Slashdot Mirror
Anti-Technology Technologies?
shanen writes "A story from the NYTimes about metering internet traffic caught my eye. I thought the exchange of information over the Internet was supposed to be a good thing? Couldn't we use technology more constructively? For example, if there is too much network traffic for video and radio channels, why don't we offset with the increased use of P2P technologies like BitTorrent? Why don't we use wireless networks to reduce the traffic on the wired infrastructure? Such technologies often have highly desirable properties. For example, BitTorrent is excellent for rapidly increasing the availability of popular files while automatically balancing the network traffic, since the faster and closer connections will automatically wind up being favored. Instead, we have an increasing trend for anti-technology technologies and twisted narrow economic solutions such as those discussed in the NYTimes article, and attempts to restrict the disruptive communications technologies. You may remember how FM radio was delayed for years; part of the security requirements of a major company includes anti-P2P software, as well as locking down the wireless communications extremely tightly — but there are still gaps for the bad guys, while the main victims are the legitimate users of these technologies. Can you think of other examples? Do you have constructive solutions?"
Bittorrent is a major part of the problem because it attempts to utalise 100% of the available bandwidth (and the client end). If every user used bittorrent, then the ISPs would have to supply 1:1 bandwidth (instead of overselling as they do at the moment), thus dramaticly forcing the price up for every user.
Follow the money. The ones with (power|control|money) want to stay on top and it's only the ones with better agility that corner the market and then become the top dog. So you're looking for a technical solution for the wrong problem.
:p
What's the problem ?
IMHO, it's the "last mile". Legislated limited monopoly controlling access with an interest in keeping that position. so there's a high barrier to access put in place.
Some of the other problems is what may work in a high density area will probably not work in a low density area. A wireless mesh may work in cities and towns but completely fails in rural. Another issue - making data retrieval a crime. "you're" responsible for someone else's actions and that kills any open public access. Some one has to pay to connect to the backbone.
If I had a solution that would work in all cases - I'd be rich
Here's a lynchpin that needs to be remove - the last mile monopoly and its bundling with "providers". Here in the Northeast (US) the power line is a separate charge on your power bill than the generation. Break that up. Internet access "line" charge $0.02 per month. ISP charge $x. Anyone should be able to send data over the lines without the big guys restricting access - for the same cost. NO AT&T ISP should be able to send data cheaper than another ISP.
It may be time for $TOWNs to own the lines, bid repair out to another party and anyone to sign up to an ISP.
BUT it won't work. See any telcom endevor.
The Duck
BTW, is Microsoft paying for the constant annoying updates of its OS, as well as Apple for the annoying connections of Quicktime (and iTunes) and Acrobat for their automated downloads too?
I read an article by a high-ranking Toyota exec in the New Yorker about how, in contrast to American companies, Japanese companies *do* think ten or twenty years in advance. He made the point that they didn't introduce hybrid cars in order to sell to hippies in ca. 2000; they introduced them because, come 2010 or 2015, gas is going to be expensive enough that lots of people are going to want them... and they wanted a mature product -- both from an engineering and from a brand standpoint -- ready to go.
Now, in 2008, Priuses (and Corollas and Yarises) are common on the road in my city, while many of the short-sighted US manufacturers are trying to retool from building 18 mpg SUV's.
The interview mentioned a Japanese business term that has no translation in English; I forget the word, but it meant something like "the faith that building products that people need and selling them for a fair price, long-term, will be profitable, long-term." That might be less true now than it once was, but it's interesting to note that Japanese companies do tend more toward the "Build useful stuff; sell it for cost + profit" model, and American ones toward "Make whatever we can market and sell it for whatever we can convince people to pay".
The main exception to this that comes to mind immediately is Sony, who can go die in a fire. They've got their hands in lots of markets and are thus successful in that regard, but they don't seem to be market leader in any of them. I follow the camera market fairly closely, and Sony's main market in the US seems to be
1) people buying point-and-shoot cameras that didn't do their research, and wind up paying >$100 more than the equivalent Canon or Panasonic that performs better;
2) digital SLR's, which aren't really Sony's; they're rebranded Konica-Minolta stuff who Sony bought out.
As an example of Sony's failing, their top-end bridge camera still doesn't offer any sort of processing controls: you're stuck with a JPG with one compression setting, one saturation setting, one contrast setting, one (excessive) noise reduction setting, etc. There's no RAW mode. The lens is *very* prone to chromatic aberration.
Canon and Panasonic's competitors are cheaper, use superior optics, and offer control over the processing; Panasonic's versions have RAW, and Canon's
But, as a marketing matter, you can't sell stuff like this to Joe Sixpack by saying "Look! Good optics! Controllable processing! RAW mode!", so Sony didn't even bother trying to do this stuff.
The problem you've identified is similar to the spam problem: I could not only annoy people, but cost them money by sending them large unsolicited emails.
A solution for this would be to charge for traffic, but charge the broadcasters, not the consumers. Home users would pay nothing for bytes received, but would be charged for every byte they send -- which is negligible for most home users but would cost prolific file-sharers and people running web sites on their home machines. (As a side effect, this would tend to discourage home file sharing.)
Large web-based businesses would see their costs go up, and those costs would be passed on to the consumer -- the price of downloading music from iTunes would increase, and Apple would pay their ISP, which would use that money to keep their network up, instead of home users paying the ISP themselves.
Tony:
The very fact that you stated "Bittorrent is slower than other methods" of file distribution shows that you have very little grasp on file distribution and limits to bandwidth.
I've had to deal with it directly, as a content producer.
There is not very much more efficient distribution out there than a peer to peer model.
Multicasting is basically about taking you back to the old paradigm where everyone watches the same thing at the same time. (ok, you can save things so people can watch it later, but it is still the old)
Bittorrent could probably benefit some from pairing peers that are locally close to each other, but that is a tracker problem, not a fundamental flaw with the protocol itself. The swarm control overhead is bigger than it has to be because with slow upstreams you need more peers for acceptable download speeds. Swarm control overhead is minimal unless you are running a rogue client. We are talking in the area of a 1% overhead here.
Btw, getting more peers has nothing to do with getting acceptable download speeds. It only spreads what precious little upload you do have out so that peers will be even less likely to trade chunks with you. Actually, if the extra peers you get are seeds, that isn't true. In that case you do get a little extra speed, although not much.
The real solution if you are planning to p2p, is to stop looking at those stupid download X mbit/s number when getting an ISP. Those numbers are only there for people who don't know that upload is what matters in 95+% of all cases. (And in the remaining cases, price is)
But your method has one glaring flaw. Cost and complexity. And planning.
One of the really odd things about p2p technology is that it also distributes the cost of distribution. Whereas you could say the ISP could add more, add more this, add more that--in some and many cases this becomes cost prohibitve to have to keep up that infrastructure, especially if the cost is burdened by one entity.
By using peer to peer technology, that cost is distributed across all of the users simiarly--not equally, but similarly. And in torrent's case, it's also pretty good at making sure users that can use higher bandwidth also need to incur a higher cost. If you can distribute more, the protocol allocates more to give you more to distribute.
It's kind of automatic in nature, self healing, distributive, with no single points of failure and no glaring flaws.
The only reason this is even "argued" is because ISPs didn't build the infrastructure to handle this. And the only other reason this is argued is because a lot of people are selfish and feel the cost shouldn't be burdened by them.
But I'd rather have say, 10 cents here, 20 cents there, than the ISP assuming I'm always going to use 20 cents and charge me accordingly for access to their distribution servers.