The Need for Speed
Via Brad Delong, Mark Cuban wonders why the music industry won't create any cool services that people actually want to use. I wonder that sometimes myself. But what I really wonder is why the internet isn't faster, damnit. Home broadband internet access (via cable or DSL) like everyone's got nowadays is pretty fast (which is good) and pretty cheap (which is also good), but normally in life if you've got something that's pretty good and pretty cheap you can pay somewhat more and get something somewhat better. And as I say, my internet access is pretty cheap. I'd happily pay somewhat more for something somewhat better. But I can't. No one will sell it to me! It's an outrage. A sin against capitalism. I can't be the only one who's eager to give telecommunications companies somewhat more money in exchange for somewhat better service. So where are the companies? This is no way to run a market economy.
Relatedly, shouldn't Apple build an iPod that's also a cell phone? It seems to me that if you take an iPod Mini and crazy glue it to one of your smaller cell phones, you've still got a pretty damn small device. And surely some bright engineers out there in California or Finland or wherever can think of a better solution than that. But just take a look at this thing and you'll see clearly that anything is possible. At any rate, I promise to get back to serious stuff in the near future.
UPDATE: I see I'm not the only one who wants an iPod Phone for Christmas.
November 15, 2004 | Permalink
TrackBack URL for this entry:
Listed below are links to weblogs that reference The Need for Speed:
Tracked on Nov 15, 2004 10:11:48 PM
Dedicated T-3, be the best on your block. Just a little bit more money.
Jeez. 300-800 kBs not fast enough.
Posted by: bob mcmanus | Nov 15, 2004 12:18:42 AM
I believe that part of the problem here is technological. Can we transmit data faster than DSL or cable can support? You betcha. However, the hardware to achieve this task is simply not present in the majority of homes. In the case of the best technologies, the technology isn't present in any homes. The reason we have DSL and cable isn't that they are the best data transmission technologies, but that they are the best data transmission technologies that work on hardware that is already installed into most or all homes. DSL works over phone lines, cable over... cable.
It's called the 'last mile' problem. The US is stupidly well wired everywhere but the last mile (more like 100 - 1000 meters in cities) that leads into your home, through the wall, and into your computer. That last mile is current bridged by, frankly, obsolete technology. The niftiness of DSL was that it made that obsolete technology (copper phone wire that is often over 50 years old) roll over and play new tricks.
So to get better bandwidth to and from homes, we need one of two things. We need better hardware going into homes, or we need newer new technology to make use of the phone wire or cable that's already there.
Aside: this is one of the reasons that serious wireless is an interesting technology. It's not just mobility, it's also the ability to turn homes on to broad(er)band at literally orders of magnitude less cost than previously.
Posted by: NBarnes | Nov 15, 2004 12:19:17 AM
You might be able to get a T1 (1.544 megabits per second) for around $1000/mo.
Posted by: sc | Nov 15, 2004 12:37:55 AM
In DC, try Starpower, the other cable provider. I just signed up for their cable/internet/phone package, although I haven't used it yet. Very cheap, and the broadband at least appears to be much faster than Comcast's.
We're actually fortunate to live in an area served by a broadband cable duopoly. The best way to encourage competition is to use it.
I'd be surprised if faster is really the issue. I believe you DC residents have decent wireless broadbandish service everywhere. Verizon I think.
Posted by: Atrios | Nov 15, 2004 12:42:47 AM
And where the heck are those flying cars?
Posted by: epistemology | Nov 15, 2004 1:24:20 AM
NBarnes is partially correct. Yes, home DSL just about maxes out the technological limits of what you can transmit over a copper pair.
But that is not true for coax. Cable modems can be pretty damn fast. What MY is asking for is the right to pay a premium to keep other people off his local cable loop. Not nearly as difficult to achieve.
Posted by: space | Nov 15, 2004 1:42:04 AM
Two points: NBarnes mentions that DSL is nifty as it is a way around the last mile problem. True. Further, the technical specs of DSL allow one to achieve data rates better than T1 (heavily dependent on the length of the loop of copper, if one is on copper. Different game if your on pair-gain). Try and get the phone company to sell it to you. Good luck. At my old house, I could go from ~600K download to around ~ 5,000k download, and easily be within spec. The price went from ~$50/mo to around ~$500+/mo, though.
Which dovetails nicely with my second point: the phone company is most certainly not your typical market economy. It is the end-state of said market: a monopoly, heavily regulated and closer to a command economy, I would guess. (The regulation is not the problem, though. It's the fact that local loop is a complete monopoly.)
Posted by: Timothy Klein | Nov 15, 2004 1:43:44 AM
If you expect Apple to make an iPod cell phone you don't understand what Apple is about.
What they aren't about is competing in markets they can't compete in profitably.
They aren't about to get into a race to the bottom
Posted by: 16 | Nov 15, 2004 1:44:28 AM
"At any rate, I promise to get back to serious stuff in the near future."
Screw serious stuff. Everyone needs some form of cat-blogging.
And BTW, you ought to have a weekly The Wire thread.
Thanks to those with more technical backgrounds helping to fill in my points.
Posted by: NBarnes | Nov 15, 2004 3:01:08 AM
". I can't be the only one who's eager to give telecommunications companies somewhat more money in exchange for somewhat better service. So where are the companies?"
Busy selling highly profitable T-1 lines whose appeal would be undermined by providing the sort of service you desire? I'm not joking - I used to cover the telecoms industry for a living, many moons ago.
The last-mile issue is real, but it's hardly as pressing as some might imagine. I was able to get 8Mbps DSL service from Verizon in Manhattan way back in 2000, and VDSL can provide up to 52 Mbps over twisted-pair wires, so it's clear that the technology to support much higher levels of service is there. The problem is that doing so would undermine large swathes of extremely profitable leased-line business, so the incumbents have every reason to drag their heels as much as possible.
I'll bet the sound quality on that penny sized nokia is crap. But then again the quality of the sound will depend on what the user needs or wants.
Most of what these people are doing is simply trying to keep up with the fad craze. The music industry is mired in the past still trying to fight cassette tape recorders, and now mp3, P2P download programs.
Cell phone makers are evolving based on both the needs and the fad craze of the more culturally developed nations. The japanese often drive this fad craze while the fast-paced business elite, aka salesmen, often drive the more practical needs.
Logic suggests that this will eventually evolve from penny-sized cell phones down to a practical integration system whereby communication systems will begin providing visual, data as well as audio communications all in a package that is hidden from the naked eye, allowing it to blend in with fashion or professional appearance needs.
While I worked for Sprint, the planners were working towards an all-in-one gadget that met the criteria I gave above. The only obstacle was compacting quality of sound, data Rx/Tx speeds, and storage space. Some are even now planning a holographic projection system, but do not foresee that for another 10+ years.
The worst, and possibly most controversial aspect is the idea of organic integration, or OI or short. The notion of creating organic circuitry that can allow a person to 'plug in' him or herself to a system and direct download to internal biological storage devices is highly controversial. This will be accompanied by the new hydrogen-based computing systems which involve single atom conductors, and the ability to use a molecule's electron spin as a digital value 0-1, or fuzzy logic. They've already had such primative systems since 97', and they were calculated as having the power to be 1-100 million times as powerful as today's fastest systems.
They're making progress, but unfortunately the music and recording industry will lag behind as long as they insist of hoarding their profit margins at their highest levels.
Posted by: MYOB | Nov 15, 2004 9:06:03 AM
Well, you actually *can* buy fiber to the premises service from Verizon -- see here. 15 MPS, $50/month... not bad (although availability is still limited).
But watching this get rolled out has made it clear how badly this system works. Verizon dragged their heels because they didn't want to have to share their new fiber network with competitors, as they are legally obligated to do in certain states. I'm not keen to legally endorse their monopoly status any further, but it's hard to see the benefit to consumers when networks are pried open -- with DSL, this practice has generally meant the same service is offered under a different brand for another $5 a month. Customer service differences are usually not sufficient to justify that price.
Broadband service requires a lot of infrastructure and competitors can't easily differentiate their products (without a massive capital investment in infrastructure, that is). Seems like a natural monopoly to me. I'm sure people who know more than me about this will shout me down, but I'd prefer to have the city government partially or wholly involved in the data pipe going into my house, the same as the gas or water line.
I know socialized solutions are supposed to stifle innovation, but as Matt notes, in this particular situation we don't need innovation -- consumer technology isn't taking advantage of several generations of network tech. We don't need to offer incentives to squeeze more performance out of decrepit copper lines, we need to build genuine urban broadband infrastructures with coordinated building.
Would Jobs demand to own the cellular network that the iPod Phone runs on?
Stick a digital satellite radio receiver in that iPod cellphone and your headset will permanently chomp onto your ear with its little blue teeth.
Posted by: Dick Eagleson | Nov 15, 2004 10:10:08 AM
You could try speakeasy. they offer 6.0down/768up (depending on your location), 24/7 support via people who are good at their jobs, and some other goodies like static IP's and no port blocking. Don't know if they would make your "internet faster" but they are a company that charges a bit more than the competition and differentiates themselves as providing extra value for that money.
Posted by: billy pilgrim | Nov 15, 2004 10:30:04 AM
tom and Timothy are right. Tom is hoping for municipal broadband because broadband is a natural monopoly. That's right. Timothy remarks: "The regulation is not the problem, though. It's the fact that local loop is a complete monopoly." Right. Others are saying that the problem is that the teleco's don't want to cannibalize their basic services (analog phone) with true broadband (which would enable VOIP). That's right, too--and the logic of a monopolist. Real competition would quickly end that. Trouble is you will only rarely get real monopoly in natural monopoly situations. Public ownership is the only way to get a monopoly business to act in the interests of the public.
My one quarrel with the post is that broadband is _not_ cheap. The folks that live across the street from me have just had to drop their phone service because they've hit a rough patch and just can't afford it. Municipal or state-wide broadband could dramatically reduce the costs and make using broadband possible for many, many on the other side of the digital divide.
We're about to build municipal fiber here in Lafayette, Louisiana -- final approval should come tomorrow night--(amazingly, see LafayetteProFiber/blog for details) and Utah and Iowa have citizen initiatives rolling to interconnect their cities' fiber utilities.
This could be a great issue for progressives if they'd only see it. Remember the _original_ political reference for progressivism? Local and state movements to reclaim their governments in a fight against the big corporate monopolists and their sleazy political pals? That can still work.
Oh yeah, Matt's point: speed. Our system should begin at 40 megs. Or so I hear.
I'd rather not have an iPod phone.
You just *know* they'd make you buy monthly packages of music minutes, and you'd have to pay out the ass if you went over.
I'd prefer a cellphone that is darn-near disposable, so that changing providers is painless. I wouldn't want to feel locked into a given cell service provider, by buying a particular device, then loading it up with personal data, songs, etc.
Posted by: Jon H | Nov 15, 2004 11:00:15 AM
And where the heck are those flying cars?
I'm not keen to legally endorse their monopoly status any further, but it's hard to see the benefit to consumers when networks are pried open -- with DSL, this practice has generally meant the same service is offered under a different brand for another $5 a month. Customer service differences are usually not sufficient to justify that price.
Actually, in my area, local third party vendors sell DSL services at the same price or cheaper than the phone company, and without the klunky co-branded Yahoo! customized browser and other fluff the phone company sells it with, and with more speed and other options for home users, and -- unlike the phone company -- they actually get it turned on, and don't give multiple and mutually inconsistent lies for a period of weeks over why they haven't gotten it activated.
Posted by: cmdicely | Nov 15, 2004 11:46:47 AM
Here's the Apple product I'm waiting to get:
A box on top of my TV -- worked by a remote -- with wireless access to my Mac's Airport card. It uses my Mac's internet access and hard drive and turns it into a DVR. I can access on-demand video via an iTunes-like interface (have to download it overnight, but still better than going to Blockbuster.)
Advantages over current DVRs? Wouldn't have to pay for the HD, since I'm using my own. Smaller; all the heavy lifting is done in the computer. You can do the programming with a keyboard and mouse, rather than the clunkier remote. Easier to install. And if it has the Apple elegance of design, it becomes a killer app.
Posted by: wagster | Nov 15, 2004 11:51:22 AM
I'm a software engineer and know a little
about networks. Everybody talks about
network bandwidth, but a lot of the
problem is network latency, and
especially round-trip latency - i.e.
if I send a 1-byte request from my
machine in Boston, MA, and it has to
get to a server in California which
then sends a 1-byte reply, the round-trip
time may be long (say 100msec) even if
the bandwidth is high.
In practice, this means that upgrading
from 800Kbit/sec to 10Mbit/sec might
not achieve the effect you would expect
Downloading large documents and pictures would go faster (because those
protocols are designed to use large
packets and tolerate high latency),
but most other activities would be
not noticeably faster.
Some of these latency effects can be
improved by cacheing and the use of
faster switches and servers; but the
bottom line is the speed of light -
if you're talking to a server 6000km
away, the round-trip is 12000km and
light goes 300000km/sec, so it takes
at least 12/300sec = 0.04sec.
Posted by: Richard Cownie | Nov 15, 2004 11:58:39 AM
NBarnes has hit a large nail on the head with his mention of the "last mile" and the capabilities of in-home appliances.
There are, however, moves afoot - although they will likely take some time to sort themselves out. Verizon in particular is slowly, carefully, almost tentatively bringing a new solution to market - called "FTTP" (for Fiber To The Premises) which will, as its name suggests, bring fiber into your home. But this is a huge task on many different levels, not the least of which being domestic equipment capable of working at optical rates - the least of which (STS1) is the equivalent of one T3, twenty-eight T1's, or seven hundred and forty eight phone lines. Simply being able to route traffic at these rates is technically demanding, but for a household appliance to be able to exploit all that bandwidth usefully, and in a way that people could unthinkingly use, remains a greater challenge still.
cmdicely: Actually, in my area, local third party vendors sell DSL services at the same price or cheaper than the phone company, and without the klunky co-branded Yahoo! customized browser and other fluff the phone company sells it with, and with more speed and other options for home users, and -- unlike the phone company -- they actually get it turned on, and don't give multiple and mutually inconsistent lies for a period of weeks over why they haven't gotten it activated.
Well, then that's impressive, and encouraging. My experience in Charlottesville, VA was very different -- it involved pricing a bunch of different companies and finding no advantage among the little guys. It was easy enough to tell that they were all working off Sprint's central office (judging by their "do you qualify" web wizards), and all of them were more expensive than Sprint's service. If they can get the service up faster then that's good, but I doubt either of us has a large enough sample size to say that for certain -- and my hunch is that the added layer of bureacracy can only complicate things.
fcb: Simply being able to route traffic at these rates is technically demanding, but for a household appliance to be able to exploit all that bandwidth usefully, and in a way that people could unthinkingly use, remains a greater challenge still.
I disagree. HDTV should be able to swallow that bandwidth pretty easily. But more to the point, the fiber service Verizon is rolling out only delivers 15 Mbps downstream (yes, small b). That's slower than current wireless standards, for christ's sake. Makes me wonder whether the fiber is just a marketing ploy. Gigabit ethernet would presumably be cheaper and still orders of magnitude faster. Hopefully they're building for the future and will magically enable speed increases as the market requires.
Finally, I think Richard Cownie's point about latency is overstated. The class of latency-sensitive applications is pretty limited -- twitch video games are the classic one; voip might be a better example for this audience. But even within this class, if you can move more data during each roundtrip, you can end up with better quality, if not better responsiveness. You'll never get much below 50 ms of latency, it's true. But it's also true that it mostly doesn't matter. Nobody is demanding faster IM. Lots of folks would like video-on-demand, though.
We will hit a usable consumer bandwidth ceiling -- humans only have so much sensory bandwidth, after all. But we're still a ways away.
The comments to this entry are closed.