And things like P2P networks would be cut way back as it's more difficult to structure them with thousands of NAT'd nodes.
But this is the POINT. Using a NAT extensively BREAKS THE INTERNET. The internet is a network that connects end devices to each other. Insterting a NAT in the network removes or limits the ability of each device to communicate directly to any other device on the network. And it's that principle (the End to End principle) that has built the internet into what it is today. With a NAT, you have to get permission from the person who runs the NAT to innovate, to start a new service, or invent new methods of communication. Which is restrictive, and would seriously hamper the ability of users to innovate.
His documents are full of "my servers...", "me" this and "I" that.
Well what's he supposed to talk about? "other guys's servers", "they" and "him"? Of course he's going to talk about his own actions, his own things, and how he was involved in this. That's the only perspective he can speak with any authority about. Franky, i though he did a good job of telling us his opinions, and letting us, the readers, draw our own conclusions about how the other members in this situation acted.
Criticizing someone for speading lies about others would be fair. Criticizing the guy for giving us his own opinions bizarre.
What else is a monopoly but the exclusive control over the rights to something?
No no no. That's not a monopoly at all. A monopoly is a market condition. A monopoly is when a single corporation offers a commodity or service and there IS NO VIABLE ALTERNATIVE. They are the sole provider of that good.
For instance, when the local power company is the ONLY company who can reasonably provide you with power. Or IBM is the only software company who can reasonably provide you with a gumba-browser (cause they got a patent on the whole idear, and now they have legally enforced monopoly).
If you have complete control over a creative work, or a painting, or something that does not have an alternative (similar music style, but different song doesn't cut it, does it?) that's not a monopoly. That's copyright. There is only one of it, so there cannot be competition.
For example, contract, which, in your words, bestows "exclusive control over the rights to something" does not create a monopoly. That's only a contract, over a specific item or work.
Most of your post was well intentioned, but please don't use words like 'monopoly' if you don't know what they mean. monopoly != control. It's much more specific than that. Monopoly is exclusive control by one group of the means of producing or selling a commodity or service.
Monopoly doesn't even imply bad or evil, either. But i'll not get off on that tangent.
Re:short term - new clients are too configurable
on
BitTorrent Guide
·
· Score: 1
Thanks for the link man. I've not yet accomplished anything, but this site contains the info i need to so all kinds of cool stuff with network shaping in linux. Man, i new reading/. comments would pay off eventually!
You're amazing. Truly amazing. Your tenacity for not giving up your ideas is phenominal.
Cutting off legitimate mail servers because they lack PTR records is not an acceptable solution...
Tell me why a legitimate mail server can't have a PTR record? There's no reason why someone running a legitimate mail server can't have the PTR record set up correctly. And that's the reason why it's so strange when they don't set it up properly. Why would someone act like a spammer (fail to set up PTR records) when they're not?
A) They're lazy B) they don't know about PTR records, which means they probably don't know enough to run a mail server C) They're thickheaded like you and think that when their friends (other friendly mail server admins) ask them nicely to behave civilly (set up PTR records) they're being forced to do something against their will, and they won't have any of that!
You seem to fall into the C) category. And this is your problem: Even though we're giving you good advice, you're ignoring it cause it's not your own.
I sincerely hope you don't run a mail server of any consequence.
I really have to question this article, and it's sources. It claims this is some kind of record, but i'm certain it's not. Perhaps it's just a bit of good ol' American Pride getting in the way here?
Guys, it's 932 Megabits per second. Not even a gigabit. For gosh sakes, you can get 10 Gigabit (that's 10,000 megabit) connections these days, if you buy really expensive high end gear. However, having a single data stream doing that fast is indeed impressive. But it's been done before.
This article outlines how YottaYotta (a storage company) along with 7 key partners (CANARIE, WestGrid, BCNET, StarLight, Netera Alliance, The Logistical Computing and Internetworking Laboratory at the University of Tennessee and The Physics Department at Carleton University.) were able to transfer data at 11.1 Gbps (that's 11,100 megabits) over 10,000 km, reading and writing on disks at either end. They used straight TCP/IP, and the main limitation was the 12 Gbps limit of their WAN link. Really they did, check the article.
So 11 times faster than this 'record', and using actual disks at either end, and standard TCP/IP. I'd say this record wins.
engage sarcasm Thanks slashdot, for another wonderfully accurate, informative story, of which we should not question, just believe blindly. disengage sarcasm
Let this be a lesson to the crack head mods to really think...
The sad think dildatron, is that the mods are us. We are the mods. The mods are just regular/. readers like you and me, randomly chosen to moderate. I've been a mod. You've probably been a mod.
What i'm saying is when you tell me that the mods are stupid, you're really saying that readers of/. are stupid. Which might very well be true.
With more pitful disregard for anything remotely resembling quality journalism, it seems slashdot editors don't even click on links to see if they work now. CORRECT SHREDDER LINK
Hemos, working as he usually does, repsonds to CmdrTaco's question about the link working. "Of course it works. Why wouldn't it work? No one would send us a broken link. That's impossible. Also, since i'm telepathic now, I don't even need to read the story to know it's front page material. Damn I'm good."
Hemos cliks *post*, grinning smugly, "Congratulations, Hemos, another feather in your cap. Job well done."
Uhm... are you SURE? You sure it's not the new 2.5 development kernel you're talking about? Cause i think if you looked into it, you'd realize that all these massive changes you mentioned, improving threading, I/O, VM and the rest are changes happening in the 2.5 kernel. They are major changes, and very significant. But they're happening in 2.5.
This release is just an incremental release. Bugfixes here, added support there... no major changes, just evolutionary changes.
Good to see the mods are looking for informative comments... however, it's sad when they mod up comments that don't reflect the facts.
I chased that link and while learning about the crash, i've yet to see anything about a rubber wheel. all ive' seen so far is: A broken wheel on the first passenger car has been fingered as the direct cause of the accident.
i'm wondering if you could point to something more specific?
I've got red hair, bright red. I've got freckles too, though fewer now than when i was a kid. I'm 20 now. Anyways, the only thing that i've noticed about having red hair is that you get sunburnt easily. Other than that, nothing. In fact, i think i have a higher than average tolerance for pain. It's hard to compare levels of pain with others, of course, but I don't find myself hurting more than others at all.
The article states that, in general, redheads felt more pain. But what if i don't? What if i feel less pain than average? Directly linking hair-color to pain reception is tenuous at best, and it seems like doctors messing with my anestheisa beacuase of my hair color could be bad, with the threat of coma.
I've never heard of hair color affecting ANYTHING before. Sure race affects some things, like reactions to certain drugs, and diseases. Sex and age as well will determine the best treatment. But hair color? What about eye color, does that factor in?
It seems like this is a broad generaliztion, and i don't think this can be the deciding criteria for pain reception. I don't buy it this first time round, if only based on personal experience.
yeah, actually, i did read it. Slashdot has that handy feature where it tells you when someone has replied to you.
Unfortunately, I am visiting my parents for [Canadian] thanksgiving, and i don't have access to Linux atm. anyways, it was interesting to read how you think it should work. I've never seen, well, probably just not noticed that behavior before, but i'll check for it. To me, it seems like quite a minor issue, as how many people scroll down page by page to the bottom of long documents, without letting up on the mouse, hence making the scrolling extremely fast. I guess if you do though, that's good enough for it to be an issue for you.
And yes, AFAIK, this is the typical windows behavior, the way it behaves now. Also, have you tried submitting a bug about this scrolling stuff? Do you even know if the devs realize it's broken? Because if you speak up, they may just fix it.
The scrollbars don't work properly.... [explanation provided]
To be breif; No, you're wrong. I'm running Mozilla 1.0 on Linux right now. Gentoo Linux, to be specific. I honestly don't know what you're saying is broken. I tried the middle button. No matter what i do, the slider always jumps to the cursor's position, and dragging from there moves it. Using the left mouse button above and below the slider scrolls up or down a page, and holding the button down scrolls the page, screen by screen, until the slider reaches the cursor, or the bottom of the page, whichever comes first.
Either you're basing your views on an old version of Mozilla, or your system is bahiving differently from mine, for whatever reason. There's another possibility, too, but i'll refrain from mentioning it.
BK is not free software. If they do somehow end up in court for this, violating anti-trust laws, you know what Bit Mover will do? They will say, fine, if this license is illegal, there is none anymore. No free license at all. And then BK will not be free as in beer, not at all. You will not be able to get BK in a free form, you will only be able to get the pay version.
And will tha be a victory? Will righteousness and truth have triumphed? Maybe, but it won't help me out one bit. We'll be worse of than we are right now. But you can hold your own opinion.
This does not work for me. I'm using Win2k with Mozilla 1.1. I also tried IE. Then i tried Lynx on a linux box. Then another linux box, on an entirely different network, and ISP. They all tell me that the host cannot be located. Which is funny, cause that's exactly what i expected to happen, since this isn't a valid IP, nor a valid DNS name. Oh, i also just asked my buddy if he could see it an no, he can't.
A) The license forbids you to use BK to further a direct competitor to BK. Distributing a competitor, while using BK, like Red Hat does, is allowed.
B) This license is the FREE license. Remember the saying, "Beggars can't be choosers?" They can't. Are you using BK for free? Then you can't expect to choose the license. If you buy the program, you can develop whatever you like with it.
C) Anyone still has the ability to be a kernel hacker without using BK whatsoever. The old tools still work, Linus and everyone else still accepts standard patches. It's just the old tools are actually worse than BK. BK was chosen purely on technical merits, it's only the license that's raising questions.
Point B) is important. Because this is the FREE license, it means that BM is not violating anti-trust laws by forbidding competition, because you can purchase the product, and get unrestricted use. Companies are not required to provide free samples of their products to competitors to help them out. Also, it means that BM is NOT acting like MS when they pulled the same stunt in their EULA. (Adding a clause stating that you cannot use MS products to harm MS in any way).
Summary: Bit Mover is acting reasonably, and completely within their rights as a company to define the acceptable uses of their free gift to users. The issue should is not whether or not Bit Mover is 'cheating' people. The issue now is whether or not to use Bit Keeper personally.
Course, maybe this BitKeeper appeals to managers more than actual developers...
I'm not a developer, but you know who is? Linus Torvalds. And you know what else? I have a sneaking feeling that he's a better developer than you. And you know what he thinks? He thinks BitKeeper is better than CVS. And you know what? All the other kernel developers agree with him. The only thing they can't reach a consensus about is the liscense.
The isssue is not whether BK is better than CVS or not. BK is better. That's been decided. The issue is about the liscense.
Also, for numbers between the dots in IP addresses, the cap is NOT usually 255.
What? What does this mean? What are you trying to say? Your sig URL doesn't work, this statement doesn't make any sense, and you don't provide any backing to it at all.
First, let me state that by no means am i an expert on IPv4 or v6, but it think i can add something here.
Your idea about assigning groups of IP addresses to certain orginizations (like universities) actually HAS happened, but not quite in the organized way you described. For instance, for a while, most cable modems you saw had a 24.x.x.x IP address. The reason for that was that @Home owned the entire 24 subnet. I belive that other companies were assigned entire 'Class A' subnets as they were called. AT&T has 12.x.x.x, i believe, Genuity has 4.x.x.x. Universities seem to usually have a 'Class B' address. The university of Alberta (in canada, where i am) has all 129.128.x.x addresses. ISP's usually own blocks of IP addresses, which are usually assigned in the least amounts that are required. So if a small isp only needs 100 IP's, they'll give them something like 205.132.211.x, which gives them 254 usable addresses. The bigger the organization, the more IP's are assigned to them. Currently, this is done by IANA in North America, and RIPE in Europe.
The other point is in response to your statement "but numbers don't really matter, so why not just let people at them?"
Well, the answer is IP Addresses DO matter. Your IP address determines how traffic is routed to you. It would be nearly impossible to keep track of how to route traffic to different ip's if 123.123.123.1 was in bejing, and 123.123.123.2 was in amsterdam. Routing logic says these 2 ip's will be routed differently only at the last step. If each IP were different, you'd have MASSIVE routing tables which would have to keep track of how to get to each ip around the word, which would be further complicate with dynamic ips, etc.
The upshot of this is that routing is very hard, and you can't decide numbers on a whim, just like you can't choose your own phone number.
>I would plan on an amp per server at least, and go no more than 16 boxes per circuit.
Now, i'm not an electrician, but this is probably the worst bit of advice here. Don't get me wrong, electricity is important, but let's do the math, and determine how good your advice is.
Amps x Volts = Wattage.
Therefore, assuming you're using 110 Volt power, and your calculation of 1 amp per server, we get:
1 x 110 = 110 Watts, per sever.
It should be painfully obvious from the above that this is obviously way too little power to be running a server on. (to say nothing of your expertise with electricity and servers.)
Now, a 20 amp circuit gives you 2220 watts of power. ( 20A x 110v = 2200W ) Assuming 500 watts per server, you can run about 4 boxes per circuit. Remember, computers run on electricity, and if you rob them of power, they are entirely useless without it. Do not skimp on power.
In fact, if you want to do power right, you will run ONE circuit PER server. This way, if you need to cut power, or a breaker blows, only one server is affected, not the entire rack. And also, you don't have to worry about not having enough power for a big server, unless your server wants more than 2000 watts, which is less likely, unless you're getting some big iron servers. And those would likely have dual (or more) power supplies. And you would want them plugged into seperate circuits, because if one fails, the other is a backup.
The overall point is, it's hard to have too much power, and with some simple math, it's easy to figure out your needs, and what to set up. Compared to the cost of a server, a curcuit is pretty cheap to install, (i'm guessing $50 each) so there's no reason to cheat yourself here. If you want to do a server room right, you've got to have good power.
The irony here is that i don't even think the author knew it was a troll. It even sparked some good comments debating the demand for broadband access. Good work.
However, the demand ISN'T THE QUESTION! @Home had 4.1 MILLION subscribers. That's 45% of the broadband market in the US, and growing at up to 100,000 new subs per month. To question the demand is silly. There is no question as to whether or not the demand is there, it most obviously is. The question is how @Home managed to screw it up. If you have 4.1 million people paying you $20/month for internet service, how to do you end up bankrupt? That's the question.
So ignore the demand question, and concentrate on the mismanagement question.
Slashdot Mirror
And things like P2P networks would be cut way back as it's more difficult to structure them with thousands of NAT'd nodes.
But this is the POINT. Using a NAT extensively BREAKS THE INTERNET. The internet is a network that connects end devices to each other. Insterting a NAT in the network removes or limits the ability of each device to communicate directly to any other device on the network. And it's that principle (the End to End principle) that has built the internet into what it is today. With a NAT, you have to get permission from the person who runs the NAT to innovate, to start a new service, or invent new methods of communication. Which is restrictive, and would seriously hamper the ability of users to innovate.
Sorry, djrogers, but you're incorrect, and the origninal corrector is accurate. See RFC 1918, Section 3.
:)
Maybe it's not your day that's slow, but your brain! heh heh
His documents are full of "my servers...", "me" this and "I" that.
Well what's he supposed to talk about? "other guys's servers", "they" and "him"? Of course he's going to talk about his own actions, his own things, and how he was involved in this. That's the only perspective he can speak with any authority about. Franky, i though he did a good job of telling us his opinions, and letting us, the readers, draw our own conclusions about how the other members in this situation acted.
Criticizing someone for speading lies about others would be fair. Criticizing the guy for giving us his own opinions bizarre.
What else is a monopoly but the exclusive control over the rights to something?
No no no. That's not a monopoly at all. A monopoly is a market condition. A monopoly is when a single corporation offers a commodity or service and there IS NO VIABLE ALTERNATIVE. They are the sole provider of that good.
For instance, when the local power company is the ONLY company who can reasonably provide you with power. Or IBM is the only software company who can reasonably provide you with a gumba-browser (cause they got a patent on the whole idear, and now they have legally enforced monopoly).
If you have complete control over a creative work, or a painting, or something that does not have an alternative (similar music style, but different song doesn't cut it, does it?) that's not a monopoly. That's copyright. There is only one of it, so there cannot be competition.
For example, contract, which, in your words, bestows "exclusive control over the rights to something" does not create a monopoly. That's only a contract, over a specific item or work.
Most of your post was well intentioned, but please don't use words like 'monopoly' if you don't know what they mean. monopoly != control. It's much more specific than that. Monopoly is exclusive control by one group of the means of producing or selling a commodity or service.
Monopoly doesn't even imply bad or evil, either. But i'll not get off on that tangent.
Thanks for the link man. I've not yet accomplished anything, but this site contains the info i need to so all kinds of cool stuff with network shaping in linux. Man, i new reading /. comments would pay off eventually!
You're amazing. Truly amazing. Your tenacity for not giving up your ideas is phenominal.
Cutting off legitimate mail servers because they lack PTR records is not an acceptable solution...
Tell me why a legitimate mail server can't have a PTR record? There's no reason why someone running a legitimate mail server can't have the PTR record set up correctly. And that's the reason why it's so strange when they don't set it up properly. Why would someone act like a spammer (fail to set up PTR records) when they're not?
A) They're lazy
B) they don't know about PTR records, which means they probably don't know enough to run a mail server
C) They're thickheaded like you and think that when their friends (other friendly mail server admins) ask them nicely to behave civilly (set up PTR records) they're being forced to do something against their will, and they won't have any of that!
You seem to fall into the C) category. And this is your problem: Even though we're giving you good advice, you're ignoring it cause it's not your own.
I sincerely hope you don't run a mail server of any consequence.
I really have to question this article, and it's sources. It claims this is some kind of record, but i'm certain it's not. Perhaps it's just a bit of good ol' American Pride getting in the way here?
Guys, it's 932 Megabits per second. Not even a gigabit. For gosh sakes, you can get 10 Gigabit (that's 10,000 megabit) connections these days, if you buy really expensive high end gear. However, having a single data stream doing that fast is indeed impressive. But it's been done before.
This article outlines how YottaYotta (a storage company) along with 7 key partners (CANARIE, WestGrid, BCNET, StarLight, Netera Alliance, The Logistical Computing and Internetworking Laboratory at the University of Tennessee and The Physics Department at Carleton University.) were able to transfer data at 11.1 Gbps (that's 11,100 megabits) over 10,000 km, reading and writing on disks at either end. They used straight TCP/IP, and the main limitation was the 12 Gbps limit of their WAN link. Really they did, check the article.
So 11 times faster than this 'record', and using actual disks at either end, and standard TCP/IP. I'd say this record wins.
engage sarcasm Thanks slashdot, for another wonderfully accurate, informative story, of which we should not question, just believe blindly. disengage sarcasm
Let this be a lesson to the crack head mods to really think...
/. readers like you and me, randomly chosen to moderate. I've been a mod. You've probably been a mod.
/. are stupid. Which might very well be true.
The sad think dildatron, is that the mods are us. We are the mods. The mods are just regular
What i'm saying is when you tell me that the mods are stupid, you're really saying that readers of
With more pitful disregard for anything remotely resembling quality journalism, it seems slashdot editors don't even click on links to see if they work now.
CORRECT SHREDDER LINK
Hemos, working as he usually does, repsonds to CmdrTaco's question about the link working. "Of course it works. Why wouldn't it work? No one would send us a broken link. That's impossible. Also, since i'm telepathic now, I don't even need to read the story to know it's front page material. Damn I'm good."
Hemos cliks *post*, grinning smugly, "Congratulations, Hemos, another feather in your cap. Job well done."
Uhm... are you SURE? You sure it's not the new 2.5 development kernel you're talking about? Cause i think if you looked into it, you'd realize that all these massive changes you mentioned, improving threading, I/O, VM and the rest are changes happening in the 2.5 kernel. They are major changes, and very significant. But they're happening in 2.5.
This release is just an incremental release. Bugfixes here, added support there... no major changes, just evolutionary changes.
Good to see the mods are looking for informative comments... however, it's sad when they mod up comments that don't reflect the facts.
RDMA? I've never heard of this. Can you explain it, or at least give a link for more information?
Well, you guys can rest easy, i sent xant an email. I said:
9 265.
:)"
"Hey xant,
I've attached the critical file you alluded to in your comment at http://slashdot.org/comments.pl?sid=43025&cid=450
Keep it on your hard drive in case we all need it.
Heh. In case his hard drive goes, maybe a couple other people should get it from here.
I chased that link and while learning about the crash, i've yet to see anything about a rubber wheel. all ive' seen so far is: A broken wheel on the first passenger car has been fingered as the direct cause of the accident.
i'm wondering if you could point to something more specific?
I've got red hair, bright red. I've got freckles too, though fewer now than when i was a kid. I'm 20 now. Anyways, the only thing that i've noticed about having red hair is that you get sunburnt easily. Other than that, nothing. In fact, i think i have a higher than average tolerance for pain. It's hard to compare levels of pain with others, of course, but I don't find myself hurting more than others at all.
The article states that, in general, redheads felt more pain. But what if i don't? What if i feel less pain than average? Directly linking hair-color to pain reception is tenuous at best, and it seems like doctors messing with my anestheisa beacuase of my hair color could be bad, with the threat of coma.
I've never heard of hair color affecting ANYTHING before. Sure race affects some things, like reactions to certain drugs, and diseases. Sex and age as well will determine the best treatment. But hair color? What about eye color, does that factor in?
It seems like this is a broad generaliztion, and i don't think this can be the deciding criteria for pain reception. I don't buy it this first time round, if only based on personal experience.
yeah, actually, i did read it. Slashdot has that handy feature where it tells you when someone has replied to you.
Unfortunately, I am visiting my parents for [Canadian] thanksgiving, and i don't have access to Linux atm. anyways, it was interesting to read how you think it should work. I've never seen, well, probably just not noticed that behavior before, but i'll check for it. To me, it seems like quite a minor issue, as how many people scroll down page by page to the bottom of long documents, without letting up on the mouse, hence making the scrolling extremely fast. I guess if you do though, that's good enough for it to be an issue for you.
And yes, AFAIK, this is the typical windows behavior, the way it behaves now. Also, have you tried submitting a bug about this scrolling stuff? Do you even know if the devs realize it's broken? Because if you speak up, they may just fix it.
The scrollbars don't work properly. ... [explanation provided]
To be breif; No, you're wrong. I'm running Mozilla 1.0 on Linux right now. Gentoo Linux, to be specific. I honestly don't know what you're saying is broken. I tried the middle button. No matter what i do, the slider always jumps to the cursor's position, and dragging from there moves it. Using the left mouse button above and below the slider scrolls up or down a page, and holding the button down scrolls the page, screen by screen, until the slider reaches the cursor, or the bottom of the page, whichever comes first.
Either you're basing your views on an old version of Mozilla, or your system is bahiving differently from mine, for whatever reason. There's another possibility, too, but i'll refrain from mentioning it.
BK is not free software. If they do somehow end up in court for this, violating anti-trust laws, you know what Bit Mover will do? They will say, fine, if this license is illegal, there is none anymore. No free license at all. And then BK will not be free as in beer, not at all. You will not be able to get BK in a free form, you will only be able to get the pay version.
And will tha be a victory? Will righteousness and truth have triumphed? Maybe, but it won't help me out one bit. We'll be worse of than we are right now. But you can hold your own opinion.
This does not work for me. I'm using Win2k with Mozilla 1.1. I also tried IE. Then i tried Lynx on a linux box. Then another linux box, on an entirely different network, and ISP. They all tell me that the host cannot be located. Which is funny, cause that's exactly what i expected to happen, since this isn't a valid IP, nor a valid DNS name. Oh, i also just asked my buddy if he could see it an no, he can't.
What's the short version?
A) The license forbids you to use BK to further a direct competitor to BK. Distributing a competitor, while using BK, like Red Hat does, is allowed.
B) This license is the FREE license. Remember the saying, "Beggars can't be choosers?" They can't. Are you using BK for free? Then you can't expect to choose the license. If you buy the program, you can develop whatever you like with it.
C) Anyone still has the ability to be a kernel hacker without using BK whatsoever. The old tools still work, Linus and everyone else still accepts standard patches. It's just the old tools are actually worse than BK. BK was chosen purely on technical merits, it's only the license that's raising questions.
Point B) is important. Because this is the FREE license, it means that BM is not violating anti-trust laws by forbidding competition, because you can purchase the product, and get unrestricted use. Companies are not required to provide free samples of their products to competitors to help them out. Also, it means that BM is NOT acting like MS when they pulled the same stunt in their EULA. (Adding a clause stating that you cannot use MS products to harm MS in any way).
Summary: Bit Mover is acting reasonably, and completely within their rights as a company to define the acceptable uses of their free gift to users. The issue should is not whether or not Bit Mover is 'cheating' people. The issue now is whether or not to use Bit Keeper personally.
Course, maybe this BitKeeper appeals to managers more than actual developers...
I'm not a developer, but you know who is? Linus Torvalds. And you know what else? I have a sneaking feeling that he's a better developer than you. And you know what he thinks? He thinks BitKeeper is better than CVS. And you know what? All the other kernel developers agree with him. The only thing they can't reach a consensus about is the liscense.
The isssue is not whether BK is better than CVS or not. BK is better. That's been decided. The issue is about the liscense.
Also, for numbers between the dots in IP addresses, the cap is NOT usually 255.
What? What does this mean? What are you trying to say? Your sig URL doesn't work, this statement doesn't make any sense, and you don't provide any backing to it at all.
Explain yourself.
First, let me state that by no means am i an expert on IPv4 or v6, but it think i can add something here.
Your idea about assigning groups of IP addresses to certain orginizations (like universities) actually HAS happened, but not quite in the organized way you described. For instance, for a while, most cable modems you saw had a 24.x.x.x IP address. The reason for that was that @Home owned the entire 24 subnet. I belive that other companies were assigned entire 'Class A' subnets as they were called. AT&T has 12.x.x.x, i believe, Genuity has 4.x.x.x. Universities seem to usually have a 'Class B' address. The university of Alberta (in canada, where i am) has all 129.128.x.x addresses. ISP's usually own blocks of IP addresses, which are usually assigned in the least amounts that are required. So if a small isp only needs 100 IP's, they'll give them something like 205.132.211.x, which gives them 254 usable addresses. The bigger the organization, the more IP's are assigned to them. Currently, this is done by IANA in North America, and RIPE in Europe.
The other point is in response to your statement "but numbers don't really matter, so why not just let people at them?"
Well, the answer is IP Addresses DO matter. Your IP address determines how traffic is routed to you. It would be nearly impossible to keep track of how to route traffic to different ip's if 123.123.123.1 was in bejing, and 123.123.123.2 was in amsterdam. Routing logic says these 2 ip's will be routed differently only at the last step. If each IP were different, you'd have MASSIVE routing tables which would have to keep track of how to get to each ip around the word, which would be further complicate with dynamic ips, etc.
The upshot of this is that routing is very hard, and you can't decide numbers on a whim, just like you can't choose your own phone number.
> Ironically, a large number of the websites were defaced shortly thereafter.
Umm... Shouldn't that read, "Expectedly, a large number of websites were defaced shortly thereafter." ?
>I would plan on an amp per server at least, and go no more than 16 boxes per circuit.
Now, i'm not an electrician, but this is probably the worst bit of advice here. Don't get me wrong, electricity is important, but let's do the math, and determine how good your advice is.
Amps x Volts = Wattage.
Therefore, assuming you're using 110 Volt power, and your calculation of 1 amp per server, we get:
1 x 110 = 110 Watts, per sever.
It should be painfully obvious from the above that this is obviously way too little power to be running a server on. (to say nothing of your expertise with electricity and servers.)
Now, a 20 amp circuit gives you 2220 watts of power. ( 20A x 110v = 2200W ) Assuming 500 watts per server, you can run about 4 boxes per circuit. Remember, computers run on electricity, and if you rob them of power, they are entirely useless without it. Do not skimp on power.
In fact, if you want to do power right, you will run ONE circuit PER server. This way, if you need to cut power, or a breaker blows, only one server is affected, not the entire rack. And also, you don't have to worry about not having enough power for a big server, unless your server wants more than 2000 watts, which is less likely, unless you're getting some big iron servers. And those would likely have dual (or more) power supplies. And you would want them plugged into seperate circuits, because if one fails, the other is a backup.
The overall point is, it's hard to have too much power, and with some simple math, it's easy to figure out your needs, and what to set up. Compared to the cost of a server, a curcuit is pretty cheap to install, (i'm guessing $50 each) so there's no reason to cheat yourself here. If you want to do a server room right, you've got to have good power.
Note all wise readers: The parent is a troll.
The irony here is that i don't even think the author knew it was a troll. It even sparked some good comments debating the demand for broadband access. Good work.
However, the demand ISN'T THE QUESTION! @Home had 4.1 MILLION subscribers. That's 45% of the broadband market in the US, and growing at up to 100,000 new subs per month. To question the demand is silly. There is no question as to whether or not the demand is there, it most obviously is. The question is how @Home managed to screw it up. If you have 4.1 million people paying you $20/month for internet service, how to do you end up bankrupt? That's the question.
So ignore the demand question, and concentrate on the mismanagement question.