The second was responsible for such things as the transistor, the solar cell, lasers, synthetic crystals, high fidelity stereo recording, sound motion pictures, radio astronomy.....
Er, do you know how much money AT&T spent to develop those things? It's a phenomenal amount, made even more poignant by the tiny number of success stories relative to failures (this is not meant to be unfair to Bell Labs, as almost all pure research labs have this necessary characteristic.)
The reason AT&T was able to do so much for the world was because the American people were footing the fabulous bill (not to mention the corporations who were and still are paying license fees for a bunch of this stuff.) If AT&T is your idea of a good business, then I'd suggest you look at what's happening to them now that they actually have to compete.
The most embarrassing thing about AT&T, while we're on this subject, is that now that they are forced to compete, it turns out that their data network is years behind the competition. Why? Because AT&T felt like they could just go on laying copper while Sprint and MCI put fiber into the ground. Oops.
Yes, and we all know how well government regulations helps competition!
Unfortunately competition and monopolies are antithetical. Unless you can come up with a plan to convince Verizon et al. to voluntarily fragment into a dozen competing ISPs, government regulation is the lesser of two evils. Or we could just deregulate, deregulate, deregulate and pick up the pieces later like we did in CA.
It's quite a stretch, isn't it? But it does make sense when you think about it. You want increased competition, the way to do it is not to give more power to the local monopoly. If they weren't a monopoly (or close to one) I'd say I was all for it. Unfortunately in this day and age we seem to have forgotten the messes that forced us into regulating our utilities in the first place.
Make the legit content more valuable (by value-added features and services and affordable cost and easy access) than the pirated content (difficult to track down, unreliable quality, no value-added features and services).
I guess you and the Rep. agree on more than you think, because that's exactly what he said in his response.
As far as control, the music companies have had several years to realize that they have absolutely no control over the current media (CDs). Any fool can rip an MP3 straight from a high-quality digital source and distribute it. Napster's shown us that. Therefore, any industry solution would be an improvement, no matter how flawed the technical or legal mechanisms protecting it. Even distributing raw MP3s wouldn't make them much worse off than they are now.
The truth is, the music companies have categorically rejected every proposal to create a secure distribution channel. They've done this in the face of some of the best minds in the security industry devoting their efforts, and in the face of the beneficial (to them) findings in the DeCSS case so far.
As a coda, it would be worthwhile to point out that so far the old-fashioned distribution methods of the recording industry have worked largely in our favor. Imagine trying to argue 'first-purchase' rights if we couldn't actually own our own music? That's the next logical step that the recording industry will take, and it will move us out into some uncharted legal territory (bad for us, good for them.)
Despite what we are given to understand, business is not a science. Much of what is being done today in the interest of shareholders will someday wind up in a textbook under the 'how not to run a company' heading. Having purchased a company, many large corporations focus on pulling as much value out of a brand as possible, even at the expense of the brand's (short or long term) future survival. Often this is seen as the 'correct' thing to do, in that it maximizes shareholder return.
Unfortunately, following a vanilla formula not only reduces the potential growth of a company, but it often threatens the value of the company to boot, by both turning off future customers and driving away the creative people who made the business profitable. Even in the short-to-mid-term, this winds up costing shareholders money, or at least reducing profits that could be made simply by applying a more delicate managerial style.
I can't quite tell if you're being cynical or serious. I guess that's my problem.
I think the fact that a lot of slackers got sucked into and vomited out of the dot-com revolution shouldn't lead you down the garden path. Lots of very creative and intelligent people at real companies (ie, creative/original/interesting) simply can't survive the corporatization process. They build the company, making it what it is. Then they leave, and the company simply becomes another piece of intellectual property, a feather in the cap of Disney, or whoever, which is equipped to bring the company into phase 2 of its lifespan (wherein it is quietly milked for ongoing profits.) Little groundbreaking work gets done, and the company lasts as long as its brand will sell action figures.
There are companies that have been run for years by the original founders, and built into empires by the same people who would otherwise have been driven away by the merger.
On a slightly different topic, have you noticed that Tivo has begun calling their boxes "Standalone Units"? Does this imply that Tivo is going to move into cable head-ends in the future? If so, we may all be enjoying the last days of fast-forwarded ads.
Face it, your individual viewing data is WORTHLESS. You're just not that important of a person.
That is so twentieth century. Actually, your personal viewing data is gold, assuming Tivo's advertisers can make use of it for significantly less than they stand to gain from it. Databases driving targeted advertising make this sort of use easily afforable, and such tactics will only become more sophisticated as time goes by.
Personally, I don't think it's such a bad thing. I'd pay money not to have to watch another SUV ad. But I am a little worried by the studies that show each additional TV-watching hour corresponds to an additional $X/mo (fill in the blank here) spent on consumer goods. One can only assume that targeted advertising will exacerbate this problem, driving many people further into debt.
The problem with the current ratings system, which uses a very small number of viewers to extrapolate the habits of the entire TV-watching public, is that it often rounds off niche audiences. There may be many thousands of people watching a particular show, but the Nielsen system can often round those people off into obscurity (especially as it does not include any demographic information on those viewers.) If you're really interested in quality TV, you should be glad that your preferences are being heard, rather than extrapolated from a bunch of couch potatoes watching "Full House".
OTOH, the bandwidth usage of gnutella searches vs. the total bandwidth available is a very small ratio. For this particular application, I don't think it'll be terribly important, but it's a good thing to think about
We did some very preliminary research on the scalability of gnutella-style search systems for a project at work. The quickest summary is yes, the search traffic is relatively low-- but it's "low" in direct proportion to the number of nodes. Since each node is supposed to hear searches from the majority of the network (or at least, those nodes within a certain distance), it's pretty easy to screw up everybody's day once the network gets large enough.
And of course, those nodes connected to narrow pipes can just be blown away by multicasting, whereas at least with the current Gnutella (as inefficient as it is), the TCP acks work as a throttle to control incoming requests.
any layer 3 multicast scheme is going to be more efficient than gnutella currently is by virtue of the fact that physical network topology can be taken into account at layer 3.
So all you're after is an architecture that's aware of network topography. You're absolutely correct in that Gnutella is currently very poor in the way it builds the network and distributes messages. Layer 3 multicast would be an improvement in this area, but unfortunately it has serious implementation flaws that make it extremely difficult to implement on a large network. However, the real problem with multicast (for this application) is that once you have subscribed to a channel, you have no reliable controls on traffic flow (outside of what your routers are able/configured to allow.) In other words, should the Gnutella client wish to throttle back traffic, you're outa luck (existing traffic reporting mechanisms are going to be inadequate for this application.)
How is this reducing the burden on routers?
Handling multicast routing puts a significant burden on a router. The more traffic, and the more complex the multicast routing table, the worse this is. Gnutella represents the worst possible case for a multicast network: many short messages from many sources. Not to mention that when a router fails, the table has to be rebuilt. If the failing component is a switch, this often results in broadcast traffic on the LAN (and this is really the least of the potential difficulties.)
Better yet, please explain to me how what gnutella does isn't multicast, and how what gnutella does is better for any segment of the network than a good multicast implementation would be?
It is a form of multicast. It's not layer 3 multicast. Was I unclear there? I'm sorry. most people call it "application level multicast", although that also implies a certain consciousness of topography. My point was simply that layer 3 multicast has serious flaws that would make it less than ideal for this purpose. That's aside from the fact that it's absolutely not going to be implemented on the major backbones, which makes the issue completely academic. The only real solution is for Gnutella to become more intelligent in the way it routes messages. It's not surprising that it's so stupid, it is after all a first generation technology.
Gnutella uses no form of multicasting at the session layer or otherwise. Multicasting, when a protocol uses IP for at the network layer, by definition must be performed at layer 2/3 or else it's unicast, period
I believe the buzzword is "application level multicast." A lot of companies (eg FastForward) are trying to implement this sort of thing as the solution to the lack of multicast support on the backbone.
Can you imagine the nightmare of implementing Gnutella using multicast (if such a thing were possible on today's net, which it's not.) You'd still have all the traffic problems you have with the current system, only now they'd be weighing down routers. Even if the traffic were routed over a breadth of multicast addresses (which would eat into the supply), imagine what happens when you subscribe to a channel.
Not only does your connection get hosed, but your ISP/IT dept. probably shuts down multicast for the afternoon and has some very stern things to say to you.
I'm not sure I understand the question in the article. If you run a Gnutella node, you're not magically exempt from responsibility for what you host. Protocol is completely irrelevant. The only reason G users can get away with more is that they're harder to track down. This could of course be solved by someone with enough money and desire to take out lots and lots of small, personal nodes.
Note that they wouldn't actually have to take anyone to court, by the way. A quick call to your ISP would be enough to take care of the problem.
There was a study which found that a huge amount of the files shared were hosted on a small number of Gnutella nodes. The RIAA was very smug in predicting that these were the people they'd attack if they ever wanted to take Gnutella out (if/when, that is.) What I don't think they've taken into account is the number of nodes that will take their place. Such a strategy might lead to a nice democratization of the network, forcing lurkers to share some content.
If I'm correct about this, the Computer Security arm of the NSA is kept slightly apart from the real thing. That is, it's located in a different place, and run by PhDs. Their security level is lower, and their stated mission is to be proactive in securing networks, not intelligence collection as with the normal NSA. I believe these are the same people mentioned in The Cuckoo's Egg.
Why not get rid of the servers altogether? If you want Jabber to reach even a fraction of the size of AIM, the total cost of running those servers is going to become substantial. Are fully distributed p2p that far off?
You could build this as a Gnutella-style system. The messages would obviously go p2p, and although there would be many more nodes, there would be a lot less traffic. The primary broadcast message would simply be "hi, I'm logged on and here's my ip, any of my buddies out there" messages. Each node could cache the IPs of thousands of other nodes, so you could probably accomplish a lot even if your pings couldn't reach the other side of the cloud. You'd probably want some way to tunnel through firewalls, but that's certainly not impossible.
In other words, IRC without the servers. I'm sure someone's already doing something like this.
finally, DeCSS does not enable copying -- you can copy bit-for-bit just fine, and then press unauthorized DVD's which will play on authorized DVD players; this is happening on a mass scale in China and other foreign countries.
It's worth pointing out that CSS is probably a trial for the more complicated anti-copying measures sure to come next. If they lose this case, they won't be able to control the players and will therefore have to try a different approach (like some sort of license agreement, or pay as you go real-time streaming.)
if it doesn't fly and 2600 is "convicted", then the precedence becomes that are own laws take precedence over treaties.
Not at all. The WIPO agreements simply left anti-circumvention restrictions up to the laws of the individual nations. It does nothing to prevent anti-circumvention laws from being passed. I could dig back through the brief and find the actual language, but I'm a little too lazy this morning.
when one of those VeriSign things pop-up, you have an options to check "Always Trust Xyz Corp"
That dialog refers to the organization that signed the certificate. Most browsers (at least, IE and Netscape) come equipped to trust any certificate signed by Verisign. When you go to a page with a Verisign cert, the browser will trust the certificate, regardless of what company actually owns it.
Since in this case the certs were purchased from Verisign, your browser won't have any problem at all with them (it'll just assume that Verisign is trustworthy.) You won't get that dialog at all. If you look at the security info for that page, it'll show the page as registered to Microsoft corporation. Generally MS signs their own certificates, so it would be a little odd to see a cert owned by MS and signed by Verisign (although they may actually do this.)
They can certainly make the layout of their site unfriendly to ad-filtering programs. These new ads are enormous. If they tweak their layout enough, they can make the pages unpleasant to read without serious re-working. I think their plan (and the $30) is a little too ambitious, though. I love Salon, but I think this is going to hurt them.
GUIDs consist of a unique system serial number and a time stamp. That's what makes each one 'unique'. The system serial number should be significantly smaller than the full GUID.
Slashdot Mirror
Will these lines be installed by the same people who do highway repair?
Er, do you know how much money AT&T spent to develop those things? It's a phenomenal amount, made even more poignant by the tiny number of success stories relative to failures (this is not meant to be unfair to Bell Labs, as almost all pure research labs have this necessary characteristic.)
The reason AT&T was able to do so much for the world was because the American people were footing the fabulous bill (not to mention the corporations who were and still are paying license fees for a bunch of this stuff.) If AT&T is your idea of a good business, then I'd suggest you look at what's happening to them now that they actually have to compete.
The most embarrassing thing about AT&T, while we're on this subject, is that now that they are forced to compete, it turns out that their data network is years behind the competition. Why? Because AT&T felt like they could just go on laying copper while Sprint and MCI put fiber into the ground. Oops.
Unfortunately competition and monopolies are antithetical. Unless you can come up with a plan to convince Verizon et al. to voluntarily fragment into a dozen competing ISPs, government regulation is the lesser of two evils. Or we could just deregulate, deregulate, deregulate and pick up the pieces later like we did in CA.
It's quite a stretch, isn't it? But it does make sense when you think about it. You want increased competition, the way to do it is not to give more power to the local monopoly. If they weren't a monopoly (or close to one) I'd say I was all for it. Unfortunately in this day and age we seem to have forgotten the messes that forced us into regulating our utilities in the first place.
I guess you and the Rep. agree on more than you think, because that's exactly what he said in his response.
As far as control, the music companies have had several years to realize that they have absolutely no control over the current media (CDs). Any fool can rip an MP3 straight from a high-quality digital source and distribute it. Napster's shown us that. Therefore, any industry solution would be an improvement, no matter how flawed the technical or legal mechanisms protecting it. Even distributing raw MP3s wouldn't make them much worse off than they are now.
The truth is, the music companies have categorically rejected every proposal to create a secure distribution channel. They've done this in the face of some of the best minds in the security industry devoting their efforts, and in the face of the beneficial (to them) findings in the DeCSS case so far.
As a coda, it would be worthwhile to point out that so far the old-fashioned distribution methods of the recording industry have worked largely in our favor. Imagine trying to argue 'first-purchase' rights if we couldn't actually own our own music? That's the next logical step that the recording industry will take, and it will move us out into some uncharted legal territory (bad for us, good for them.)
Unfortunately, following a vanilla formula not only reduces the potential growth of a company, but it often threatens the value of the company to boot, by both turning off future customers and driving away the creative people who made the business profitable. Even in the short-to-mid-term, this winds up costing shareholders money, or at least reducing profits that could be made simply by applying a more delicate managerial style.
I think the fact that a lot of slackers got sucked into and vomited out of the dot-com revolution shouldn't lead you down the garden path. Lots of very creative and intelligent people at real companies (ie, creative/original/interesting) simply can't survive the corporatization process. They build the company, making it what it is. Then they leave, and the company simply becomes another piece of intellectual property, a feather in the cap of Disney, or whoever, which is equipped to bring the company into phase 2 of its lifespan (wherein it is quietly milked for ongoing profits.) Little groundbreaking work gets done, and the company lasts as long as its brand will sell action figures.
There are companies that have been run for years by the original founders, and built into empires by the same people who would otherwise have been driven away by the merger.
On a slightly different topic, have you noticed that Tivo has begun calling their boxes "Standalone Units"? Does this imply that Tivo is going to move into cable head-ends in the future? If so, we may all be enjoying the last days of fast-forwarded ads.
That is so twentieth century. Actually, your personal viewing data is gold, assuming Tivo's advertisers can make use of it for significantly less than they stand to gain from it. Databases driving targeted advertising make this sort of use easily afforable, and such tactics will only become more sophisticated as time goes by.
Personally, I don't think it's such a bad thing. I'd pay money not to have to watch another SUV ad. But I am a little worried by the studies that show each additional TV-watching hour corresponds to an additional $X/mo (fill in the blank here) spent on consumer goods. One can only assume that targeted advertising will exacerbate this problem, driving many people further into debt.
The problem with the current ratings system, which uses a very small number of viewers to extrapolate the habits of the entire TV-watching public, is that it often rounds off niche audiences. There may be many thousands of people watching a particular show, but the Nielsen system can often round those people off into obscurity (especially as it does not include any demographic information on those viewers.) If you're really interested in quality TV, you should be glad that your preferences are being heard, rather than extrapolated from a bunch of couch potatoes watching "Full House".
We did some very preliminary research on the scalability of gnutella-style search systems for a project at work. The quickest summary is yes, the search traffic is relatively low-- but it's "low" in direct proportion to the number of nodes. Since each node is supposed to hear searches from the majority of the network (or at least, those nodes within a certain distance), it's pretty easy to screw up everybody's day once the network gets large enough.
And of course, those nodes connected to narrow pipes can just be blown away by multicasting, whereas at least with the current Gnutella (as inefficient as it is), the TCP acks work as a throttle to control incoming requests.
So all you're after is an architecture that's aware of network topography. You're absolutely correct in that Gnutella is currently very poor in the way it builds the network and distributes messages. Layer 3 multicast would be an improvement in this area, but unfortunately it has serious implementation flaws that make it extremely difficult to implement on a large network. However, the real problem with multicast (for this application) is that once you have subscribed to a channel, you have no reliable controls on traffic flow (outside of what your routers are able/configured to allow.) In other words, should the Gnutella client wish to throttle back traffic, you're outa luck (existing traffic reporting mechanisms are going to be inadequate for this application.)
How is this reducing the burden on routers?
Handling multicast routing puts a significant burden on a router. The more traffic, and the more complex the multicast routing table, the worse this is. Gnutella represents the worst possible case for a multicast network: many short messages from many sources. Not to mention that when a router fails, the table has to be rebuilt. If the failing component is a switch, this often results in broadcast traffic on the LAN (and this is really the least of the potential difficulties.)
Better yet, please explain to me how what gnutella does isn't multicast, and how what gnutella does is better for any segment of the network than a good multicast implementation would be?
It is a form of multicast. It's not layer 3 multicast. Was I unclear there? I'm sorry. most people call it "application level multicast", although that also implies a certain consciousness of topography. My point was simply that layer 3 multicast has serious flaws that would make it less than ideal for this purpose. That's aside from the fact that it's absolutely not going to be implemented on the major backbones, which makes the issue completely academic. The only real solution is for Gnutella to become more intelligent in the way it routes messages. It's not surprising that it's so stupid, it is after all a first generation technology.
I believe the buzzword is "application level multicast." A lot of companies (eg FastForward) are trying to implement this sort of thing as the solution to the lack of multicast support on the backbone.
Not only does your connection get hosed, but your ISP/IT dept. probably shuts down multicast for the afternoon and has some very stern things to say to you.
Note that they wouldn't actually have to take anyone to court, by the way. A quick call to your ISP would be enough to take care of the problem.
There was a study which found that a huge amount of the files shared were hosted on a small number of Gnutella nodes. The RIAA was very smug in predicting that these were the people they'd attack if they ever wanted to take Gnutella out (if/when, that is.) What I don't think they've taken into account is the number of nodes that will take their place. Such a strategy might lead to a nice democratization of the network, forcing lurkers to share some content.
If I'm correct about this, the Computer Security arm of the NSA is kept slightly apart from the real thing. That is, it's located in a different place, and run by PhDs. Their security level is lower, and their stated mission is to be proactive in securing networks, not intelligence collection as with the normal NSA. I believe these are the same people mentioned in The Cuckoo's Egg.
Why not get rid of the servers altogether? If you want Jabber to reach even a fraction of the size of AIM, the total cost of running those servers is going to become substantial. Are fully distributed p2p that far off?
In other words, IRC without the servers. I'm sure someone's already doing something like this.
It's worth pointing out that CSS is probably a trial for the more complicated anti-copying measures sure to come next. If they lose this case, they won't be able to control the players and will therefore have to try a different approach (like some sort of license agreement, or pay as you go real-time streaming.)
Except as far as I know there is no jury. Fortunately, instead we have a passle of young, techno-savvy judges just dying to take up 2600's cause.
Not at all. The WIPO agreements simply left anti-circumvention restrictions up to the laws of the individual nations. It does nothing to prevent anti-circumvention laws from being passed. I could dig back through the brief and find the actual language, but I'm a little too lazy this morning.
That dialog refers to the organization that signed the certificate. Most browsers (at least, IE and Netscape) come equipped to trust any certificate signed by Verisign. When you go to a page with a Verisign cert, the browser will trust the certificate, regardless of what company actually owns it.
Since in this case the certs were purchased from Verisign, your browser won't have any problem at all with them (it'll just assume that Verisign is trustworthy.) You won't get that dialog at all. If you look at the security info for that page, it'll show the page as registered to Microsoft corporation. Generally MS signs their own certificates, so it would be a little odd to see a cert owned by MS and signed by Verisign (although they may actually do this.)
They can certainly make the layout of their site unfriendly to ad-filtering programs. These new ads are enormous. If they tweak their layout enough, they can make the pages unpleasant to read without serious re-working. I think their plan (and the $30) is a little too ambitious, though. I love Salon, but I think this is going to hurt them.
GUIDs consist of a unique system serial number and a time stamp. That's what makes each one 'unique'. The system serial number should be significantly smaller than the full GUID.