The real fight, not the cosmetic fight over ISPs censoring content which they can't do anyways, is over the government setting peering and interconnect prices even though this has always been set by the free market. In this case, the hardline Net Neutrality proponents want to set ISP peering rates to zero, or at least heavily regulated by the ISPs. The FCC tried to compromise by putting out incoherent regulations that would outlaw paid prioritization but not outlaw paid peering which are essentially the same thing (see http://www.digitalsociety.org/2010/12/fccs-utter-incoherence-on-paid-prioritization/). The FCC thought that compromise rules wouldn't get them sued by the ISPs and slammed by most of congress, but that happened anyways.
What happens when someone resolves an IPv6 address and their software and/or IP configuration won't support it? The point is that websites have to be 100% backwards compatible with IPv4 but an IPv6 presence will be optional. The point is that everything/everyone will have to maintain IPv4 compatibility which means there is simply no incentive to go dual-stack.
No, IPv4 will become valuable real-estate. Ask anyone if they prefer a 10 digit phone number or a 40 digit phone number because all the 10 digits ran out. Or a short domain name over a long domain name. The IPv4 addresses will become the original short IPs that all the large companies have already horded. Websites will prefer to be reachable over IPv4 and IPv6 visibility will be secondary. That's because IPv4 is accessible by the whole world one way or another and IPv6 will only be visible to a few early adopters.
The people complaining about not enough young people are the same ones who didn't produce enough offspring. They're having a hard time finding mates due to their unrealistic expectations. They might want to relax the rules for more immigrant spouses from other countries. Yes that dilutes their Japanese gene pool, but it's better than going to zero at the rate they're headed.
Even if you switch to a pubic IPv6 address, all your internal stuff will still be IPv4. My home print server and IP telephony adapter are all IPv4. The problem with IPv6 is that you can't entirely switch to it and just shut down IPv4. You have to run dualstack for the foreseeable future. That's why every IT consultant and IT manager and CIO I've spoken to says they don't give a crap about IPv6 because every adopter of IPv6 will have to be backward compatible with IPv4 so why bother running dual stack. Even after all the addresses are assigned, not a single IPv4 device or network will stop working.
The choice is between IPv4 single-stack or IPv4/IPv6 dual-stack. Given those as the only choices, people are choosing the former instead of the latter. There is no possibility of running IPv6 single-stack. IPv6 will essentially become the new "private IP addresses" that have to translate to "public" IPv4 addresses used by 99% of the IP devices in the world. The only difference is that IPv6 devices will be able to talk to each other without a NAT across organizations.
Your math works out if everyone lived in an apartment complex where your last mile costs are covered by the rent. In that situation, the "last mile" is really the last 100 meters which can be covered with CAT-5 cabling. But even when we look at your math, not a lot of people are interested in paying $389/month for broadband. The typical subscriber is willing to pay $40 and 1/3 of homes don't want broadband at all.
The cost of providing broadband for single family homes is substantially different, and it is in the last mile. AT&T and Verizon for example each have nearly 300,000 employees. When you're providing a single high speed circuit, it's substantially cheaper than providing 20-40 individual circuits to single unit homes. Each home costs about $10K to wire and you have to maintain each of those lines, and provide customer service to ~30 separate user accounts.
"why can so many service providers in northern Europe and southeast Asia provide an extremely consistent 100+ mbps, even at night when virtually everyone is online (say in South Korea), to every single household for anywhere from $10-$50/month?"
They don't advertise at least x Mbps, they advertise "up to" 6 Mbps for example. I got my mom a 6 Mbps U-verse connection and found that their advertising wasn't accurate. Turned out that they gave her 7 Mbps which is generally sustainable even over a long duration. However, I don't expect 7 or 6 Mbps to be an "at least" number.
So you're saying that "up to" means "at least"? Do you not realize that broadband bits cost 20-40 times less than commercial bandwidth, precisely because it's shared 20-40 times? Now you want the government to change the service level of a shared circuit to that of a dedicated circuit? Any idea what this does to prices? Any idea how you'd actually achieve this, since it's impossible to build a core network that can handle all the concurrent data that the end points can throw at it?
I hate having to choose between an SSD and an HDD for a laptop and really want one of each. I want a nice big 500+ GB HDD but I always want a 40+ GB SSD for a boot/OS/applications/page partition i.e., the "C" drive. Then you really get the best of both worlds because you get the insanely fast IO speeds of SSD but you have somewhere to put large data files.
Having gone back to college for a few brush up classes in 2008 after being out for 10 yers, it was a bit of culture shock to me. The computers are way too much of a distraction and they should be banned for most of the class, and at least during the lecture session. I was one of the laptop users and I wouldn't have been angry if the teacher told me to stop, and I know a lot of people like that.
Fast typing has very little importance to developers. However, anyone who spends the kind of time with computers as programmers will likely have at least moderately fast typing skills. So when we see someone hunt and peck, we generally presume that this person is not well versed in computers much less a programmer and it's probably true more than 99% of the time. In this day and age, kids are exposed to typing early so it's very unlikely they aren't familiar with the qwerty keyboard. But it's quite possible that they have sloppy technique where they make a lot of errors and have to use the back space a lot or they don't use all their fingers, especially the pinky.
The best thing to do when a person is starting out is to spend 2 weeks, 1 hour a day developing fundamental typing skills using a training program like Mavis. Sure it's a chore but the amount of pain you save in the long run is amazing.
When the phone reboots a few times a day, boot time matters. Yes of course, the phone (android) shouldn't crash but it does.
OSes on the other hand do need to be boot more, especially laptops if you don't want to lose charge in standby or you care about crypto which only works if you power off. In standby mode, people can freeze the memory and transplant it to another computer.
The other problem is that OSes that boot slow also tend to run slow. The user interface slowness is something developers and product managers neglect all too often. That has worked for traditional computing products, but the public clearly likes a fast feeling OS. That's why the iPad does so well.
If you can make and algorithm 43000 times faster, that means the first version was utter garbage to begin with. By this standard, all a person needs to do is write an algorithm that is one million times less efficient than the hardware would permit. Then over 18 months, make the software progressively faster until it gets one million times faster and claim that you performed a miracle. I think this is the Montgomery "Scotty" Scott school of computer programing where you under promise a thousand fold but deliver a half ass effort to claim that you're a genius
The ugly reality is that mainstream applications and OSes have slowed down more than the hardware has improved. That's disturbing considering the fact that hardware these days is about 10,000 times faster yet modern OSes are more sluggish than ever. Fortunately, people are waking up to this fact from instant computing devices like smartphones and tablets and once they've used instant computing, waiting 5 minutes for a computer to boot or even 45 seconds on a lean fresh install just won't be acceptable.
The claim about Scott Lowe (which one never specified) was that he was on the FBI's dole to write how to implement OpenBSD based VPN VMware tutorials. Writing tutorials doesn't make him an OpenBSD coder. The claim was that "Jason Wright and others" were the ones who inserted the backdoor into the source code of OpenBSD. I haven't heard any refutation from Jason Wright and the story doesn't even claim that.
I mean individual companies. An ISP should selling to multiple companies is different. Of course there would still be exceptions of an individual company needed more than 256 per site. My point is that most of these legacy companies are sitting on blocks of 65K or 16.8M addresses and they're using it as an internal IP scheme. That's the problem I have.
It's ridiculous that companies own more than a/24 (256 IP addresses) since they're not using it for public visibility. Even a web site can mask thousands of servers behind a single IP. If they have multiple sites, let them have a/24 per site. This business of letting companies use multiple/16 or/8 (that's 16.8 million IPs) for private networks is ridiculous.
People who say "just switch to IPv6" simply don't get it. The reason is that even after you "switch", you really haven't switched at all because you still have to have backward compatibility. So what's the point of adding IPv6 when you still have to fully support IPv4 to reach the remaining 99.9% of the Internet? At the end of the day, we're still going to run IPv4 for the next decade or more which is why nobody in the real world cares about IPv6. If it was truly a switch over, then fine. But nobody is fooled into believing it's anything but some type of dual-stack for the foreseeable future.
We would all be better off if we made more efficient use of IPv4. We'll have to do that anyway even if we do switch to IPv6 (which won't happen) because of the need for backward compatibility.
Make sure you know the proper form. If you don't know it, most gyms offer at least 1 free training session and you can learn the correct form there. Also, do the full stroke. These guys that slap 1000 lbs on an incline sled and raise it 2 inches are fooling themselves. It boosts their ego, but doesn't do much for their muscle development. 10 good form full range squats with 135 lb free weight barbells beats those machines any day of the week.
Slashdot Mirror
Georgia gives new meaning to the word "back ho"
I explained the dispute between Level3/Netflix and Comcast in better detail here and it's not as simple as you made it out to be. http://www.digitalsociety.org/2010/12/video-level-3-versus-comcast-peering-dispute/ http://www.digitalsociety.org/2010/12/division-of-labor-between-broadband-and-cdn/
The real fight, not the cosmetic fight over ISPs censoring content which they can't do anyways, is over the government setting peering and interconnect prices even though this has always been set by the free market. In this case, the hardline Net Neutrality proponents want to set ISP peering rates to zero, or at least heavily regulated by the ISPs. The FCC tried to compromise by putting out incoherent regulations that would outlaw paid prioritization but not outlaw paid peering which are essentially the same thing (see http://www.digitalsociety.org/2010/12/fccs-utter-incoherence-on-paid-prioritization/). The FCC thought that compromise rules wouldn't get them sued by the ISPs and slammed by most of congress, but that happened anyways.
What happens when someone resolves an IPv6 address and their software and/or IP configuration won't support it? The point is that websites have to be 100% backwards compatible with IPv4 but an IPv6 presence will be optional. The point is that everything/everyone will have to maintain IPv4 compatibility which means there is simply no incentive to go dual-stack.
No, IPv4 will become valuable real-estate. Ask anyone if they prefer a 10 digit phone number or a 40 digit phone number because all the 10 digits ran out. Or a short domain name over a long domain name. The IPv4 addresses will become the original short IPs that all the large companies have already horded. Websites will prefer to be reachable over IPv4 and IPv6 visibility will be secondary. That's because IPv4 is accessible by the whole world one way or another and IPv6 will only be visible to a few early adopters.
The people complaining about not enough young people are the same ones who didn't produce enough offspring. They're having a hard time finding mates due to their unrealistic expectations. They might want to relax the rules for more immigrant spouses from other countries. Yes that dilutes their Japanese gene pool, but it's better than going to zero at the rate they're headed.
Even if you switch to a pubic IPv6 address, all your internal stuff will still be IPv4. My home print server and IP telephony adapter are all IPv4. The problem with IPv6 is that you can't entirely switch to it and just shut down IPv4. You have to run dualstack for the foreseeable future. That's why every IT consultant and IT manager and CIO I've spoken to says they don't give a crap about IPv6 because every adopter of IPv6 will have to be backward compatible with IPv4 so why bother running dual stack. Even after all the addresses are assigned, not a single IPv4 device or network will stop working.
The choice is between IPv4 single-stack or IPv4/IPv6 dual-stack. Given those as the only choices, people are choosing the former instead of the latter. There is no possibility of running IPv6 single-stack. IPv6 will essentially become the new "private IP addresses" that have to translate to "public" IPv4 addresses used by 99% of the IP devices in the world. The only difference is that IPv6 devices will be able to talk to each other without a NAT across organizations.
Your math works out if everyone lived in an apartment complex where your last mile costs are covered by the rent. In that situation, the "last mile" is really the last 100 meters which can be covered with CAT-5 cabling. But even when we look at your math, not a lot of people are interested in paying $389/month for broadband. The typical subscriber is willing to pay $40 and 1/3 of homes don't want broadband at all.
The cost of providing broadband for single family homes is substantially different, and it is in the last mile. AT&T and Verizon for example each have nearly 300,000 employees. When you're providing a single high speed circuit, it's substantially cheaper than providing 20-40 individual circuits to single unit homes. Each home costs about $10K to wire and you have to maintain each of those lines, and provide customer service to ~30 separate user accounts.
No, I've posted one of the more thoughtful posts here. http://www.digitalsociety.org/2009/09/the-need-for-a-broadband-transparency-standard/. It's not a straightforward answer.
"why can so many service providers in northern Europe and southeast Asia provide an extremely consistent 100+ mbps, even at night when virtually everyone is online (say in South Korea), to every single household for anywhere from $10-$50/month?"
The actual data shows US providers more honest. US promise index was 93%. EU promise index was 84.3%.
http://www.digitalsociety.org/2010/10/ookla-data-debunks-fcc-report-us-isps-exonerated/
They don't advertise at least x Mbps, they advertise "up to" 6 Mbps for example. I got my mom a 6 Mbps U-verse connection and found that their advertising wasn't accurate. Turned out that they gave her 7 Mbps which is generally sustainable even over a long duration. However, I don't expect 7 or 6 Mbps to be an "at least" number.
So you're saying that "up to" means "at least"? Do you not realize that broadband bits cost 20-40 times less than commercial bandwidth, precisely because it's shared 20-40 times? Now you want the government to change the service level of a shared circuit to that of a dedicated circuit? Any idea what this does to prices? Any idea how you'd actually achieve this, since it's impossible to build a core network that can handle all the concurrent data that the end points can throw at it?
These priority lanes have existed since the beginning of the broadband era.
http://www.digitalsociety.org/2010/12/video-level-3-versus-comcast-peering-dispute/
http://www.digitalsociety.org/2010/12/division-of-labor-between-broadband-and-cdn/
I hate having to choose between an SSD and an HDD for a laptop and really want one of each. I want a nice big 500+ GB HDD but I always want a 40+ GB SSD for a boot/OS/applications/page partition i.e., the "C" drive. Then you really get the best of both worlds because you get the insanely fast IO speeds of SSD but you have somewhere to put large data files.
Having gone back to college for a few brush up classes in 2008 after being out for 10 yers, it was a bit of culture shock to me. The computers are way too much of a distraction and they should be banned for most of the class, and at least during the lecture session. I was one of the laptop users and I wouldn't have been angry if the teacher told me to stop, and I know a lot of people like that.
Fast typing has very little importance to developers. However, anyone who spends the kind of time with computers as programmers will likely have at least moderately fast typing skills. So when we see someone hunt and peck, we generally presume that this person is not well versed in computers much less a programmer and it's probably true more than 99% of the time. In this day and age, kids are exposed to typing early so it's very unlikely they aren't familiar with the qwerty keyboard. But it's quite possible that they have sloppy technique where they make a lot of errors and have to use the back space a lot or they don't use all their fingers, especially the pinky.
The best thing to do when a person is starting out is to spend 2 weeks, 1 hour a day developing fundamental typing skills using a training program like Mavis. Sure it's a chore but the amount of pain you save in the long run is amazing.
When the phone reboots a few times a day, boot time matters. Yes of course, the phone (android) shouldn't crash but it does.
OSes on the other hand do need to be boot more, especially laptops if you don't want to lose charge in standby or you care about crypto which only works if you power off. In standby mode, people can freeze the memory and transplant it to another computer.
The other problem is that OSes that boot slow also tend to run slow. The user interface slowness is something developers and product managers neglect all too often. That has worked for traditional computing products, but the public clearly likes a fast feeling OS. That's why the iPad does so well.
If you can make and algorithm 43000 times faster, that means the first version was utter garbage to begin with. By this standard, all a person needs to do is write an algorithm that is one million times less efficient than the hardware would permit. Then over 18 months, make the software progressively faster until it gets one million times faster and claim that you performed a miracle. I think this is the Montgomery "Scotty" Scott school of computer programing where you under promise a thousand fold but deliver a half ass effort to claim that you're a genius
The ugly reality is that mainstream applications and OSes have slowed down more than the hardware has improved. That's disturbing considering the fact that hardware these days is about 10,000 times faster yet modern OSes are more sluggish than ever. Fortunately, people are waking up to this fact from instant computing devices like smartphones and tablets and once they've used instant computing, waiting 5 minutes for a computer to boot or even 45 seconds on a lean fresh install just won't be acceptable.
Better title is "NSA not delusional of impenetrable networks".
Thank you for link.
The claim about Scott Lowe (which one never specified) was that he was on the FBI's dole to write how to implement OpenBSD based VPN VMware tutorials. Writing tutorials doesn't make him an OpenBSD coder. The claim was that "Jason Wright and others" were the ones who inserted the backdoor into the source code of OpenBSD. I haven't heard any refutation from Jason Wright and the story doesn't even claim that.
I mean individual companies. An ISP should selling to multiple companies is different. Of course there would still be exceptions of an individual company needed more than 256 per site. My point is that most of these legacy companies are sitting on blocks of 65K or 16.8M addresses and they're using it as an internal IP scheme. That's the problem I have.
That is an interesting question. I don't know.
It's ridiculous that companies own more than a /24 (256 IP addresses) since they're not using it for public visibility. Even a web site can mask thousands of servers behind a single IP. If they have multiple sites, let them have a /24 per site. This business of letting companies use multiple /16 or /8 (that's 16.8 million IPs) for private networks is ridiculous.
People who say "just switch to IPv6" simply don't get it. The reason is that even after you "switch", you really haven't switched at all because you still have to have backward compatibility. So what's the point of adding IPv6 when you still have to fully support IPv4 to reach the remaining 99.9% of the Internet? At the end of the day, we're still going to run IPv4 for the next decade or more which is why nobody in the real world cares about IPv6. If it was truly a switch over, then fine. But nobody is fooled into believing it's anything but some type of dual-stack for the foreseeable future.
We would all be better off if we made more efficient use of IPv4. We'll have to do that anyway even if we do switch to IPv6 (which won't happen) because of the need for backward compatibility.
Make sure you know the proper form. If you don't know it, most gyms offer at least 1 free training session and you can learn the correct form there. Also, do the full stroke. These guys that slap 1000 lbs on an incline sled and raise it 2 inches are fooling themselves. It boosts their ego, but doesn't do much for their muscle development. 10 good form full range squats with 135 lb free weight barbells beats those machines any day of the week.