Simson's right in denying IPv6's short-term inevitability, but he's still being too easy on it! IPv6 is just plain dumb. He should say it.
IPv6 creates much larger headers, so there's more overhead, particularly, as a percentage, on short packets (voice, ACK's, etc.). So it'll waste bandwidth, or lower effective throughput on fixed bandwidths. We need this? It is not even using its 128 bits efficiently. The general approach is to use the top half to identify the network and the bottom half to include the 48-bit MAC address of the computer. That was a clever hack in 1985 when proposed for DECnet Phase V (which never caught on) and became an approach in OSI CLNP. But that was not for a public spammer-ridden insecure Internet. Now it is a security and privacy hole to do that. It also means the 128 bits are not used efficiently -- we are tight with 32 bits, but an address for every atom?
IPv6 also does nothing for QoS (ignore the hype, which is based on a misunderstanding) and nothing for security (IPsec works just fine with v4). It just wastes bandwidth. So it does something for, oh, MCI. No wonder Vint (the Chauncey Gardner of the Internet) likes it! And Sprint, AT&T and VeriZontal. Great.
IPv4 could use a decent replacement some day, but IPv6 is everything you don't like about v4, and more. Eccch. A dozen years since it was "adopted" and it's gone nowhere, for good reason. The Asians weren't so involved with IETF at the time, to know the messy politics behind it. And btw the whole thing about their not having addresses is false; there is plenty of space left in the IPv4 space waiting to be allocated where needed. China can have more, as they provide more and more spam relays for the h3rb@1-v14gr4 crowd.
The problem with VoIP is that it isn't half as good as people think it is -- there are certainly good niche applications, and ways to use it profitably, but it simply isn't the be-all and end-all. Why do people fawn over it so much? I think it's largely because "IP" has that "k3w1" quality of the Internet in general, while phones are passe -- hardly a good way to make rational decisions.
This paper is pretty useful: http://klamath.stanford.edu/~nickm/papers /HotNets0 2-IP_conquest_of_the_world_with_authors.pdf
In the meantime, VoIP grows because some countries allow it to be used for a sort of regulatory arbitrage. It popped up before the rules covered it, or they didn't know how to deal with it, so it got special favored treatment. That's not the same as saying it is "unregulated"! In the USA, long distance is almost unregulated, but the local telephone monoplies are regulated -- they have a stake in how much they can charge for VoIP calls that use their networks the same way other long distance calls do. Expect an interesting year at the FCC while this is debated.
I do not expect computer-to-computer VoIP to be regulated (in the USA) at all; it's simply not anyone's but the users' to deal with. But of course some cable or DSL providers might try to block it, in order to sell their own phone services -- that'll be interesting to watch.
The Ofcom release goes beyond the March article, but not far enough in explaining the details.
They have adopted "light touch" licensing, which will be an on-line process, with a fee of one quid per user, fifty quid minimum. Hardly the type of fees garnered in the 3G auctions! It seems to me that this is primarily intended to keep track of the units, in case there's interference to a radar installation.
What's missing from the press release, or ofcom's web site (that I can see), are details like power limits, ERP limits, etc. The USA, for instance, has this as the "U-NII" band, theoretically authorized for unlicensed high-speed metropolitan use, with a power limit of 4 watts ERP. That sounds fast but the 5 GHz band is very sensitive to foliage fade and other obstructions. So it hasn't been of much practical use; it's largely used for backhaul between 2.4 GHz access points.
Weber State is famous for having launched its own satellite, Webersat, one of the OSCAR series. These kids are really at home with UHF and microwave radios!
In practical terms, the range of a microwave link, such as 2.4 GHz, is based on having line of sight without attenuation. The radio line of sight path is based on the horizon, with a simple guideline of roughly horizon (miles) = 1.4 * sqrt(height-in-feet). So if you have totally flat ground and 100-foot towers, your range to the horizon is 14 miles. The range of a hop is the sum of both sides' horizons. Now if you have a 2500-foot-high mountain to stand on, then your horizon is stretched to 70 miles.
The path loss is a function of distance, which antenna gain can make up for. The legality of doing this with unlicensed WiFi is a different question. Ham radio operators do this stuff routinely, but ham power limits are much higher, and there's no ERP limit. The 10 GHz band in particular is said to be popular in England. The crowded 144 and 430 MHz bands respond to similar rules. Attenuation by moisture in the air (serious form: rain fade) can get in the way, though. So if you're really looking for good distance, a nice place might be, oh, the Utah desert. Flat and no humidity.
So while it's possible to hack a good range with enough effort, conventional WiFi equipment is still not reliable getting from one side of my house to the other. It's really not a threat to the phone companies, especially in non-rural areas.
Well, no; CallSmart is pushing the rules, quite obviously and quite openly.
Under FCC guidelines (not formally "rules" yet, a phone to phone nonlocal call that crosses state lines is properly treated as long distance, whether or not it happens to use IP, ATM, or smoke signals in the middle of it. The local phone companies are thus allowed to charge "switched access" charges, rather than local; for a rural company, those charges can be quite stiff. CallSmart is claiming that their phone-to-phone service is not what it is. MCI, btw, claimed the same thing about its Execunet service in 1976; that case was eventually settled by the creation of the carrier access charge system that remains in effect. (Before that, AT&T Long Lines and the Bell Operating Companies used a "separation of revenues" formula to pass subsidies to the latter.)
Rural phone companies face very heavy capital costs (often over $10k/line), which cannot be covered by "affordable" monthy rates. So there's an elaborate subsidy system in place. CallSmart is evading it. Now it's easy to understand why they'd want to, but somebody has to pay for those rural phone lines, and monthly bills typically cover less than 1/5 of their cost! The rest comes from bloated LD access charges (what CallSmart is evading) and direct subsidies (that "Universal Service" line item on your LD bill largely goes to rural telcos).
I don't think the current system is right -- it encourages inefficiency, excessively discourages wireless substitution, and leads to arbitrage hacks like this. But CallSmart is not Vonage, which gets by because it requires a computing device (AT-186) at the customer site and is thus not "phone to phone" but technically "computer to phone". And yes, a PBX can be subjected to switched access charges too; I know of cases where that was arranged.
Where I am, the Comcast plant, ex-MediaOne, is in good shape, and well maintained. I get phone service off of it and reliability is better than VeriZontal's. This is not VoIP, just TDM/FDM cable telephony. The cable modem's pretty reliable too, though there are sometimes outage -- most often upstream, getting to the Internet backbone. On the other hand I've known cable systems whose raison d'etre seemed to be to make VeriZontal look good. It doesn't hurt that I'm fairly near the cable head end and nowhere near the telco central office.
The FCC really has no interest in regulating VoIP per se. A telephone call between two IP nodes (computer to computer) is simply not of interest to them. It is competition, sure, but that's okay; they're supposed to support competition. They generally don't, but they're supposed to.
Here's Chairman Mike "the lesser" Powell on the subject, from http://hraunfoss.fcc.gov/edocs_public/attachmatch/ DOC-241750A1.pdf
And I think the first thing is to truly commit ourselves to understanding the technology
and the market and the nature and the way in which it's unfolding so that we're not fearful --- we don't approach this with fear, but we approach it with excitement and optimism that this is inherently a good thing for the world, a good thing for the country, a major breakthrough in telecommunications and communications, and a great new opportunity and promise. We shouldn't be afraid of that.
I personally think we should be embracing it. And so I think that that's an important part of the debate. And so we're very excited about it.
I have decided that the Commission is now going to start exerting itself in this area much more directly. And that is not to say regulating it either, only to put a marker down that it's time to start having these policy questions in forums that matter. I really don't. And I think that we run the risk that if we don't move quickly to at least show that we're focused on it, then if you don't have a state jurisdiction do it, you will have a court do it.
There are some big issues still unresolved. The current FCC policies, which are largely supported by the language of the Telecom Act, classify calls made through regulated local telephone companies (VZ, SBC, etc.) as "telephone exchange service" (basically, local) and "exchange access service" (basically, end legs of a toll call). Those have different prices; LD carriers usually pay more for "access". VoIP is sometimes used as a way around that. So it threatens that subsidy mechanism, which is particularly important for rural telephone companies.
So the big questions focus on when does a VoIP call become long distance "access" rather than "local" or ISP-bound "exempt information access" (ISP access dialup calls, for now, are legally classified as not local. but telcos are usually required to treat them as if they were). And if VoIP calls are exempt, when is a call exempt? If AT&T sticks IP headers on the middle of its LD trunks, transparent to the user, does it become exempt? If the trunks are dedicated VoIP circuits? If the calls sound crappy enough?
I'm not sure the FCC is going to come up with any great answers in a hurry, but they have enough problems figuring out what the telephone companies can charge VoIP users without having to worry about messing around with Internet user traffic.
The RF interference issue is of somewhat broader significance than to just ham and CB radio operators, who will essentially be knocked off the air by BPL:
But some broadcasters use electrical wires as antennas for radio signals and are concerned that the internet signals could interfere with radio and television reception.
Broadcast expert Jacques Bouliane said the internet signal could completely ruin television reception.
"Even if you don't subscribe to the service, you would get interference from it," he said.
Hydro said it won't be a problem, and pointed out that interference doesn't occur over cables that provide both television and internet service.
Hydro's argument there, of course, is utterly specious! Cable and internet are carried over coaxial cable, which is shielded. In the USA -- I assume Canadian rules are similar -- cable operators have to go around and "sweep" their coax for RF leaks twice a year, a costly process. Power lines are not shielded or even designed to carry RF without leaking. They're big distributed antennas.
BPL creates loud radio noise across the shortwave spectrum, above AM radio but below FM. Some systems might intefere with VHF TV (channels 2-6). The key is that they knock out shortwave (3-30 MHz), where foreign broadcasts can be received. Not a lot of Americans, percentage wise, listen to shortwave, though it's popular in many other countries. But it's a useful counter to monopoly media. Clear Channel, Viacom and Fox want to control what you hear. Bush likes that, and they like him. Shortwave lets in foreign ideas and different viewpoints.
Remember, BPL Internet is not a "common carrier" service. The Powerline ISP will have the right of censorship, just as other ISPs do today. You can read Slashdot freely because there's too much ISP competition today to allow a censored ISP to get very far -- "Christian" censored ISPs do have some customers, though, especially in Dixie. If and when the FCC and the ILECs are done knocking out competitive ISPs, you can expect a lot less freedom on the net.
The truth is, the FCC doesn't particularly want to regulate VoIP -- nor do they regulate much long distance. Those are not monopolies, and there's a competitive market for them already.
What the FCC regulates is the behavior of monopoly local phone companies (ILECs) towards ISPs, VoIP providers, and long distance companies. The incumbent LECs charge more to LD companies than by retail users. So they want VoIP and Internet for that matter to be classified as long distance, so they can be permitted to charge more when it uses their dial-up lines. That's the primary controversy.
PC-to-PC is off the radar, except to the extent that it so threatens the ILECs that they use it as an excuse to demand that dial-up ISPs be classified as LD companies. That was proposed by the FCC in 1987 and dubbed the "modem tax", and only stopped because a Democratic Congress made very, very clear that the FCC would be punished severely if they went ahead with it. Congress makes the laws the FCC's rules are supposed to follow, and (more importantly) funds them.
I agree that the broadcast flag is a mess, but that's a different Bureau of the FCC, and they're so compartmentalized that if a huge fire broke out in one bureau, the one in the next office suite over would probably keep working until ordered out by the fire department. Not their problem, after all.
The FCC has no intention, I am quite sure, of regulating private VoIP, or any computer-to-computer applications. They're really, really, not interested in going there. (I do this stuff for a living. I'm not a lawyer, but I do regulatory work, and often write formal Comments on FCC proceedings. So I stay on top of this sort of thing.) Theoretically, they do have a lot of authority that they do not exercise. But for the past 25 years or so, their direction has been to exercise authority to prevent monopolies from impeding progress. The Internet itself only exists, for instance, because the FCC ordered AT&T, in the 1970s, to remove a restriction on "sharing and resale" of leased line circuits.
The FCC is however interested in a number of very sticky questions that relate to VoIP. The telephone network itself is subject to fairly strict regulation, particularly the amount of money that each carrier is allowed to charge the other carriers on a given call. So when somebody in Virginia calls somebody in California over Qwest's network, how much does VZ in VA get from Q, how much does SBC in CA get from Q? Those are covered by detailed tariffs.
A local leg of an interstate call is not treated the same as a local call. The current regulatory system is based on a system of classification, and that system is obsolete. VoIP increases the pressure on it.
VoIP threatens that because it's so easy to sneak around the usual processes. The current FCC not-quite-rule (an April 1999 "Report to Congress", which is an unofficial policy statement) says that "phone to phone" VoIP calls are just plain calls, subject to the same payments as other calls. PC-to-phone calls, however, are undefined. And there are all sorts of variations. The big phone companies know it, and want to use their influence to make things go their own way. Small, rural local phone companies actually have the most to lose, because they get a much bigger share of their revenue from long distance settlements. Rural state regulators and legislators are very protective of these companies.
Well, I've never actually used BitTorrent before, and it was not that easy to get going -- the d*mned dependency tree for wxWindows, or something like that, couldn't resolve until I picked the *right* package to install, whose mirrors were okay, and then it updated 17 packages, including TuxRacer fercripesake, but finally it was in. Then the curses version, but not the gui, worked, and started the download.
I've got a cable modem and typically ftp at about 150-200 kbyte/sec. I was only getting 20k down, while feeding about 25k up! Ugh. So I come to Slashdot to look at the gripefest. I see Thing 1's note. I go into my Netgear (ZyXel in disguise) router and turn on port forwarding for 6881 to 6889. In seconds, the download rate rises to about 100 kbyte/s! (It's around 135k now.) Projected time falls from 48H to 6H!
I don't understand how BitTorrent works or how it kinda sorta works but only slowly when port forwarding is off. But I do appreciate Thing 1's note, and would moderate it up if I had mod privs today. Thanks!
Using a pay-to-send relay agent is crude, but it doesn't really address all the issues. Who runs the relays, and who decides if the relay is valid, or is a spammer's? Who decides the price of relaying? What about mailing lists?
I prefer a system of micropostage, in which there is no single postage-issuing authority, but the mail receiver maintains a whitelist of acceptable ones. (If one becomes compromised by spammers, then it gets deleted from the whitelist, a quick anc clean form of RBL.) Micropostage is only needed when the mail comes from a stranger; users can put their friends, correspondents, cow-orkers (whole domains) and mailing lists onto a whitelist. Postage is only checked at the receiving end, where something that arrives without a valid stamp and is not from someone on a whitelist will be rejected.
In this micropostage scheme, micropostage is very cheap (fractions of a cent). It takes the form of one-time digital signatures. The recipient has to query every stampette-bearing email against its issuing micropostage authority, which determines if it isboth valid and has not already been used. The micropostage authority recovers its costs via the sale of stampettes. And if it sets its price low enough for spammers, then its stampettes don't get whitelisted. If it sets its price too high, a competitors' stampettes get used instead.
End users should get allotments of stampettes from their ISPs. If they're compromised by a virus, the allotment will run out, and the ISP will demand virus removal before giving them more.
Let's face it; SMTP sucks, was never intended for a big public network, and needs major replacement. But the insistence that email be "free as in beer" will doom any alternative. Cheap, yes but when a million mails to strangers cost nearly zero (especially with spammers stealing service as they do), it's too tempting to spam. Some tiny cost is needed.
I didn't say it didn't exist. I did suggest that not many people had heard of it. Whether that matters or not is a question of judgement. It had little influence on future products. The Altair, on the other hand, spawned the S-100 bus, which was an industry of its own for a few years, and, e gads, Microsoft itself!
And the Altair folks, and the people who widely disseminated knowledge about it, knew about Scelbi. I am not insulting MCM's product, but its impact on the future was limited.
The MCM was a little-known machine in its day. I was following the birth of the micro as it happened; I have a few copies of Byte Magazine #1 (July, 1975) to prove it. Nobody in that circle ever heard of MCM, I suspect.
The Scelbi 8-H (1974) was often considered the first hacker microcomputer; here's a picture: http://online.sfsu.edu/~hl/c.Scelbi8H.html . It was $580, though by itself it did little. It used the 8008. Very few were made. In early 1975, MITS came out with the Altair 8800, using the Intel 8080 CPU. Much nicer than the 8008, but the Altair was modeled on minicomputers (PDP-8, for instance) and used lots of sense switches and blinkenlights on the front panel. External teletype for I/O, for instance, and cassette. Or the SWTP "TV Typewriter" kit.
Bill Gates went to Albuquerque that summer to work at MITS, with the "4K BASIC" and "8K BASIC" interpreters. He never went back to finish his studies at Harvard. He went home to Seattle to sell software. Apple started in 1996, competing with MITS and SWTP, who by then had a popular 6800-based unit. Ohio Scientific and a few others tried to sell micros to business. IBM actually had a BASIC-speaking CRT/keyboard business micro, the 5100, in 1977 or so -- the original "PC" was the 5150, reflecting its family heritage.
The original RCA corp. no longer exists. Sarnoff's baby, which made Nipper the dog a household icon, was acquired some years ago by GE. In turn, GE licensed the RCA name, along with the GE name (for consumer electronics products), to Thomson. So today's RCA TV sets are made by Thomson. If I'm not mistaken, they're starting to play in the PVR space too.
I'll explain *specifically* what I witnessed myself, as a recipient of these calls.
My home phone rang, and I read the caller ID. It was an unfamiliar local number, showing no name, just the city. The caller, however, was in another city. He was making an ordinary LD call using his local phone company (no VoIP hacks) and MCI as the LD company. I specifically noted this on two different occasions, with the caller's real locations on opposite sides of the country (NY and CA).
The local number at my end belonged to Focal, a CLEC. So my local carrier, AT&T, saw the call as a local one from a Focal subscriber. I presume it had a Focal point code (Signaling System 7 network address; these are three-octet numbers, like 005.103.204). Thus AT&T did not see the call as MCIs, and did not see the call as long distance. Focal may have owed them a little fee for local termination, if anything, but not the fee that applies to long distance termination.
Under the current FCC rules, the rate that a carrier pays for a half-call depends on whether the call was toll or local or ISP-bound. And don't ask about VoIP's status, which is fuzzy. Having different rates for the same thing invites fraud. But in MCI's case, it was blatant -- the incorrect CallerID gave the whole thing away.
Carterfone allowed "foreign attachments" but only using a "protective coupler", which was an extra-rent box from The Phone Company. So you still couldn't plug in your own phones directly and be legal (though of course many of us did so anyway). Around 1977, the FCC introduced Registration, which did away with the protective coupler requirement, provided that the equipment was either a) Registered with the FCC, having passed tests (this is what everything does now); or b) Grandfathered -- if it's something that any phne company used themselves prior to the start of registration, then every phone company had to allow it to be attached directly. This legalized all those standard phones.
Prices for cable modems and DSL-based ISP service are not regulated in the United States!
Raw Bell co. DSL bit-pipe service to independent ISPs is tariffed, but there is no particular approval cycle on the tariffs. The FCC accepts what they file, and that's that. The FCC is considering removing that requirement, so that the Bell company is as free to set prices and cut off competitive access to their wires as cable companies are. Cable is regulated under a different law, which is why it's different.
SBC and the other incumbent telephone companies grew up with protected monopolies. They grumbled when the FCC's Carterfone ruling in 1969 forced them to allow "foreign" attachments of customer-owned equipment like telephone sets, PBXs, answering machines and modems. (Before that, you could only rent equipment from them. A 300-baud modem was $25/month.) They grumbled when long distance competition was authorized.
They would have grumbled when local telephone service competition was authorized in 1996, but they got, in return, permission to offer long distance service and "advanced" services such as Internet. So having gotten much in return, they're trying to weasel out of their half of the bargain. Powell's FCC has rolled back competition. They're making it next to impossible for CLECs to lease the high-frequency part of copper that's needed to offer consumer DSL service, and even cutting off some access to plain old full-price copper wire. So the CLECs like Covad won't be able to offer the ISPs a substitute for ILEC (SBC, VZ, etc.) DSL. Game! Powell also has a pending proposal that removes common carrier status from ILEC DSL, which is what this case is about -- SBC won't be required by federal regulation to offer raw DSL bit-pipe service to competitors of its Prodigy ISP service. Set! And even dial-up is coming under increased attack; many dial-up ISPs are becoming reclassified as toll calls, as the ILECs try to worm in a back-door "modem tax". It's happening -- I'm involved in some of these cases. Match!
So the independent ISPs are being squeezed hard. Under the old pre-1996 regulations, the ILECs were not subject to much antitrust review, because regulation controlled them. Now, they're being unshackled, but they still have their inherited monopolies on essential facilities -- that's a term of art in the antitrust business. They're blatantly using these monopolies (the copper loop) to leverage sales of what should be fully-competitive businesses (ISPs like Prodigy and VZ Online). That is certainly a red flag in antitrust.
Since the regulators (FCC) have stepped aside, relief will have to happen in the courts. A number of cases are pending now; this one looks to be particularly important. Its fate will help determine if the American public will have free access to the Internet, or whether we'll be stuck behind a corporate-administered Great Firewall of Bell, paying top dollar for limited choice.
And with an Internet in monopoly hands, the FCC's excuse for broadcast ownership deregulation (extreme concentration of ownership of the media) is proven a lie. But Powell hopes we don't notice.
This is standard process in the database biz, including things like mailing lists and (as others have noted here) maps. The term for it is "salting". Calling them "honeytokens" is applying the wrong seasoning... and treating it as new on/. is also silly.
Excellent post. I was going to write a similar note but I'll second this one and add to it.
Circuit-switched telephone networks aren't actually costly to build nowadays -- the competitive-bid price of circuit-switching (TDM) gear is a small fraction of what it was 15 years ago. Lucent and Nortel stock suffers as a result. Undersea cable bandwidth is also much, much cheaper. If one evaluates the cost of building a new wireline network from scratch, then TDM/circuit is not much costlier than VoIP; it could even be cheaper. BTW I do this for a living so I see the real prices.
However, most underdeveloped countries do not treat the telephone network rationally. It is a government agency, and its primary purposes are employment and taxation. Actually providing good service is secondary.
So the price of international calls is kept high. Outgoing rates are high because it's viewed as a luxury -- only a small fraction of the public can afford it, and those foreign businesses are rich, so they can subsidize the telehone network's employment role with very high prices. The price of incoming calls is held high too -- they charge high "settlement" fees to international LD carriers, who reflect them in their rates. That's why international call rates vary so widely between nearby countries. The price, then, has nothing to do with cost, and more to do with what a monopoly can get away with.
VoIP's role here is not technological goodness, it's stealth. Users can cloak phone calls inside IP packets that aren't metered at the country's borders. Thus they get around settlement fees on inbound calls, and don't pay the ridiculous outbound rates. This is good, from an economic perspective (arbitrage removes price distortion), but it makes VoIP look relatively better than it should. Government action distortions outside perspectives on technology.
If these countries really wanted to join the wired (and BTW even cell phones are "wired" in this sense) technological world, then they should rethink their adminstrative policies towards telecommunications in general.
Many of these countries own their own telephone company, in whole or with a partner. In some Caribbean states (I don't know about St. Lucia offhand), Cable & Wireless, a privatized formerly-British state-owned company, has the monopoly, and has its tentacles deep into the regulators.
So a WISP would be a competitor, and that's simply not allowed. All communications money must flow to the government or C&W (who pays off the right people). That used to be the system in the USA, and the RBOCs are trying to get it back, but wireless has too many friends here now. Unlicensed packet radios (802.11) are a relatively recent change in the USA. In other countries, the mere right to be an ISP is again a government franchise, and competition is Not Allowed. This helps them censor opposition thought, and keep the price high (again, for the sake of the monopolist).
Berkeley did not invent LSD, and I don't think that the original quote was intending to denigrate Hoffman's work. Berkeley did, however, play some role in the product's development. While Dr. Leary of Harvard, not Berkeley, was the foremost advocate of various "off-label" uses, Berkeley was the site of considerable uh experimentation, yeah, that's the ticket, which was for a time quite influential. Yep, trust Berkeley to have plenty of volunteers for experiments with mind-altering pharmaceuticals....
Digression... actually, post-mortem recounts of the entire state show that Gore would have won had the entire state been recounted, as the Bush side had once suggested. Had the recount only occurred in areas where Gore suggested it was necessary, Bush would have ended up ahead. But even that leaves out the impact of the butterfly ballot, of the roadblock keeping African-American voters away from the polls on the panhandle, and from the thousands of eligible African-American voters stricken from the rolls under a contract to ChoicePoint, while at-the-poll appeals were primarily made available to Cubans.
Just trying to fight a defective meme before it spreads.
As a consultant to the telecom industry, I have to pay attention to what works, not what sells papers. This year's big hype is WiFi, which was designed for room-sized LANs but somehow seems to have captured the imagination of the public as if it could actually hurt the telcos (ILECs). VoIP has been hyped since '96 or so, and has eaten tons of vulture capital, and while it has nice niche applications, it is still no substitute for the Real Thing.
Yes, you can run voice over IP. Yes, you can run IP over wireless. Heck, 16 years ago I was running IP over 1200 bps Aloha AX.25 packet radio links. Very instructional, because Phil Karn's NOS let me watch a decoded protocol trace of passing packets, and they came so slowly that I could study all of them in real time. Think about it. The point is that you can't run voice over *any* IP, just some paths.
Circuit-switched telephony is cheap to build. Sure the existing telco networks are made of gear that they paid a lot for, but ILECs depreciate gear over 20+ years. So the Lucent 5ESS and Nortel DMS-100 are VAX-era hardware. What did a MIPS cost when the 5E was designed? Modern circuit switching (which CLECs and some small ILECs buy, not to mention the PBX market) uses modern parts. The switching hardware is only a little costlier than IP stuff, and it sounds better. All the sexy call control features are in the control software, which in a modern system is agnostic about physical-layer protocol. So you can do nice things on circuit, ATM, or IP. Just a different card in the switch.
WiFi's limits are obvious -- there's finite spectrum, and it's shared with domestic cookers (microwave ovens are right in the middle of its band!), cordless phones, VCR "multipliers", baby monitors, and all sorts of other crap. WiFi5 is cleaner spectrum, though the lower-volume gear is costlier. The 5 GHz band will benefit from a recent FCC rule change that adds 275 MHz more bandwidth. But unlicensed still means low power, and either very short range *or* directional antennas (which take more work). And you have to worry about things like hills and trees.
I'm always looking for alternatives to Bell wire -- that's really a big part of my job! But WiFi ain't it. There are non-WiFi radios that are better for "last mile" purposes (and slower, because they have to trade speed for range -- see Shannon). The FCC is contemplating making some additional frequencies available, and in rural/exurban areas, especially flatland, wireless can do wonders. In hilly or woody areas, it's tougher. In urban areas, spectrum is too limited. Fiber optic bandwidth is infinite -- there's lots of sand out there, and only one radio spectrum.
Slashdot Mirror
Simson's right in denying IPv6's short-term inevitability, but he's still being too easy on it! IPv6 is just plain dumb. He should say it.
IPv6 creates much larger headers, so there's more overhead, particularly, as a percentage, on short packets (voice, ACK's, etc.). So it'll waste bandwidth, or lower effective throughput on fixed bandwidths. We need this? It is not even using its 128 bits efficiently. The general approach is to use the top half to identify the network and the bottom half to include the 48-bit MAC address of the computer. That was a clever hack in 1985 when proposed for DECnet Phase V (which never caught on) and became an approach in OSI CLNP. But that was not for a public spammer-ridden insecure Internet. Now it is a security and privacy hole to do that. It also means the 128 bits are not used efficiently -- we are tight with 32 bits, but an address for every atom?
IPv6 also does nothing for QoS (ignore the hype, which is based on a misunderstanding) and nothing for security (IPsec works just fine with v4). It just wastes bandwidth. So it does something for, oh, MCI. No wonder Vint (the Chauncey Gardner of the Internet) likes it! And Sprint, AT&T and VeriZontal. Great.
IPv4 could use a decent replacement some day, but IPv6 is everything you don't like about v4, and more. Eccch. A dozen years since it was "adopted" and it's gone nowhere, for good reason. The Asians weren't so involved with IETF at the time, to know the messy politics behind it. And btw the whole thing about their not having addresses is false; there is plenty of space left in the IPv4 space waiting to be allocated where needed. China can have more, as they provide more and more spam relays for the h3rb@1-v14gr4 crowd.
The problem with VoIP is that it isn't half as good as people think it is -- there are certainly good niche applications, and ways to use it profitably, but it simply isn't the be-all and end-all. Why do people fawn over it so much? I think it's largely because "IP" has that "k3w1" quality of the Internet in general, while phones are passe -- hardly a good way to make rational decisions.
s /HotNets0 2-IP_conquest_of_the_world_with_authors.pdf
This paper is pretty useful:
http://klamath.stanford.edu/~nickm/paper
In the meantime, VoIP grows because some countries allow it to be used for a sort of regulatory arbitrage. It popped up before the rules covered it, or they didn't know how to deal with it, so it got special favored treatment. That's not the same as saying it is "unregulated"! In the USA, long distance is almost unregulated, but the local telephone monoplies are regulated -- they have a stake in how much they can charge for VoIP calls that use their networks the same way other long distance calls do. Expect an interesting year at the FCC while this is debated.
I do not expect computer-to-computer VoIP to be regulated (in the USA) at all; it's simply not anyone's but the users' to deal with. But of course some cable or DSL providers might try to block it, in order to sell their own phone services -- that'll be interesting to watch.
The Ofcom release goes beyond the March article, but not far enough in explaining the details.
They have adopted "light touch" licensing, which will be an on-line process, with a fee of one quid per user, fifty quid minimum. Hardly the type of fees garnered in the 3G auctions! It seems to me that this is primarily intended to keep track of the units, in case there's interference to a radar installation.
What's missing from the press release, or ofcom's web site (that I can see), are details like power limits, ERP limits, etc. The USA, for instance, has this as the "U-NII" band, theoretically authorized for unlicensed high-speed metropolitan use, with a power limit of 4 watts ERP. That sounds fast but the 5 GHz band is very sensitive to foliage fade and other obstructions. So it hasn't been of much practical use; it's largely used for backhaul between 2.4 GHz access points.
Weber State is famous for having launched its own satellite, Webersat, one of the OSCAR series. These kids are really at home with UHF and microwave radios!
In practical terms, the range of a microwave link, such as 2.4 GHz, is based on having line of sight without attenuation. The radio line of sight path is based on the horizon, with a simple guideline of roughly horizon (miles) = 1.4 * sqrt(height-in-feet). So if you have totally flat ground and 100-foot towers, your range to the horizon is 14 miles. The range of a hop is the sum of both sides' horizons. Now if you have a 2500-foot-high mountain to stand on, then your horizon is stretched to 70 miles.
The path loss is a function of distance, which antenna gain can make up for. The legality of doing this with unlicensed WiFi is a different question. Ham radio operators do this stuff routinely, but ham power limits are much higher, and there's no ERP limit. The 10 GHz band in particular is said to be popular in England. The crowded 144 and 430 MHz bands respond to similar rules. Attenuation by moisture in the air (serious form: rain fade) can get in the way, though. So if you're really looking for good distance, a nice place might be, oh, the Utah desert. Flat and no humidity.
So while it's possible to hack a good range with enough effort, conventional WiFi equipment is still not reliable getting from one side of my house to the other. It's really not a threat to the phone companies, especially in non-rural areas.
Well, no; CallSmart is pushing the rules, quite obviously and quite openly.
Under FCC guidelines (not formally "rules" yet, a phone to phone nonlocal call that crosses state lines is properly treated as long distance, whether or not it happens to use IP, ATM, or smoke signals in the middle of it. The local phone companies are thus allowed to charge "switched access" charges, rather than local; for a rural company, those charges can be quite stiff. CallSmart is claiming that their phone-to-phone service is not what it is. MCI, btw, claimed the same thing about its Execunet service in 1976; that case was eventually settled by the creation of the carrier access charge system that remains in effect. (Before that, AT&T Long Lines and the Bell Operating Companies used a "separation of revenues" formula to pass subsidies to the latter.)
Rural phone companies face very heavy capital costs (often over $10k/line), which cannot be covered by "affordable" monthy rates. So there's an elaborate subsidy system in place. CallSmart is evading it. Now it's easy to understand why they'd want to, but somebody has to pay for those rural phone lines, and monthly bills typically cover less than 1/5 of their cost! The rest comes from bloated LD access charges (what CallSmart is evading) and direct subsidies (that "Universal Service" line item on your LD bill largely goes to rural telcos).
I don't think the current system is right -- it encourages inefficiency, excessively discourages wireless substitution, and leads to arbitrage hacks like this. But CallSmart is not Vonage, which gets by because it requires a computing device (AT-186) at the customer site and is thus not "phone to phone" but technically "computer to phone". And yes, a PBX can be subjected to switched access charges too; I know of cases where that was arranged.
This depends a lot on your cable company.
Where I am, the Comcast plant, ex-MediaOne, is in good shape, and well maintained. I get phone service off of it and reliability is better than VeriZontal's. This is not VoIP, just TDM/FDM cable telephony. The cable modem's pretty reliable too, though there are sometimes outage -- most often upstream, getting to the Internet backbone. On the other hand I've known cable systems whose raison d'etre seemed to be to make VeriZontal look good. It doesn't hurt that I'm fairly near the cable head end and nowhere near the telco central office.
Here's Chairman Mike "the lesser" Powell on the subject, from http://hraunfoss.fcc.gov/edocs_public/attachmatch
There are some big issues still unresolved. The current FCC policies, which are largely supported by the language of the Telecom Act, classify calls made through regulated local telephone companies (VZ, SBC, etc.) as "telephone exchange service" (basically, local) and "exchange access service" (basically, end legs of a toll call). Those have different prices; LD carriers usually pay more for "access". VoIP is sometimes used as a way around that. So it threatens that subsidy mechanism, which is particularly important for rural telephone companies.
So the big questions focus on when does a VoIP call become long distance "access" rather than "local" or ISP-bound "exempt information access" (ISP access dialup calls, for now, are legally classified as not local. but telcos are usually required to treat them as if they were). And if VoIP calls are exempt, when is a call exempt? If AT&T sticks IP headers on the middle of its LD trunks, transparent to the user, does it become exempt? If the trunks are dedicated VoIP circuits? If the calls sound crappy enough?
I'm not sure the FCC is going to come up with any great answers in a hurry, but they have enough problems figuring out what the telephone companies can charge VoIP users without having to worry about messing around with Internet user traffic.
The RF interference issue is of somewhat broader significance than to just ham and CB radio operators, who will essentially be knocked off the air by BPL:
But some broadcasters use electrical wires as antennas for radio signals and are concerned that the internet signals could interfere with radio and television reception.
Broadcast expert Jacques Bouliane said the internet signal could completely ruin television reception.
"Even if you don't subscribe to the service, you would get interference from it," he said.
Hydro said it won't be a problem, and pointed out that interference doesn't occur over cables that provide both television and internet service.
Hydro's argument there, of course, is utterly specious! Cable and internet are carried over coaxial cable, which is shielded. In the USA -- I assume Canadian rules are similar -- cable operators have to go around and "sweep" their coax for RF leaks twice a year, a costly process. Power lines are not shielded or even designed to carry RF without leaking. They're big distributed antennas.
BPL creates loud radio noise across the shortwave spectrum, above AM radio but below FM. Some systems might intefere with VHF TV (channels 2-6). The key is that they knock out shortwave (3-30 MHz), where foreign broadcasts can be received. Not a lot of Americans, percentage wise, listen to shortwave, though it's popular in many other countries. But it's a useful counter to monopoly media. Clear Channel, Viacom and Fox want to control what you hear. Bush likes that, and they like him. Shortwave lets in foreign ideas and different viewpoints.
Remember, BPL Internet is not a "common carrier" service. The Powerline ISP will have the right of censorship, just as other ISPs do today. You can read Slashdot freely because there's too much ISP competition today to allow a censored ISP to get very far -- "Christian" censored ISPs do have some customers, though, especially in Dixie. If and when the FCC and the ILECs are done knocking out competitive ISPs, you can expect a lot less freedom on the net.
The truth is, the FCC doesn't particularly want to regulate VoIP -- nor do they regulate much long distance. Those are not monopolies, and there's a competitive market for them already.
What the FCC regulates is the behavior of monopoly local phone companies (ILECs) towards ISPs, VoIP providers, and long distance companies. The incumbent LECs charge more to LD companies than by retail users. So they want VoIP and Internet for that matter to be classified as long distance, so they can be permitted to charge more when it uses their dial-up lines. That's the primary controversy.
PC-to-PC is off the radar, except to the extent that it so threatens the ILECs that they use it as an excuse to demand that dial-up ISPs be classified as LD companies. That was proposed by the FCC in 1987 and dubbed the "modem tax", and only stopped because a Democratic Congress made very, very clear that the FCC would be punished severely if they went ahead with it. Congress makes the laws the FCC's rules are supposed to follow, and (more importantly) funds them.
I agree that the broadcast flag is a mess, but that's a different Bureau of the FCC, and they're so compartmentalized that if a huge fire broke out in one bureau, the one in the next office suite over would probably keep working until ordered out by the fire department. Not their problem, after all.
The FCC has no intention, I am quite sure, of regulating private VoIP, or any computer-to-computer applications. They're really, really, not interested in going there. (I do this stuff for a living. I'm not a lawyer, but I do regulatory work, and often write formal Comments on FCC proceedings. So I stay on top of this sort of thing.) Theoretically, they do have a lot of authority that they do not exercise. But for the past 25 years or so, their direction has been to exercise authority to prevent monopolies from impeding progress. The Internet itself only exists, for instance, because the FCC ordered AT&T, in the 1970s, to remove a restriction on "sharing and resale" of leased line circuits.
The FCC is however interested in a number of very sticky questions that relate to VoIP. The telephone network itself is subject to fairly strict regulation, particularly the amount of money that each carrier is allowed to charge the other carriers on a given call. So when somebody in Virginia calls somebody in California over Qwest's network, how much does VZ in VA get from Q, how much does SBC in CA get from Q? Those are covered by detailed tariffs.
A local leg of an interstate call is not treated the same as a local call. The current regulatory system is based on a system of classification, and that system is obsolete. VoIP increases the pressure on it.
VoIP threatens that because it's so easy to sneak around the usual processes. The current FCC not-quite-rule (an April 1999 "Report to Congress", which is an unofficial policy statement) says that "phone to phone" VoIP calls are just plain calls, subject to the same payments as other calls. PC-to-phone calls, however, are undefined. And there are all sorts of variations. The big phone companies know it, and want to use their influence to make things go their own way. Small, rural local phone companies actually have the most to lose, because they get a much bigger share of their revenue from long distance settlements. Rural state regulators and legislators are very protective of these companies.
Well, I've never actually used BitTorrent before, and it was not that easy to get going -- the d*mned dependency tree for wxWindows, or something like that, couldn't resolve until I picked the *right* package to install, whose mirrors were okay, and then it updated 17 packages, including TuxRacer fercripesake, but finally it was in. Then the curses version, but not the gui, worked, and started the download.
I've got a cable modem and typically ftp at about 150-200 kbyte/sec. I was only getting 20k down, while feeding about 25k up! Ugh. So I come to Slashdot to look at the gripefest. I see Thing 1's note. I go into my Netgear (ZyXel in disguise) router and turn on port forwarding for 6881 to 6889. In seconds, the download rate rises to about 100 kbyte/s! (It's around 135k now.) Projected time falls from 48H to 6H!
I don't understand how BitTorrent works or how it kinda sorta works but only slowly when port forwarding is off. But I do appreciate Thing 1's note, and would moderate it up if I had mod privs today. Thanks!
Using a pay-to-send relay agent is crude, but it doesn't really address all the issues. Who runs the relays, and who decides if the relay is valid, or is a spammer's? Who decides the price of relaying? What about mailing lists?
I prefer a system of micropostage, in which there is no single postage-issuing authority, but the mail receiver maintains a whitelist of acceptable ones. (If one becomes compromised by spammers, then it gets deleted from the whitelist, a quick anc clean form of RBL.) Micropostage is only needed when the mail comes from a stranger; users can put their friends, correspondents, cow-orkers (whole domains) and mailing lists onto a whitelist. Postage is only checked at the receiving end, where something that arrives without a valid stamp and is not from someone on a whitelist will be rejected.
In this micropostage scheme, micropostage is very cheap (fractions of a cent). It takes the form of one-time digital signatures. The recipient has to query every stampette-bearing email against its issuing micropostage authority, which determines if it isboth valid and has not already been used. The micropostage authority recovers its costs via the sale of stampettes. And if it sets its price low enough for spammers, then its stampettes don't get whitelisted. If it sets its price too high, a competitors' stampettes get used instead.
End users should get allotments of stampettes from their ISPs. If they're compromised by a virus, the allotment will run out, and the ISP will demand virus removal before giving them more.
Let's face it; SMTP sucks, was never intended for a big public network, and needs major replacement. But the insistence that email be "free as in beer" will doom any alternative. Cheap, yes but when a million mails to strangers cost nearly zero (especially with spammers stealing service as they do), it's too tempting to spam. Some tiny cost is needed.
I didn't say it didn't exist. I did suggest that not many people had heard of it. Whether that matters or not is a question of judgement. It had little influence on future products. The Altair, on the other hand, spawned the S-100 bus, which was an industry of its own for a few years, and, e gads, Microsoft itself!
And the Altair folks, and the people who widely disseminated knowledge about it, knew about Scelbi. I am not insulting MCM's product, but its impact on the future was limited.
The MCM was a little-known machine in its day. I was following the birth of the micro as it happened; I have a few copies of Byte Magazine #1 (July, 1975) to prove it. Nobody in that circle ever heard of MCM, I suspect.
The Scelbi 8-H (1974) was often considered the first hacker microcomputer; here's a picture: http://online.sfsu.edu/~hl/c.Scelbi8H.html . It was $580, though by itself it did little. It used the 8008. Very few were made. In early 1975, MITS came out with the Altair 8800, using the Intel 8080 CPU. Much nicer than the 8008, but the Altair was modeled on minicomputers (PDP-8, for instance) and used lots of sense switches and blinkenlights on the front panel. External teletype for I/O, for instance, and cassette. Or the SWTP "TV Typewriter" kit.
Bill Gates went to Albuquerque that summer to work at MITS, with the "4K BASIC" and "8K BASIC" interpreters. He never went back to finish his studies at Harvard. He went home to Seattle to sell software. Apple started in 1996, competing with MITS and SWTP, who by then had a popular 6800-based unit. Ohio Scientific and a few others tried to sell micros to business. IBM actually had a BASIC-speaking CRT/keyboard business micro, the 5100, in 1977 or so -- the original "PC" was the 5150, reflecting its family heritage.
How did RCA get into this?
The original RCA corp. no longer exists. Sarnoff's baby, which made Nipper the dog a household icon, was acquired some years ago by GE. In turn, GE licensed the RCA name, along with the GE name (for consumer electronics products), to Thomson. So today's RCA TV sets are made by Thomson. If I'm not mistaken, they're starting to play in the PVR space too.
I'll explain *specifically* what I witnessed myself, as a recipient of these calls.
My home phone rang, and I read the caller ID. It was an unfamiliar local number, showing no name, just the city. The caller, however, was in another city. He was making an ordinary LD call using his local phone company (no VoIP hacks) and MCI as the LD company. I specifically noted this on two different occasions, with the caller's real locations on opposite sides of the country (NY and CA).
The local number at my end belonged to Focal, a CLEC. So my local carrier, AT&T, saw the call as a local one from a Focal subscriber. I presume it had a Focal point code (Signaling System 7 network address; these are three-octet numbers, like 005.103.204). Thus AT&T did not see the call as MCIs, and did not see the call as long distance. Focal may have owed them a little fee for local termination, if anything, but not the fee that applies to long distance termination.
Under the current FCC rules, the rate that a carrier pays for a half-call depends on whether the call was toll or local or ISP-bound. And don't ask about VoIP's status, which is fuzzy. Having different rates for the same thing invites fraud. But in MCI's case, it was blatant -- the incorrect CallerID gave the whole thing away.
To clarify...
Carterfone allowed "foreign attachments" but only using a "protective coupler", which was an extra-rent box from The Phone Company. So you still couldn't plug in your own phones directly and be legal (though of course many of us did so anyway). Around 1977, the FCC introduced Registration, which did away with the protective coupler requirement, provided that the equipment was either
a) Registered with the FCC, having passed tests (this is what everything does now); or
b) Grandfathered -- if it's something that any phne company used themselves prior to the start of registration, then every phone company had to allow it to be attached directly. This legalized all those standard phones.
Prices for cable modems and DSL-based ISP service are not regulated in the United States!
Raw Bell co. DSL bit-pipe service to independent ISPs is tariffed, but there is no particular approval cycle on the tariffs. The FCC accepts what they file, and that's that. The FCC is considering removing that requirement, so that the Bell company is as free to set prices and cut off competitive access to their wires as cable companies are. Cable is regulated under a different law, which is why it's different.
SBC and the other incumbent telephone companies grew up with protected monopolies. They grumbled when the FCC's Carterfone ruling in 1969 forced them to allow "foreign" attachments of customer-owned equipment like telephone sets, PBXs, answering machines and modems. (Before that, you could only rent equipment from them. A 300-baud modem was $25/month.) They grumbled when long distance competition was authorized.
They would have grumbled when local telephone service competition was authorized in 1996, but they got, in return, permission to offer long distance service and "advanced" services such as Internet. So having gotten much in return, they're trying to weasel out of their half of the bargain. Powell's FCC has rolled back competition. They're making it next to impossible for CLECs to lease the high-frequency part of copper that's needed to offer consumer DSL service, and even cutting off some access to plain old full-price copper wire. So the CLECs like Covad won't be able to offer the ISPs a substitute for ILEC (SBC, VZ, etc.) DSL. Game! Powell also has a pending proposal that removes common carrier status from ILEC DSL, which is what this case is about -- SBC won't be required by federal regulation to offer raw DSL bit-pipe service to competitors of its Prodigy ISP service. Set! And even dial-up is coming under increased attack; many dial-up ISPs are becoming reclassified as toll calls, as the ILECs try to worm in a back-door "modem tax". It's happening -- I'm involved in some of these cases. Match!
So the independent ISPs are being squeezed hard. Under the old pre-1996 regulations, the ILECs were not subject to much antitrust review, because regulation controlled them. Now, they're being unshackled, but they still have their inherited monopolies on essential facilities -- that's a term of art in the antitrust business. They're blatantly using these monopolies (the copper loop) to leverage sales of what should be fully-competitive businesses (ISPs like Prodigy and VZ Online). That is certainly a red flag in antitrust.
Since the regulators (FCC) have stepped aside, relief will have to happen in the courts. A number of cases are pending now; this one looks to be particularly important. Its fate will help determine if the American public will have free access to the Internet, or whether we'll be stuck behind a corporate-administered Great Firewall of Bell, paying top dollar for limited choice.
And with an Internet in monopoly hands, the FCC's excuse for broadcast ownership deregulation (extreme concentration of ownership of the media) is proven a lie. But Powell hopes we don't notice.
This is standard process in the database biz, including things like mailing lists and (as others have noted here) maps. The term for it is "salting". Calling them "honeytokens" is applying the wrong seasoning... and treating it as new on /. is also silly.
Excellent post. I was going to write a similar note but I'll second this one and add to it.
Circuit-switched telephone networks aren't actually costly to build nowadays -- the competitive-bid price of circuit-switching (TDM) gear is a small fraction of what it was 15 years ago. Lucent and Nortel stock suffers as a result. Undersea cable bandwidth is also much, much cheaper. If one evaluates the cost of building a new wireline network from scratch, then TDM/circuit is not much costlier than VoIP; it could even be cheaper. BTW I do this for a living so I see the real prices.
However, most underdeveloped countries do not treat the telephone network rationally. It is a government agency, and its primary purposes are employment and taxation. Actually providing good service is secondary.
So the price of international calls is kept high. Outgoing rates are high because it's viewed as a luxury -- only a small fraction of the public can afford it, and those foreign businesses are rich, so they can subsidize the telehone network's employment role with very high prices. The price of incoming calls is held high too -- they charge high "settlement" fees to international LD carriers, who reflect them in their rates. That's why international call rates vary so widely between nearby countries. The price, then, has nothing to do with cost, and more to do with what a monopoly can get away with.
VoIP's role here is not technological goodness, it's stealth. Users can cloak phone calls inside IP packets that aren't metered at the country's borders. Thus they get around settlement fees on inbound calls, and don't pay the ridiculous outbound rates. This is good, from an economic perspective (arbitrage removes price distortion), but it makes VoIP look relatively better than it should. Government action distortions outside perspectives on technology.
If these countries really wanted to join the wired (and BTW even cell phones are "wired" in this sense) technological world, then they should rethink their adminstrative policies towards telecommunications in general.
Many of these countries own their own telephone company, in whole or with a partner. In some Caribbean states (I don't know about St. Lucia offhand), Cable & Wireless, a privatized formerly-British state-owned company, has the monopoly, and has its tentacles deep into the regulators.
So a WISP would be a competitor, and that's simply not allowed. All communications money must flow to the government or C&W (who pays off the right people). That used to be the system in the USA, and the RBOCs are trying to get it back, but wireless has too many friends here now. Unlicensed packet radios (802.11) are a relatively recent change in the USA. In other countries, the mere right to be an ISP is again a government franchise, and competition is Not Allowed. This helps them censor opposition thought, and keep the price high (again, for the sake of the monopolist).
Interference has nothing to do with it.
Berkeley did not invent LSD, and I don't think that the original quote was intending to denigrate Hoffman's work. Berkeley did, however, play some role in the product's development. While Dr. Leary of Harvard, not Berkeley, was the foremost advocate of various "off-label" uses, Berkeley was the site of considerable uh experimentation, yeah, that's the ticket, which was for a time quite influential. Yep, trust Berkeley to have plenty of volunteers for experiments with mind-altering pharmaceuticals....
Digression... actually, post-mortem recounts of the entire state show that Gore would have won had the entire state been recounted, as the Bush side had once suggested. Had the recount only occurred in areas where Gore suggested it was necessary, Bush would have ended up ahead. But even that leaves out the impact of the butterfly ballot, of the roadblock keeping African-American voters away from the polls on the panhandle, and from the thousands of eligible African-American voters stricken from the rolls under a contract to ChoicePoint, while at-the-poll appeals were primarily made available to Cubans.
Just trying to fight a defective meme before it spreads.
As a consultant to the telecom industry, I have to pay attention to what works, not what sells papers. This year's big hype is WiFi, which was designed for room-sized LANs but somehow seems to have captured the imagination of the public as if it could actually hurt the telcos (ILECs). VoIP has been hyped since '96 or so, and has eaten tons of vulture capital, and while it has nice niche applications, it is still no substitute for the Real Thing.
Yes, you can run voice over IP. Yes, you can run IP over wireless. Heck, 16 years ago I was running IP over 1200 bps Aloha AX.25 packet radio links. Very instructional, because Phil Karn's NOS let me watch a decoded protocol trace of passing packets, and they came so slowly that I could study all of them in real time. Think about it. The point is that you can't run voice over *any* IP, just some paths.
Circuit-switched telephony is cheap to build. Sure the existing telco networks are made of gear that they paid a lot for, but ILECs depreciate gear over 20+ years. So the Lucent 5ESS and Nortel DMS-100 are VAX-era hardware. What did a MIPS cost when the 5E was designed? Modern circuit switching (which CLECs and some small ILECs buy, not to mention the PBX market) uses modern parts. The switching hardware is only a little costlier than IP stuff, and it sounds better. All the sexy call control features are in the control software, which in a modern system is agnostic about physical-layer protocol. So you can do nice things on circuit, ATM, or IP. Just a different card in the switch.
WiFi's limits are obvious -- there's finite spectrum, and it's shared with domestic cookers (microwave ovens are right in the middle of its band!), cordless phones, VCR "multipliers", baby monitors, and all sorts of other crap. WiFi5 is cleaner spectrum, though the lower-volume gear is costlier. The 5 GHz band will benefit from a recent FCC rule change that adds 275 MHz more bandwidth. But unlicensed still means low power, and either very short range *or* directional antennas (which take more work). And you have to worry about things like hills and trees.
I'm always looking for alternatives to Bell wire -- that's really a big part of my job! But WiFi ain't it. There are non-WiFi radios that are better for "last mile" purposes (and slower, because they have to trade speed for range -- see Shannon). The FCC is contemplating making some additional frequencies available, and in rural/exurban areas, especially flatland, wireless can do wonders. In hilly or woody areas, it's tougher. In urban areas, spectrum is too limited. Fiber optic bandwidth is infinite -- there's lots of sand out there, and only one radio spectrum.