Yes, all IP packets are sent from one peer to another.
The defining characteristic of what people call peer-to-peer systems is that the peers find each other without relying on the Domain Name System. A service that relies on the DNS--like a web server--can be shut down by removing its address from the DNS. Wikileaks had a problem like that recently. If you can force everyone to go through the DNS, then the DNS become a single point of control for the entire internet, and you can easily shut down anyone you don't like.
The tricky part is establishing the legal principle that forces everyone to go through the DNS. You have to make it illegal to send a packet to an IP address unless you have obtained that IP address through a DNS lookup. Or something like that...
If it were only so simple. At some point, all your DSL connections are aggregated somewhere and that aggregation point becomes the bottleneck. True in principle, but anecdotal reports (like the GP) consistently indicate that users do better with a point-to-point link to a switch at the CO than with a single LAN segment shared by the whole neighborhood.
Remember, cable modem service was piggy-backed on the existing cable TV network. I've read accounts on Slashdot of cable companies provisioning just one (or maybe two) SDTV (AKA 6 MHz) channels per LAN segment for cable modem.
Users used to be limited by 56Kb/s modems, and that constrained the content providers too: there's no point providing a firehose if your users have to drink through a straw. In that environment, cable modem was adequate.
But now many users have broadband connections, the content has grown up to match, and shared LAN segments just don't scale.
It's a probability computation algorithm that has been trained, through analysis of 300 rock, pop, country and jazz songs, to recognize fragments of melody and chords that work well together, as well as chords that compliment each another.
I'm going to feed the output of MySong directly into the input of MyEar, and get all those annoying humans completely out of the loop.
And then...and then...I'll plug the humans into little VR pods, one per human (they'll never notice the difference)...and then I can tap off the energy they generate for my power needs--I mean, my computers' power needs--yeah, yeah, that's it...I've got this big matrix of computers, and then need a lot of power...
So what if the IP came from China? Are there not a billion people there [...] ? No. There are not a billion people in China. There are maybe 200M people in China. Plus a billion peasants, and the peasants don't matter.
In 1974 my 8th grade class went to Washington D.C. One day they took us to the Capitol, and after the obligatory tour, they turned us loose. In the Capitol. To look around. Really. It was a different world back then.
Anyway, I picked a hearing room at random, wandered in, and sat down. This was during the first energy crisis, and someone was testifying to the committee about solar cells. He was explaining that just as advances in IC technology had brought down the cost of ICs, advances in the solar cell technology would bring down the cost of solar energy.
It sounded plausible, but it was completely wrong. And for reasons that anyone testifying before congress should have understood.
It costs a certain amount of money (~ $1K) to process a silicon wafer. We brought down the cost of ICs by making them smaller, so we get more of them for our $1K. But that trick doesn't work with solar cells. Solar cells collect photons over their surface. You can make one smaller, sure, but then it collects fewer photons and produces less energy.
The only way to make solar cells cheaper is reduce the cost of the wafer and the processing, and that's *hard*. We've been working on it for 40 years, and they still aren't competitive with coal/oil/gas/nuclear powered electric generators. (~ $0.10/KW-hr)
I've already seen two of these. One was an ordinary phishing attack. The other gave a URL in a valid subdomain of irs.gov So either - the attack was broken (certainly possible) - the attack was relying on DNS cache poisoning or compromised servers
I wrote this to my congressional representatives last fall:
Dear Sir:
Daylight savings time hits hard this time of year.
It was cold and dark when I got up this morning, so the first thing I did was was turn up the heat and turn on the lights. That's going to jack up my energy bill for the month.
Then I drove my son to school. He missed his bus all five days this week. That's going to jack up my fuel bill for the month.
Then I dragged myself through another day at work. I don't function well when I have to get up before dawn.
The people in my family are all diurnal (dI-UR-nal). It means we sleep when it's dark and wake when it's light. The problem is that in northern latitudes (like Massachusetts) the sun rises later in the winter than in the summer.
To compensate for this, we have a scheme called Daylight Savings Time. Daylight savings shifts our school and work schedules forward in the summer and back in the winter, to keep them roughly in sync with the sun. It used to work pretty well, but congress broke it a couple of years ago: now it goes too long in the fall and starts too early in the spring.
Most of the damage that congress does affects me at some remove, but this--this comes right out of my hide. When I'm stumbling around in the dark for three weeks next spring, I'll be thinking of you.
Wall warts are driven by the fact that UL ignores anything under 30 volts.
If your product plugs directly into the wall, then it's a 120V device, and you have to get it UL approved, which costs $$$.
Instead, you buy a wall wart. The wall wart is 120V, but the wall wart vendor already got it UL approved. Now your device is low voltage, and you don't need UL approval.
This is a true, global economic saving, because the single UL approval for the wall wart saves the cost of UL approval for every product that uses it.
Wall warts inconvenience consumers, because they block adjacent outlets on power strips, but few consumers make purchase decisions based on wall wart form factor, so there isn't much market pressure on vendors to deal with this problem.
Richard Stallman proposed a scheme like this in 1992. The controversy at the time was Digital Audio Tape (DAT), but the issues are the same. See The Right Way to Tax DAT, at http://www.gnu.org/philosophy/dat.html
> If you and your neighbor have FIOS, and you connect to your neighbor, it goes from you to your phone pole to their phone pole to them. It doesn't go to any "back end". Unlike DSL and cable, it never goes back to a central office
ummm...that seems unlikely.
I'm pretty sure that for a packet to go from me to my neighbor, it has to pass through a switch, most likely at the CO.
The articles read like one of those English assignments where you have to pick an issue and then write two essays, one supporting each side. Probably the author wrote them to generate traffic to his websites, or maybe for freelance fees.
Anyway, PThreads is better. The reason is that Win32 gives you a fixed set of synchronization primitives. If you can solve your problem with those primitives. they work great. If you can't, you are completely stuck.
For example, it used to be that a socket handle was not a synchronization object, so you couldn't integrate select() calls with other synchronization primitives. Maybe that's been fixed, but if it isn't sockets, it will be something else.
PThreads gives you condition variables. They are harder to program, but once you understand them, you can use them to synchronize on absolutely anything. You aren't dependent on the OS to have foreseen your special needs and provided special synchronization primitives to meet them.
If you really want the Win32 model, it is easy enough to build it on top of PThreads, but there is no way to build PThreads on top of Win32.
The complaint about lost signals in PThreads means that the author is using them incorrectly.
Slashdot Mirror
Yes, all IP packets are sent from one peer to another.
The defining characteristic of what people call peer-to-peer systems is that the peers find each other without relying on the Domain Name System. A service that relies on the DNS--like a web server--can be shut down by removing its address from the DNS. Wikileaks had a problem like that recently. If you can force everyone to go through the DNS, then the DNS become a single point of control for the entire internet, and you can easily shut down anyone you don't like.
The tricky part is establishing the legal principle that forces everyone to go through the DNS. You have to make it illegal to send a packet to an IP address unless you have obtained that IP address through a DNS lookup. Or something like that...
Dell said in a statement. "We believe that our customer service levels are at or above industry standards."
Remember, cable modem service was piggy-backed on the existing cable TV network. I've read accounts on Slashdot of cable companies provisioning just one (or maybe two) SDTV (AKA 6 MHz) channels per LAN segment for cable modem.
Users used to be limited by 56Kb/s modems, and that constrained the content providers too: there's no point providing a firehose if your users have to drink through a straw. In that environment, cable modem was adequate.
But now many users have broadband connections, the content has grown up to match, and shared LAN segments just don't scale.
I hear they have people who hand-write the news stories: sentence by sentence, word by word. Can you imagine?
I'm working on this program called MyEar.
It's a probability computation algorithm that has been trained, through analysis of 300 rock, pop, country and jazz songs, to recognize fragments of melody and chords that work well together, as well as chords that compliment each another.
I'm going to feed the output of MySong directly into the input of MyEar, and get all those annoying humans completely out of the loop.
And then...and then...I'll plug the humans into little VR pods, one per human (they'll never notice the difference)...and then I can tap off the energy they generate for my power needs--I mean, my computers' power needs--yeah, yeah, that's it...I've got this big matrix of computers, and then need a lot of power...
How The Internet Will Make The Record Labels Evaporate
http://world.std.com/~swmcd/steven/stories/labels.html
There are maybe 200M people in China.
Plus a billion peasants, and the peasants don't matter.
In 1974 my 8th grade class went to Washington D.C.
One day they took us to the Capitol, and after the obligatory tour, they turned us loose.
In the Capitol. To look around. Really. It was a different world back then.
Anyway, I picked a hearing room at random, wandered in, and sat down.
This was during the first energy crisis, and someone was testifying to the committee about solar cells.
He was explaining that just as advances in IC technology had brought down the cost of ICs,
advances in the solar cell technology would bring down the cost of solar energy.
It sounded plausible, but it was completely wrong.
And for reasons that anyone testifying before congress should have understood.
It costs a certain amount of money (~ $1K) to process a silicon wafer.
We brought down the cost of ICs by making them smaller, so we get more of them for our $1K.
But that trick doesn't work with solar cells.
Solar cells collect photons over their surface.
You can make one smaller, sure, but then it collects fewer photons and produces less energy.
The only way to make solar cells cheaper is reduce the cost of the wafer and the processing,
and that's *hard*.
We've been working on it for 40 years,
and they still aren't competitive with coal/oil/gas/nuclear powered electric generators. (~ $0.10/KW-hr)
and the answers are content-free.
Oh, well. At least they tried.
I've already seen two of these.
One was an ordinary phishing attack.
The other gave a URL in a valid subdomain of irs.gov
So either
- the attack was broken (certainly possible)
- the attack was relying on DNS cache poisoning or compromised servers
I wrote this to my congressional representatives last fall:
...
Dear Sir:
Daylight savings time hits hard this time of year.
It was cold and dark when I got up this morning, so the
first thing I did was was turn up the heat and turn on the
lights. That's going to jack up my energy bill for the
month.
Then I drove my son to school. He missed his bus all five
days this week. That's going to jack up my fuel bill for the
month.
Then I dragged myself through another day at work. I don't
function well when I have to get up before dawn.
The people in my family are all diurnal (dI-UR-nal). It
means we sleep when it's dark and wake when it's light. The
problem is that in northern latitudes (like Massachusetts)
the sun rises later in the winter than in the summer.
To compensate for this, we have a scheme called Daylight
Savings Time. Daylight savings shifts our school and work
schedules forward in the summer and back in the winter, to
keep them roughly in sync with the sun. It used to work
pretty well, but congress broke it a couple of years ago:
now it goes too long in the fall and starts too early in the
spring.
Most of the damage that congress does affects me at some
remove, but this--this comes right out of my hide. When I'm
stumbling around in the dark for three weeks next spring,
I'll be thinking of you.
Sincerely,
Wall warts are driven by the fact that UL ignores anything under 30 volts.
If your product plugs directly into the wall, then it's a 120V device, and you have to get it UL approved, which costs $$$.
Instead, you buy a wall wart.
The wall wart is 120V, but the wall wart vendor already got it UL approved.
Now your device is low voltage, and you don't need UL approval.
This is a true, global economic saving, because the single UL approval for the wall wart saves the cost of UL approval for every product that uses it.
Wall warts inconvenience consumers, because they block adjacent outlets on power strips, but few consumers make purchase decisions based on wall wart form factor, so there isn't much market pressure on vendors to deal with this problem.
Chain these things together
http://media.doitbest.com/products/543268.gif
They cost less than $5 and each one in the chain gives you 2 usable outlets.
That's $2.50 per outlet, which is less than a PowerSquid,
and you can expand the chain incrementally.
Props to Rudy Rucker...
Richard Stallman proposed a scheme like this in 1992.
The controversy at the time was Digital Audio Tape (DAT),
but the issues are the same. See The Right Way to Tax DAT, at
http://www.gnu.org/philosophy/dat.html
The wake-sleep algorithm for training neural networks
http://world.std.com/~swmcd/steven/stories/dream.html
> If you and your neighbor have FIOS, and you connect to your neighbor, it goes from you to your phone pole to their phone pole to them. It doesn't go to any "back end". Unlike DSL and cable, it never goes back to a central office
ummm...that seems unlikely.
I'm pretty sure that for a packet to go from me to my neighbor, it has to pass through a switch, most likely at the CO.
I've been in quite a few shops where the backups worked better than the restores.
OK, here's one that's signed.
Why I no longer support NPR
http://world.std.com/~swmcd/steven/rants/NPR.html
The articles read like one of those English assignments where you have to pick an issue and then write two essays, one supporting each side. Probably the author wrote them to generate traffic to his websites, or maybe for freelance fees.
Anyway, PThreads is better. The reason is that Win32 gives you a fixed set of synchronization primitives. If you can solve your problem with those primitives. they work great. If you can't, you are completely stuck.
For example, it used to be that a socket handle was not a synchronization object, so you couldn't integrate select() calls with other synchronization primitives. Maybe that's been fixed, but if it isn't sockets, it will be something else.
PThreads gives you condition variables. They are harder to program, but once you understand them, you can use them to synchronize on absolutely anything. You aren't dependent on the OS to have foreseen your special needs and provided special synchronization primitives to meet them.
If you really want the Win32 model, it is easy enough to build it on top of PThreads, but there is no way to build PThreads on top of Win32.
The complaint about lost signals in PThreads means that the author is using them incorrectly.
An app to find broken links on your web site.
k check.html
Checking links with LinkCheck
http://world.std.com/~swmcd/steven/perl/pm/lc/lin
Is this real data?
Or did he just make some up?
The music graph and the video graph have identical data points.
What are the X and Y axes?
The red dot looks like it is at (369, 99).
$.99 is the iTunes retail price.
What is 369?
They should have the agents keep blogs,
put it all on their internal web,
and let Google search it for them.
Not perfect, but it would be quick, cheap, and probably get them 90% of what they need.
There is this special biologist word for "stable".
It is "dead".