connections. So old Ma Bell was doing well by doing good.
Poorer countries like the Phillipines don't have the widespread networks yet, and obviously need a better means of getting there than the old "tax LD to hell" routine. But it can't be done overnight. This is a capital-intensive business.
"Unlimited" calling plans exist only because regulators require them -- in most countries, where the telcos were owned by the govt. and thus not regulated, all calls were measured, usually at a price many times any reasonably-estimated cost. Note the huge controversies over dial-up Internet access in Europe, where it's very costly, though telcos (now privatized and regualted) are beginning to offer "unlimited" ISP access plans under pressure.
An American "unlimited" plan is actually a rate-averaged plan, wherein the price is supposed to cover the average local usage. Toll calls have always been charged for at a higher rate, again well above cost, in order to subsidize basic local service (the base monthly residential rate rarely covers cost; they make up for it via tolls, optional features, and much higher business-line local charges). That's done to promote "universal service".
If people use dial-up voice calls to access somebody who carries calls a long distance, then they're making long distance calls. It shouldn't matter whether the LD haul is coming via PCM fiber optic circuits, ancient analog microwave, the Internet, or modulated smoke signals. That's the LD carrier's business. Letting LD carriers use "the Internet" (which is NOT a clearly-defined term, and can be easily stretched to refer to semi-dedicated voice circuits) to carry voice, without paying the same as other LD carriers, is simply a way of subsidizing bad-quality carriers at the expense of good ones.
Note that if dial-up ISP calls become identified with LD, then it will be all the easier for the telco to demand toll charges for them. That's incredibly counterproductive.
In a country like the Phillipines, they haven't gotten as far as the USA has (not all that far!) in demonopolizing the phone business. So there is a real sensitivity to VoIP, which costs the local telco (PLDT) a lot of its international settlement revenue. And that will make it harder to provide basic service in what's basically a fairly low-income country, where most people can't even afford a phone.
Monopolies are generally bad and the old telcos made their own beds, but short-term disruption can hurt lots of people, even ISPs and their customers. In the long run this all shouldn't matter, but you have to be very sensitive to the economic interactions when an old monopoly faces competition in unexpected ways.
AT&T isn't going to adopt GSM voice where it now has IS-136. Not that IS-136 is much better, but a) you can run IS-136 on the US 800 MHz band and not GSM, thanks to the latter's screw-the-US design, and b) they're both obsolete.
CDMA is by far the better technology, using less battery power (because spreading lowers Rayleigh Fading) and more spectrum efficiency (because it allows adjacent-cell frequency reuse). W-CDMA is one mode of 3G but so is cdma2000, upward-compatible with IS-95. GSM was hurried to market because Europe needed something fast (its analog networks were in compatible, unlike US AMPS) while CDMA was based on later technology (really fancy silicon).
The UMTS club tried to keep Qualcomm *and the USA* out. They proposed a chipping rate that would just barely NOT fit into a USA 5 MHz frequency allocation. They ended up lowering it just a little bit but it's still too large, so there's a separate cdma2000-based 5 MHz air interface (3.68 Mcps) and an Ericsson one (3.88 Mcps, I think). How truly petty.
BTW, GSM is just as "closed" as CDMA. It's protected by a patent portfolio that is shared among its club of members, but outsiders have little advantage picking it over CDMA (and thus paying Qualcomm its patent royalties).
Note also that WAP and Imode are application-layer things, separate from air interface questions. "3G" means a higher bit rate, if you can afford to pay for it. WAP was optimized for <= 9600 bps on tiny phone screens. Imode is a bit more webbish, but designed for handheld screens too.
If anybody bothered to read the CNET article and/or research the topic, they'd understand that AT&T is no more adopting GSM than Morse code.
GSM is an obsolescent technology, a clever hack for the 1980s but terribly inefficient by today's standards. The GSM community is planning to migrate to new "3G" protocols, which are designed to accommodate GSM migration while using a CDMA-derived technology.
AT&T is planning to use EDGE (a data upgrade of their current IS-136 TDMA protocol) and migrate to new formats that will, in the 3G spirit, interwork better with GSM. And they've got the common sense to use imode rather than WAP, which is pretty awful. So in a while their sets will be data-compatible with where GSM networks will be. But that's NOT the same as adopting GSM in the latter's twilight of existence.
This isn't all so great. You have exactly two licenses for US satellite radio systems, so there are only two corporations who will choose the programming. That's even worse than terrestrial AM/FM radio, which has been horribly consolidated. So what crap will satradio have? The 70's pop rock channel, the 80's pop rock channel, the Don Ho channel?
Terrestrial digital radio is likewise only a couple of licensees, so it too will offer all the variety of McDonalds. The cheese and no cheese channels.
I don't think that Clarke has a great partner in this deal; he's probably being taken advantage of.
From http://www.theregister.co.uk/content/7/14971.html :
A 28-year-old man has fled the country to escape his creditors after his technology business collapsed around his ears.
Joe de Saram started his software company, Rhodium, a year ago with a loan of £2500. The company specialised in banking software and encryption technology.
At the height of the technology boom he was worth a cool £25 million. He drove a Ferrari 355 F1 and was the 62nd richest Asian in the UK.
He had offices in Sheffield and London and was planning to launch an online bank and share trading system. His company name was changed to "I Love My Encryption Technology".
But as the dotcom bubble deflated, his company ran into financial difficulties, and was finally wound up in Leeds Registry Court.
Lawyers acting for Saram's creditors said that the young ex-millionaire was thought to be in Sri Lanka, having been traced via his mobile phone.
One creditor told London freebie paper Metro that he was quite a character. She said: "There are all sorts of stories and rumours circulating about him. People are even saying that the Tamil Tigers are after him."
Leeds county court said an official liquidator will be appointed within five
days of the winding up. ®
Hurd is based on Mach, so it is full of context changes. Very slow on some machines, especially RISCy ones. Thus I don't suspect it'll ever be a good performer.
Chorus apparently includes message-passing within its microkernel, substantially reducing the number of context switches. That should greatly improve its performance doing network-y things. Sun is selling Chorus now as a real-time system. It has Solaris compatibility (to simplify development) but lets you build tight small-code embedded systems. I don't know what the license is on the Chorus microkernel itself, before you buy the whole ChorusOS from Sun.
>So why wouldn't it be considered a monopoly for the cable company?
Because under the federal law (47CFR), Cable companies are notcommon carriers, while telephone companies are. Cable companies do not, as others have noted, have statutory monopolies; it's just tough to be the second one in town.
Since telephone companies are common carriers, they are obligated to provide access to anybody. So you can use DSL to reach any connected ISP, or dial-up, or leased telco lines. But cablecos are under a different law.
That they provide anything but video is a bonus. If you impose excessive obligations on t hem, they just won't do data at all. Better to let them learn how to be "open" before imposing premature regulation on them.
So you're saying that because of mobile support, every packet has to get about 40 bytes larger, thereby raising traffic on the backbones, LANs and everybody else's networks? Hardly a good idea!
I do note that "Class A" address space 64-126 was never issued, so a LOT of CIDR blocks can be released there.
While I'm not the final expert on these matters, I suspect that the opposite of what you suggest is true. CISC machines tend to be much better at context switches. It does depend on the machine... but the VAX (about as CISC as you can get) could task-switch in a handful of cycles, while some RISC machines take hundreds.
RISC machines are in general really fast at big monolithic tasks, like number crunching, not at task-switching. So if you go to a microkernel that needs to do a lot of context switches, RISC performance will probably be bad.
If the microkernel had better message-passing within it (I'm told Chorus is good at this), then the frequent task switching becomes unnecessary and performance improves. But Mach got it wrong.
I've seen this story elsewhere, and while I haven't seen the AOL 6 software in use yet, at least one article contradicts the assumption that underlist most of the comments here.
It appears that AOL changed the <i>manner</i> in which one's home page is set. It used to be very easy (so easy, even an AOLer could do it!:-) ) but now it is changed by going in to some preferences menu or such. In other words, it's not fixed, just not saying "change me" so loud.
What's scary is the way some people use this minor change to call for governmental intervention. Do slashdotters really want the government to be regulating the design of software? (The MS antitrust case is utterly different, starting with its monopoly status, but it also wasn't about regulating the code per se.) Do slashdotters really want to government to be regulating Internet content providers?
AOL doesn't own wires. They are basically a time-sharing house with a clever display technology. All of the non-web content is hosted in a big data center in Virginia. That's not even a classical ISP model. It's content. They lease their modems from UUNET, Genuity and Sprint. Like them or hate them, asking to regulate them is asking for trouble.
> IPv4 over MPLS can only get you QoS while you're running over MPLS, which most links aren't. IPv6 carries the flow label in the header, so you can perform QoS end-to-end.
No. First off, end-to-end QoS is a misnomer; if a router in the middle drops or delays a packet, 'tain't nuttin' the ends can do to make it arrive on time. Second, the flow label in IPv6 is too short to be globally significant, so it's basically useless to the intermediate routers; they will still need MPLS!
NATs cause problems but solve others (like security; some corporate nets use them even with ample space just for that reason). If you want to address every shelf of your refrigerator, I'd think a NAT address would be adequate. IPv6 does offer more address space, but it's not well thought out; TUBA does it better.
The 64-126 space, assigned via CIDR, could do quite a bit to preserve IPv4 until something WORTHWHILE is done. The cost of IPv6 is just too high for the few benefits that it actually delivers.
IPv6 and 3G Wireless are perfect for each other because both are blind alleys, technologies that are vastly oversold. Neither is half as good as its supporters imply, and IPv6 itself is a huge loser.
Cut to the chase: The main reasons that the referenced article states that IPv6 is useful are 1) security, and 2) end-to-end voice. NEITHER is materially better in v6 than in v4! IPSec (which has its own problems) works in v4; v6 doesn't make it more secure. And contrary to the bushian Big Lies going around these days, IPv6 does NOTHING for QoS by itself! It got flows wrong too. You can run IPv4 over MPLS and get QoS; IPv6 by itself doesn't get you QoS. QED.
IPv6 is a bad protocol, a political compromise from the NSFnet days of yore, which has a high cost of adoption and few benefits. The IPv4 address space ain't done yet -- first off, NATs solve a lot of problems. Second, the former "class A" space 64-126 is still reserved; it could free up a LOT of CIDR space, if ICANN weren't playing games.
TUBA was a better proposal, but the IETF dropped it at the last minute because it was tainted by the whiff of OSI -- kind of a redbaiting approach to protocol design. What's needed now isn't IPv6 but a rethought Internet layer, one that's not devoted to fighting the religious battles of the 1970s and 1980s, but one that's designed to carry diverse applications with reasonable overhead.
And wireless? Well, 3G has more bandwidth, but not cheap bandwidth -- look up what the spectrum auctions netted and figure out how much you'll need to pay to cover that bill! It's fine for voice, but don't expect to watch on-demand-TV on your cellphone's screen, unless you're willing to pay macro {dollars,euros} per minute. Now the LAST think wireless (land of costly bandwidth) needs is a huge wasteful header overhead! IPv6 assumes basically free bandwidth. Wireless needs its overhead the way the Sahara needs more sand.
The original article was utterly full of hyperbole, but it's not quite as bad as you make it sound.
By now, the vast majority of telco central offices, even rural ones, have fiber optic connectivity. I think even hapless old VeriZontal/New England Tel has glass to their COs everywhere. BUT it rarely goes any farther. It's the local loop, to the subscriber, that's always copper. Sure, glass loops exist for businesses that require DS3+ (45Mbps) or multiple T1s, but the entry cost is indeed high.
"Fiber to the home" was a big catch phrase a decade ago, but almost dead now. They're trying to make DSL do the job, which it often can't (because the old copper was installed for voice and more often than not can't carry DSL).
But even then, the Cogent analysis is wrong: Telcos will bring you T1 for under $150/mo (from their CO), often much less. Its ISP fees that are higher, charges that are <i>above</i> the loop cost, and ISPs lose money as it is.
Flat out untrue. The Cable Act of 1992 forbade local governments from giving any monopoly franchises, not that most had official monopolies before then.
Ever wonder why some cities (like L.A.) have such a crazy-quilt of cable areas? No monopoly! Each of several non-exclusive franchisees just built until they met somebody else's buildout. They didn't have to stop, but it's a hell of a lot more profitable to be the first on a block than the second, so they logically avoided each other.
Telcos had monopolies enforced in most states until 1996. Very different.
The original article has a decidedly anti-cableco bias. While TWE has obviously got a lot of learning to do about dealing with ISPs, there's much to be said about cable modems:
* Cablecos are not, under law, common carriers, so they don't have to carry anything they don't want to. Congress said so.
* Telephone companies are common carriers. Under the law, they have to serve all comers, even ISPs. The incumbent local telcos (ILECs) are trying to avoid their obligations.
* The above situation is not so unfair. Telcos basically got guaranteed returns on their monopolies, while cablecos haven't, and have been much more marginal, financially. And cablecos aren't legally monopolies (nor are telcos anymore); it's just hard to have more than one cableco in a city. RCN and Knology are trying to make a go of "overbuilding".
* The cablecos got into cable modems a few years ago by having third parties (@Home and Roadrunner, mostly) set it up for them. They know entertainment, not data, so they were giving the public an alternative source of Internet access (not using telcos) and should be thanked for that.
* There was no "open access" movement then; the cablecos were basically creating self-provisioned ISPs.
* Now that the cablecos are successful at selling modems, dial ISPs are feeling competition and want in. Understandable. But remember, cablecos are not common carriers, telcos are.
* AOL charges $10 for "bring your own" access, which is 25% of the typical $40/month cable modem fee. Of the $40, probably less than $10 goes to the backbone and to run the servers which @Home provides.
TW's proposed revenue-sharing model is obviously wrong. They should be selling the pipe for a fee, probably around $30/month (which is the range of wholesale ADSL), content neutral. The homepage crap follows from the wrong model. It is in all cablecos best interests to have the ISPs use them, rather than telcos, for their access pipes. But the cablecos are just learning about selling bandwidth, rather than content. It's a foreign concept to them. Don't blast them for trying. They will probably get it right the third or fourth time around.
The Alpha motherboard is close, but probably won't win the cigar.
Motherboards include ROMs, for things like setup, BIOS and booting up in the first place. Alpha uses a different instruction set, so a 386-instruction chip like Athlon wouldn't read it.
The mobo vendor could probably do an AMD port of the board. API is unlikely to since their job is to sell Alpha processors, not mobos. Somebody else with an Alpha SMP mobo would be a better candidate.
Slashdot Mirror
North American Numbering Plan Forum presents my alternative. It ends up with 8-digit local dialing (4+4), which should be adequate for all but the largest metro areas (which will need two area codes). It also sorts the area codes into geographic, nongeographic/functional and an expanded freephone (800, etc.) space.
Implementation of such a change will take years, because there are necessary transition phases, the first of which is to move to mandatory 10-digit local dialing. Assume that that is inevitable; the only question is when. The plan I discuss has no flag days, plenty of "permissive" time for every stage of transition, and makes new numbers predictable.
The web site is a Delphi board, which means that it has its own discussion forum. I dont' know if it's a good idea to advertise it on /. :-) but hey it could use some activity!
You can't have variable (or, technically, nondeterministic) phone number length in the USA. It works in Germany because the network uses very different internal technology. German networks use "compelled signaling", wherein the terminating exchange asks for more digits until it is satisfied. American exchange use "en-bloc signaling", wherein the originating exchange collects the entire number and then passes it along. That is not going to change, and it requires the number length to be predictable by the originating exchange.
They ran out of prefix codes.
This happened because there are lots of rate centers (billing areas), and each rate center needs its own prefix (NXX) code. Now, with competition, each carrier needs its own NXX within each rate center that it's in. That eats up NXXs faster than Pac-Man.
There are technical solutions. NXXs can be shared among up to ten carriers by using hte next digit to identify carriers, so for instance 924-5 goes to VeriZontal and 924-6 goes to AT&T Local. Some states have ordered it, but some haven't, thus worsening things.
Even better, telephone numbers could be shared from a common pool, since they are already portable (can be moved to a different local carrier). A phone number is now like a DNS name, NOT an IP address. So a neutral pool registry could let carriers assing numbers from the common pool, and East Frog Dick (population 133) wouldn't need five prefix codes. BUT the FCC hasn't ordered that (the incumbent telcos are opposed to giving up their numbering advantages).
I agree that TCP/IP per se is a little bit tricky over high-speed wireless, but AX.25 is NOT the answer.
AX.25 is a variant of LAP-B, the X.25 layer 2 protocol. It is optimized for 300-1200 bps local links only, and if I felt like wasting the space, I could recite a litany of what's wrong with it. Indeed, AX.25 is even prone to congestion collapse, because it lacks even the most basic congestion control mechanisms. Been there, done that, back in the 1980s, on 2-meter amateur packet radio (both "raw" AX.25 and TCP/IP over AX.25, which btw usually used "unassured" mode, which worked better).
It is indeed possible to design a wireless subnetwork/datalink protocol that compensates for the problems of wireless. But AX.25 ain't it. I'm somewhat embarassed when I see commercial products still attempting to use it, although it's workable for the simple one-hop mobile dispatch application.
Amen!
Most people don't understand how fragile TCP/IP or the Internet is. VERY fragile! The TCP congestion control algorithm is all that stands between it and total collapse. VoIP and other (video, for instance) streaming over UDP doesn't participate, so it keeps streaming away while TCP slows down and thus loses its share of bandwidth.
That's one reason why voice shouldn't be allowed to be a major share of Internet traffic: You can run voice over an "old-fashioned" five-nines phone network, but you can't run the Internet over it.
>Owned by the government means not regulated?
Yes, because there was no independent regulator. The PTT (Post, Telegraph and Telephone) was the ministry in charge, and they both ran the network and set the rules. They set their rates, terms and conditions. This is still the case in some countries, mostly smaller ones. And yes, some even regulate ISPs, or maintain monopolies over Internet service.
Ma Bell dealt with state commissions and the FCC. Most were docile, but at least gave a fig leaf for consumer protection. We gringos had it very, very good, phone-wise, compared to most of the world. Competition has caught on in a lot of countries (it's now basically mandatory in the E.U., for instance) and their service is catching up with ours. (And in the meantime, the entrenched RBOCs here are getting worse, as they fight competition by abusing their monopoly market power, not by improving.)
As a matter of federal policy since at least 1934 if not 1927 or earlier, basic telephone service has been treated, in the United States at least, as a basic human need. The method has been to subsidize local service via overpriced LD bills. That is economically inefficient and is being phased out, or at least phased down. But there are still taxes and cross-subsidies.
This was originally done for both altruistic and commercial means: Universal telephone service made the phone that much more valuable to business customers, who always paid above cost. The old network effect -- the value of each connection to a network rises with the number of connections. So old Ma Bell was doing well by doing good.
Poorer countries like the Phillipines don't have the widespread networks yet, and obviously need a better means of getting there than the old "tax LD to hell" routine. But it can't be done overnight. This is a capital-intensive business.
An American "unlimited" plan is actually a rate-averaged plan, wherein the price is supposed to cover the average local usage. Toll calls have always been charged for at a higher rate, again well above cost, in order to subsidize basic local service (the base monthly residential rate rarely covers cost; they make up for it via tolls, optional features, and much higher business-line local charges). That's done to promote "universal service".
If people use dial-up voice calls to access somebody who carries calls a long distance, then they're making long distance calls. It shouldn't matter whether the LD haul is coming via PCM fiber optic circuits, ancient analog microwave, the Internet, or modulated smoke signals. That's the LD carrier's business. Letting LD carriers use "the Internet" (which is NOT a clearly-defined term, and can be easily stretched to refer to semi-dedicated voice circuits) to carry voice, without paying the same as other LD carriers, is simply a way of subsidizing bad-quality carriers at the expense of good ones.
Note that if dial-up ISP calls become identified with LD, then it will be all the easier for the telco to demand toll charges for them. That's incredibly counterproductive.
In a country like the Phillipines, they haven't gotten as far as the USA has (not all that far!) in demonopolizing the phone business. So there is a real sensitivity to VoIP, which costs the local telco (PLDT) a lot of its international settlement revenue. And that will make it harder to provide basic service in what's basically a fairly low-income country, where most people can't even afford a phone.
Monopolies are generally bad and the old telcos made their own beds, but short-term disruption can hurt lots of people, even ISPs and their customers. In the long run this all shouldn't matter, but you have to be very sensitive to the economic interactions when an old monopoly faces competition in unexpected ways.
Typical Euro-centric blather.
AT&T isn't going to adopt GSM voice where it now has IS-136. Not that IS-136 is much better, but a) you can run IS-136 on the US 800 MHz band and not GSM, thanks to the latter's screw-the-US design, and b) they're both obsolete.
CDMA is by far the better technology, using less battery power (because spreading lowers Rayleigh Fading) and more spectrum efficiency (because it allows adjacent-cell frequency reuse). W-CDMA is one mode of 3G but so is cdma2000, upward-compatible with IS-95. GSM was hurried to market because Europe needed something fast (its analog networks were in compatible, unlike US AMPS) while CDMA was based on later technology (really fancy silicon).
The UMTS club tried to keep Qualcomm *and the USA* out. They proposed a chipping rate that would just barely NOT fit into a USA 5 MHz frequency allocation. They ended up lowering it just a little bit but it's still too large, so there's a separate cdma2000-based 5 MHz air interface (3.68 Mcps) and an Ericsson one (3.88 Mcps, I think). How truly petty.
BTW, GSM is just as "closed" as CDMA. It's protected by a patent portfolio that is shared among its club of members, but outsiders have little advantage picking it over CDMA (and thus paying Qualcomm its patent royalties).
Note also that WAP and Imode are application-layer things, separate from air interface questions. "3G" means a higher bit rate, if you can afford to pay for it. WAP was optimized for <= 9600 bps on tiny phone screens. Imode is a bit more webbish, but designed for handheld screens too.
If anybody bothered to read the CNET article and/or research the topic, they'd understand that AT&T is no more adopting GSM than Morse code.
GSM is an obsolescent technology, a clever hack for the 1980s but terribly inefficient by today's standards. The GSM community is planning to migrate to new "3G" protocols, which are designed to accommodate GSM migration while using a CDMA-derived technology.
AT&T is planning to use EDGE (a data upgrade of their current IS-136 TDMA protocol) and migrate to new formats that will, in the 3G spirit, interwork better with GSM. And they've got the common sense to use imode rather than WAP, which is pretty awful. So in a while their sets will be data-compatible with where GSM networks will be. But that's NOT the same as adopting GSM in the latter's twilight of existence.
Ken Olsen, founder of DEC, was once asked how many people worked there.
"About half."
This isn't all so great. You have exactly two licenses for US satellite radio systems, so there are only two corporations who will choose the programming. That's even worse than terrestrial AM/FM radio, which has been horribly consolidated. So what crap will satradio have? The 70's pop rock channel, the 80's pop rock channel, the Don Ho channel?
Terrestrial digital radio is likewise only a couple of licensees, so it too will offer all the variety of McDonalds. The cheese and no cheese channels.
I don't think that Clarke has a great partner in this deal; he's probably being taken advantage of.
From http://www.theregister.co.uk/content/7/14971.html :
A 28-year-old man has fled the country to escape his creditors after his technology business collapsed around his ears.
Joe de Saram started his software company, Rhodium, a year ago with a loan of £2500. The company specialised in banking software and encryption technology.
At the height of the technology boom he was worth a cool £25 million. He drove a Ferrari 355 F1 and was the 62nd richest Asian in the UK.
He had offices in Sheffield and London and was planning to launch an online bank and share trading system. His company name was changed to "I Love My Encryption Technology".
But as the dotcom bubble deflated, his company ran into financial difficulties, and was finally wound up in Leeds Registry Court.
Lawyers acting for Saram's creditors said that the young ex-millionaire was thought to be in Sri Lanka, having been traced via his mobile phone.
One creditor told London freebie paper Metro that he was quite a character. She said: "There are all sorts of stories and rumours circulating about him. People are even saying that the Tamil Tigers are after him."
Leeds county court said an official liquidator will be appointed within five
days of the winding up. ®
Hurd is based on Mach, so it is full of context changes. Very slow on some machines, especially RISCy ones. Thus I don't suspect it'll ever be a good performer.
Chorus apparently includes message-passing within its microkernel, substantially reducing the number of context switches. That should greatly improve its performance doing network-y things. Sun is selling Chorus now as a real-time system. It has Solaris compatibility (to simplify development) but lets you build tight small-code embedded systems. I don't know what the license is on the Chorus microkernel itself, before you buy the whole ChorusOS from Sun.
>So why wouldn't it be considered a monopoly for the cable company? Because under the federal law (47CFR), Cable companies are not common carriers, while telephone companies are. Cable companies do not, as others have noted, have statutory monopolies; it's just tough to be the second one in town. Since telephone companies are common carriers, they are obligated to provide access to anybody. So you can use DSL to reach any connected ISP, or dial-up, or leased telco lines. But cablecos are under a different law. That they provide anything but video is a bonus. If you impose excessive obligations on t hem, they just won't do data at all. Better to let them learn how to be "open" before imposing premature regulation on them.
So you're saying that because of mobile support, every packet has to get about 40 bytes larger, thereby raising traffic on the backbones, LANs and everybody else's networks? Hardly a good idea!
I do note that "Class A" address space 64-126 was never issued, so a LOT of CIDR blocks can be released there.
While I'm not the final expert on these matters, I suspect that the opposite of what you suggest is true. CISC machines tend to be much better at context switches. It does depend on the machine... but the VAX (about as CISC as you can get) could task-switch in a handful of cycles, while some RISC machines take hundreds.
RISC machines are in general really fast at big monolithic tasks, like number crunching, not at task-switching. So if you go to a microkernel that needs to do a lot of context switches, RISC performance will probably be bad.
If the microkernel had better message-passing within it (I'm told Chorus is good at this), then the frequent task switching becomes unnecessary and performance improves. But Mach got it wrong.
I've seen this story elsewhere, and while I haven't seen the AOL 6 software in use yet, at least one article contradicts the assumption that underlist most of the comments here.
:-) ) but now it is changed by going in to some preferences menu or such. In other words, it's not fixed, just not saying "change me" so loud.
It appears that AOL changed the <i>manner</i> in which one's home page is set. It used to be very easy (so easy, even an AOLer could do it!
What's scary is the way some people use this minor change to call for governmental intervention. Do slashdotters really want the government to be regulating the design of software? (The MS antitrust case is utterly different, starting with its monopoly status, but it also wasn't about regulating the code per se.) Do slashdotters really want to government to be regulating Internet content providers?
AOL doesn't own wires. They are basically a time-sharing house with a clever display technology. All of the non-web content is hosted in a big data center in Virginia. That's not even a classical ISP model. It's content. They lease their modems from UUNET, Genuity and Sprint. Like them or hate them, asking to regulate them is asking for trouble.
Assuming the summary is correct...
Finance Open Source projects through licensing fees?
Further proof that the Eurocrats are basically nitwits who couldn't get work in their own countries!
> IPv4 over MPLS can only get you QoS while you're running over MPLS, which most links aren't. IPv6 carries the flow label in the header, so you can perform QoS end-to-end.
No. First off, end-to-end QoS is a misnomer; if a router in the middle drops or delays a packet, 'tain't nuttin' the ends can do to make it arrive on time. Second, the flow label in IPv6 is too short to be globally significant, so it's basically useless to the intermediate routers; they will still need MPLS!
NATs cause problems but solve others (like security; some corporate nets use them even with ample space just for that reason). If you want to address every shelf of your refrigerator, I'd think a NAT address would be adequate. IPv6 does offer more address space, but it's not well thought out; TUBA does it better.
The 64-126 space, assigned via CIDR, could do quite a bit to preserve IPv4 until something WORTHWHILE is done. The cost of IPv6 is just too high for the few benefits that it actually delivers.
IPv6 and 3G Wireless are perfect for each other because both are blind alleys, technologies that are vastly oversold. Neither is half as good as its supporters imply, and IPv6 itself is a huge loser.
Cut to the chase: The main reasons that the referenced article states that IPv6 is useful are 1) security, and 2) end-to-end voice. NEITHER is materially better in v6 than in v4! IPSec (which has its own problems) works in v4; v6 doesn't make it more secure. And contrary to the bushian Big Lies going around these days, IPv6 does NOTHING for QoS by itself! It got flows wrong too. You can run IPv4 over MPLS and get QoS; IPv6 by itself doesn't get you QoS. QED.
IPv6 is a bad protocol, a political compromise from the NSFnet days of yore, which has a high cost of adoption and few benefits. The IPv4 address space ain't done yet -- first off, NATs solve a lot of problems. Second, the former "class A" space 64-126 is still reserved; it could free up a LOT of CIDR space, if ICANN weren't playing games.
TUBA was a better proposal, but the IETF dropped it at the last minute because it was tainted by the whiff of OSI -- kind of a redbaiting approach to protocol design. What's needed now isn't IPv6 but a rethought Internet layer, one that's not devoted to fighting the religious battles of the 1970s and 1980s, but one that's designed to carry diverse applications with reasonable overhead.
And wireless? Well, 3G has more bandwidth, but not cheap bandwidth -- look up what the spectrum auctions netted and figure out how much you'll need to pay to cover that bill! It's fine for voice, but don't expect to watch on-demand-TV on your cellphone's screen, unless you're willing to pay macro {dollars,euros} per minute. Now the LAST think wireless (land of costly bandwidth) needs is a huge wasteful header overhead! IPv6 assumes basically free bandwidth. Wireless needs its overhead the way the Sahara needs more sand.
The original article was utterly full of hyperbole, but it's not quite as bad as you make it sound.
By now, the vast majority of telco central offices, even rural ones, have fiber optic connectivity. I think even hapless old VeriZontal/New England Tel has glass to their COs everywhere. BUT it rarely goes any farther. It's the local loop, to the subscriber, that's always copper. Sure, glass loops exist for businesses that require DS3+ (45Mbps) or multiple T1s, but the entry cost is indeed high.
"Fiber to the home" was a big catch phrase a decade ago, but almost dead now. They're trying to make DSL do the job, which it often can't (because the old copper was installed for voice and more often than not can't carry DSL).
But even then, the Cogent analysis is wrong: Telcos will bring you T1 for under $150/mo (from their CO), often much less. Its ISP fees that are higher, charges that are <i>above</i> the loop cost, and ISPs lose money as it is.
Flat out untrue. The Cable Act of 1992 forbade local governments from giving any monopoly franchises, not that most had official monopolies before then.
Ever wonder why some cities (like L.A.) have such a crazy-quilt of cable areas? No monopoly! Each of several non-exclusive franchisees just built until they met somebody else's buildout. They didn't have to stop, but it's a hell of a lot more profitable to be the first on a block than the second, so they logically avoided each other.
Telcos had monopolies enforced in most states until 1996. Very different.
The original article has a decidedly anti-cableco bias. While TWE has obviously got a lot of learning to do about dealing with ISPs, there's much to be said about cable modems:
* Cablecos are not, under law, common carriers, so they don't have to carry anything they don't want to. Congress said so.
* Telephone companies are common carriers. Under the law, they have to serve all comers, even ISPs. The incumbent local telcos (ILECs) are trying to avoid their obligations.
* The above situation is not so unfair. Telcos basically got guaranteed returns on their monopolies, while cablecos haven't, and have been much more marginal, financially. And cablecos aren't legally monopolies (nor are telcos anymore); it's just hard to have more than one cableco in a city. RCN and Knology are trying to make a go of "overbuilding".
* The cablecos got into cable modems a few years ago by having third parties (@Home and Roadrunner, mostly) set it up for them. They know entertainment, not data, so they were giving the public an alternative source of Internet access (not using telcos) and should be thanked for that.
* There was no "open access" movement then; the cablecos were basically creating self-provisioned ISPs.
* Now that the cablecos are successful at selling modems, dial ISPs are feeling competition and want in. Understandable. But remember, cablecos are not common carriers, telcos are.
* AOL charges $10 for "bring your own" access, which is 25% of the typical $40/month cable modem fee. Of the $40, probably less than $10 goes to the backbone and to run the servers which @Home provides.
TW's proposed revenue-sharing model is obviously wrong. They should be selling the pipe for a fee, probably around $30/month (which is the range of wholesale ADSL), content neutral. The homepage crap follows from the wrong model. It is in all cablecos best interests to have the ISPs use them, rather than telcos, for their access pipes. But the cablecos are just learning about selling bandwidth, rather than content. It's a foreign concept to them. Don't blast them for trying. They will probably get it right the third or fourth time around.
The Alpha motherboard is close, but probably won't win the cigar.
Motherboards include ROMs, for things like setup, BIOS and booting up in the first place. Alpha uses a different instruction set, so a 386-instruction chip like Athlon wouldn't read it.
The mobo vendor could probably do an AMD port of the board. API is unlikely to since their job is to sell Alpha processors, not mobos. Somebody else with an Alpha SMP mobo would be a better candidate.