Their legacy services are connection-oriented. They have quality of service guarantees on packet delivery that can NOT be met on a best-effort network, period.
Bullshit. Imagine a point-to-point connection where every link in the network connecting them was unsaturated, all the routers had spare CPU capacity, and there was 0% packet loss. Under such conditions, which exist today and are purchasable, (though quite expensive for a home user) you can guarantee that the quality of a voice call would not falter at any time.
The main reason is the same as why QOS means nothing for an unsaturated link. Quality of service (QOS) is any mechanism by which packet delivery is prioritized. However, a prioritization decision can only be made when there is a queue. So, if you have 10 packets in a buffer, waiting to be sent, you can send packet 3 first, then 7, then 2, etc. You can essentially re-order the order in which the packets are transmitted, based upon some criteria.
This contrasts with the old-school, "dumb" way of handling packet delivery: FIFO. First-in, first-out is the simplest kind of buffer. As the name states, the first packet into the buffer, will be the first packet to leave. So in a 10 packet queue, packet 1 (being at the front of the line) is sent first, then packet 2, 3, 4, and so on. The ordering completely ignores any data in the packet, including QOS tags, port number, IP... it ignores it all. This is typically how routers on the internet behave.
However, when links are unsaturated, packets are processed immediately. So, when packet 1 is received by the router, it is transmitted with delays that are in the microsecond range. There is "room" on the line, and so the packet is sent without queuing alongside other packets. So, with a queue of 0 or 1 packet, the most complex, super-sophisticated QOS algorithm behave exactly the same way as FIFO: They will deliver the single packet immediately.
To come back to the point, you can provide 100% quality on a best-effort network so long as no packets are dropped due to noise, saturated links, or saturated routers.
I apologize for not having any refernces to cite, but i assure you this is true.
Notice how the article ends on the tired "it'll be good to the consumer" strawman:
underinvestment by some operators may "drive quality traffic to quality networks.
I think you may have misunderstood that quote. He's not saying it's good for the consumer, he's saying that competition in the ISP industry is the answer to capacity problems.
That quote is from a Level(3) excutive. Level(3) is a competitor to Qwest, and is saying that Qwest is out of money, and can't upgrade capacity due to financial reasons, not technical ones. He says, "With appropriate continuing investment, the Internet is capable of handling any application. What we're starting to see is a distinction between those operators who have the capital to fund expansion and those that don't."
He's saying that Qwest customers will move to Level(3) because Level(3) will have a superior, non-overloaded network, which will "drive quality traffic to quality networks." So Youtube won't purchase bandwidth from the old, slow ISP, but the modern, fast one. Same with the savvy home or business customer.
In that sense, i thought the article ended well, but started poorly, especially the headline.
But it also reminds me of those people in the fun-fun situation of not having a competitive ISP to choose. This happens in the US and Canada quite a lot, from what i hear. If your default cable/DSL ISP is good, then you're in luck. But if they are poor, well... You could pick the other, slower DSL/cable company, but that's not not really competitive choice, which is what's required.
I believe that if the goverment should be interfering with the ISP market in any way, it would be to foster competition. Technology is improving just fine, thanks, it's not technology's fault that Qwest is having capacity problems.
IDEA: City government builds fibre to your home, but offers no services. Instead, it allows free-market, competitive companies to offer whatever service they like (telephone, TV, Internet, radio) over the city-owned fibre from the CO (the competition point) to your house.
Thank you for the excellent explanation of how things work from the ISP side. However, i think you have betrayed the ISP by citing what you did when AOL started to suck: because AOL woefully misforecast those ratios, it became next to impossible to connect to AOL for quite a while until they caught up with modem provisioning (That's when I got rid of my AOL account and got my first real ISP acccount, yay!). Looks like everything old is new again. (emphasis mine)
This is exactly the point. If qwest is starting to offer shitty service, it's proposterous to blame the customer, and then talk as if the internet itself is breaking cause of these damnable users.
If company A is not capable of delivering a good product, i'm sure company B will have something you'd be more interested in.
Following this logic, you come back to the situation that many of us in Canada and the US are faced with: Lack of competive choices for an ISP, resulting, in this case, with shitty service being blamed on the customer. I hope enourmously that no one in goverment is buying this tortured logic, and making policy decisions based on it.
Do any of them not use Gecko? Yes, KHTML Seriously, man, he pre-answered your question, and yet you were still too lazy to look it up. How much hand-holding do you need?
Hey dude, i can see you like Yahoo's IM. You tell us, over and over, how great it is, how nothing will ever beat it, and, if i can quote you, "And did I mention free?" It's free, you boast. Free!
Free, eh? Tell me, what software besides Yahoo IM are you running? Is it Windows? Is Windows free? If so, please be quiet, and listen to was the Free Software guys are talking about, when they talk about open protocols, and Free (really) software.
Additionally, "...you may as well return your DVD collection too." DVDs are not sponsored by advertising, which is why they aren't free. Compare with your TV example. (Strangely, most people pay for TV, though...;)
It should be noted that $5k is for companies with more than $1 million revenue from the product with "Linux" in the name, not from the entire operation of the company. This is even more reasonable than the parent's assumption.
Also, remember that Linus didn't write copyright law, he must work within the legal framework he is given.
I see no issue with Govt. providing that infrastructure, but I do take issue with govt. providing the services passed with the infrastructure.
I agree with this, and i'd like to share how i think this could be done. This isn't my own idea, but it's the most appealing i've heard so far.
Like the phone company, you establish a central office in an urban area. Now, depending on the size of your city, you will probably need more than one. Let's guess each CO could service 10,000 customers.
This CO is owned and operated by the Fibre Utility, which is government owned. From the CO, the utility runs dedicated a fibre pair to each house (business, apartment, etc). Maybe even 2 pairs, while they're at it. So the result is 1 building with 10,000 fibre terminations.
In the CO, the fibre utility doesn't really have any equipment, they are not capable of providing a service to anyone, really. However, at fair and equitable rates, any company can place any equipment they like in the racks. And then, they can sell the service provided by that equipment to the customer. It could be television. It could be pure IP, nothing more. It could be the next future type service that no one's even thought of.
When you get your bill, you see something like:
FooBar Full Service
Internet: $25
TV bundle 2: $32 Fibre Utility
2 Active pairs, 2x$5: $10
The fibre utility could bill this to the service provider, i suppose, but i prefer the transparency of billing me directly for used fibre.
Now, the primary reason why i like this idea is: It fosters competition.
Over and over it has been shown that copetition is good for the consumer. And the more competition, the better. In this proposed scenario, a new competitor would only have to put equipment in 1 place to have the potential of servicing 10,000 customers. If you imagine a city of 500,000, that's 50 installs for the service provider, to cover an entire city. The idea here is the the infrastructure is common, but not the service.
If any of you remember the first dial-up days, (i'm thinking about 1992-2000) there were TONS of Dial-up ISPs, all competing based on price, online time, and various other criteria. It was excellent for the consumer, giving us plenty of choice, and driving prices down. Common infrastructure, competing service.
I don't know about the rest of you, but when was the last time your broadband connection price went down? Mine hasn't. And yet we can be sure that all the equipment costs are falling just like all computer hardware. Funny, isn't it?
I hear Verizon and Bellsouth are maybe rolling out fibre. This is cool, but that just entrenches their monopoly EVEN MORE, which is, of course, terrible for the consumer. This concerns me greatly. Our goal should be fantastic public infrastructure, not private ownership of all-important communications systems.
You're correct in stating that bandwidth does not equal latency. That's exactly correct. However, they are both important in determining the throughput of your transfer, and also they are interrelated, not orthogonal.
and the lag is pretty constant from 56k dialup to 100mbit lan... Now this is very wrong, and I'll prove it.
First, let us decide we have 1000 bytes of data to send. (we'll forget additional headers, etc, for now) This is 8000 bits.
At 56,000 bits per second, 8000 bits will take 143ms to be serialized on the line. That is, even if the line is 2 metres long, between your computer and the next, at 56kbps, it will always take a minimum of 143ms to send. And remember, this is one way, RTT (round trip time) is double, at 286ms.
Let us now contrast that with 100Mbps, or 100,000,000 bits per second. The serialization delay is now 0.08ms, or 80us (microseconds). The point, then, is that serialzation delay is now not really much of a factor in the overall latency of the connection, and instead propagation delay is the main factor. That is, how quickly the signal propagates through the medium. For fibre, this is about 2/3 the speed of light, or about 200,000 km per second. in 10ms, then, it will travel 2000 km.
What that means then, is that using remote X on 56k will give you at LEAST 300ms latency (which is noticable, and poor) wheras using 100Mbps fibre (given good conditions; no slow routers, overloaded links, etc) will provide you with a sub 10ms latency anywhere within the metro area of a city. The office to your house should have latency nearly as low as a LAN in your own home.
4 sources of latency: A) Propagation delay B) Serialization delay C) Routing/switching processing D) Buffering
In a good network, A is your main source of latency, which cannot be overcome, since it's so hard to go faster than the speed of light, you know.
My point, then, is that the network admins are wrong. They are not using the network as designed, and as such, things are breaking, ceasing to work... such as p2p apps, FTP, and a whole host of other things that are done on the internet.
The network is not aware of the structure of the data crossing it. This is by design, not accidental, not some kind of error.
Blaming the protocol because the network is absurdly implemented isn't fair, nor wise.
See also: active vs. passive FTP. Any protocol that requires remote hosts to connect back to your client is going to make your network admins hate you.
What should the network care? There's data going in, out, and all directions, all the time. Why does the network care about protocol?
(keep in mind that active FTP has, and continues to work, just fine for thousands of users, for like 20 years now)
first, your post is a bit incoherent, it's hard to tell what you're saying. However, the reality is this:
As we updated the User-Agent, we considered application-compatibility issues, historical precedent, and feedback from the community. We arrived at a very simple string.
IE7 running on Longhorn will send the following User-Agent header:Mozilla/4.0 (compatible; MSIE 7.0b; Windows NT 6.0)
And that doesn't seem like they've fixed things at all. They're just doing the same old, same old.
Counterpoint: Windows drivers are AWFUL; there are a myriad of problem, many differing, maybe better, maybe worse drivers, for the exact same device.
Proper Linux drivers are very, very good. They always work, are amazingly reliable, and are continually maintianed, fixed, and often vastly improved, with specific performance tuning for Linux, without the vendor even lifting a finger. All this available to you with no effort, just the upgrade of your kernel.
It boils down to the fact that linux drivers are just better, because of the way in which they're developed. Just like the Linux kernel itself, in fact. Only if you believe that Windows is actually superior to a GNU/Linux system can you hold the opinion that Linus (and all the other developers along with him) is wrong.
If 2- doesn't happen, then it's not theft. If, for example, I go to your house and break all your windows, then, 1- follows, you lost your property (the windows), but 2- doesn't, because I gained nothing. Then I would not be a thieve, I would be a mad man that breaks windows, a window destroyer, an aggresor, or I don't know what.
why don't you PRINT THE NAME of this company? They are a public company, are they not? Their name is known. Mentioning their name in their post, as an ISP that you should not choose, is a service in the public good.
Hiding them, keeping their evil ways secret is HELPING THEM, not us. Please, when you have a comment about a corporation, TELL US THE NAME so that we might avoid them. Concepts like freedom of speech protect you, and all of us, from persecution for telling the truth. Since you are telling the truth, no harm will come to you for speaking the whole truth, instead of aiding and abbetting these villians.
Please, slashbots, be BOLD in calling these corporations on their shennanigans.
But, at the end of the day, a LOT of people out there only care if it works or not.
Yes, i agree with you. However, people are short-sighted in their wants. They do not realize that in wanting things to 'just work', and thus using IE, they have actually made things worse for the rest of the world that doesn't want to pay money to microsoft. And that same user may be surprised that when they go to their public library which uses a Linux in order to save money, that the video doesn't work, because of their selfish decision to use IE just like everyone else.
A chief objection amongst the Open Source users to proprietary software is that proprietary software eventually leads to the day when you get screwed because the corporation you were relying on is greedy, and they don't actually like you. Their interest is always to take more money from you, not help you out. So despite the fact that their effort has made it 'just work', you might not be getting the best bargain in the long run.
I'm saying that people are responsible for the outcomes of their actions. And thus, people using microsoft software are responsible for the sucess of MS, and all that MS has brought with them, like horrible business practises, support of DRM, and software patents like the one that killed email reform, which would probably have been the most important development in email in the last 20 years.
So despite the logic "the user just wants things to work" supporting using IE, in the long run, supporting open standards will help you and your fellow human EVEN MORE, despite any short term pain.
Ah i see, you're actually using DNS as a fail-over procedure. This is valid, i guess, and in my other comment (just above, it should appear) i actually use this in my own setup, but only because i have a shoestring budget.
For failure, using DNS like this is poor, (your response time is still 5 minutes) but using it as a maintenence method to move traffic to a different cluster should work just fine.
I suppose i didn't expect your fail-over procedures to involve DNS, because i do not feel DNS is a very good solution. But it makes sense. Thanks for the response.
You are aware that TTLs are set on a per-record basis, right?
So the process goes like this. You realize that some server will be moving to a new IP (for example). The TTL is currently 24 hours. Change the TTL to 5 minutes. Wait 24 hours for all old cached records to expire. Then, move your server, change the record, and voila, no one will remember the old record for more than 5 minutes. The updated record with the new IP can have it's TTL set back to some high number, like 24 hours.
The only reason you would want to put a low TTL on a record and leave it there is if you expect that record to change, and thus the clients to need new information. As an example, a small IRC network i run has servers that are prone to failing. The TTL on the round-robin DNS record is 10 mins, cause i want people to have a minimum amount of time (less than 10 mins) where they could attempt to connect to a downed server. However, the IN A record of the actual server (who's IP isn't changing, they're just down) is much higher, i think 6 hours.
In additional response to the parent: DNS is not instant. You can chose the length of others caching your info, so you should chose judiciously. Nothing in your comment contradics this.
That being said, I have a PC, and I just extracted the AIFF files... never saw the license;)
Wait, do you mean you have windows?
When did personal computer, aka PC, come to mean windows? WTF is going on here? A Mac is a PC, a Dell is a PC, an my Linux box is a PC. Can i make use of the data on my Linux box?
I would also like to add that even if the scenario is playing out as those who made the scenario would have us believe--This is, artists are fairly compensated for lost sales--that the reality is still divorced from this.
It is important to note that the tariff is delivered to COPYRIGHT HOLDERS. Pick up a CD, any CD, and examine who the copyright is owned by. Then pick up a book and do the same thing. Chances are that with the CD, the copyright holder is a corporation, a record company (after all, they are in the business of selling records, which they buy from the artists, then own themselves) and in the book, it's probably the author.
So now, in reality, my tariffs are not really ever going to Celine Dion, Nickelback, The Hip, or Avril, even though they are rich and probably their music is downloaded more than anyone. The reason is because they don't own the copyright, the record company does. Therefore, it's BMI, and Sony, and Vivendi that are getting the tariff. You may note that these companies have never authored a creative work, or enriched your lives in any way through the creation or performance of music. Humans create music, not corporations. (You may argue that corps are now considered people, but that's a seperate discussion)
Directly, legally, and with moral justfication, large record companies have found a way to convinced us to allow them to tax the public. I'm amazed that any goverment empolyee ever though this was a good idea. But corporations have become very good at lying, and i think we're really starting to see the effects of it.
(btw, i've based my argument in this post not on any personal or aquired knowledge of the process, (in fact, i'm not sure anyone has clear knowledge of the process, disturbing in itself) but merely by observing the facts in front of me. The rule says 'copyright holder' and the copyright holder is not the band, not the artist, but the record company, as is clear from the markings on the CD. So i assume that if they are doing what they say they are doing, the money goes to record companies, not artists. But i really wish i could find out if this is true, with, you know, evidence)
As a businessman, you should understand that businesses are not nice, they are not generous, and they cannot be trusted at their word.
Linux and the OSS community have given given given given given things away, and asked for nothing in return but that you continue to share, as they shared with you. BitMover's goal is to profit, the same as any business. Somehow you figure that the OSS community is acting unfairly?
You say the OSDL went back on it's "word". Not like all those noble and honourable corporations who keep their word and come through on promises, right? Oh wait, they DID just backtrack and take back their product, something the OSS community will NEVER do, for in fact they have stated from the start that NO ONE is allowed to take it back.
It seems that businesses play hardball, they play tough, they play to win. There is no time to feel bad for a corp who's not doing well, it's just business. BitMover has decided to pull the product from being free. That's just business.
Oh, and please don't try to tell me that corporations have morals, or could hold the moral high ground on any issue. *Humans* have morals, principles, and integrity. Not corporations. Corporations are merely a profit machine, and it's only the fact that humans run them that they are kept in check.
This comment may be late, and my get buried, but i just wanted to correct the slashdot title for this article. (Which is strange cause/. is so reliable for facts)
it is: X Window system it is not: X Windows system
Can you see the difference? There is no s on 'window'. I know that MS has taught us all to use the word 'windows', but we should keep our heads and use the correct names for technology.
As a reference, i will cite the X.org Website where they make reference to the "X Window System" extensively.
Thanks Zonk. You couldn't even copy from the submitter's words, who got it correct.
I would suggest that it's very hard to argue that Windows is ready for the internet. In fact, when it comes to the internet, Windows it the worst OS, bar none. It gets viruses, worms, is remotely rooted, becomes a horrible piece of junk that spews spam, and all this after merely an hour connected to the internet. You know, the real internet, without firewalls and anti-virus, and anti-spyware and popup blockers.
I do not contest that Windows has a pretty good UI. It's just that all the technical stuff is so bad, i cannot use windows without embarassing myself.
Slashdot Mirror
Their legacy services are connection-oriented. They have quality of service guarantees on packet delivery that can NOT be met on a best-effort network, period.
Bullshit. Imagine a point-to-point connection where every link in the network connecting them was unsaturated, all the routers had spare CPU capacity, and there was 0% packet loss. Under such conditions, which exist today and are purchasable, (though quite expensive for a home user) you can guarantee that the quality of a voice call would not falter at any time.
The main reason is the same as why QOS means nothing for an unsaturated link. Quality of service (QOS) is any mechanism by which packet delivery is prioritized. However, a prioritization decision can only be made when there is a queue. So, if you have 10 packets in a buffer, waiting to be sent, you can send packet 3 first, then 7, then 2, etc. You can essentially re-order the order in which the packets are transmitted, based upon some criteria.
This contrasts with the old-school, "dumb" way of handling packet delivery: FIFO. First-in, first-out is the simplest kind of buffer. As the name states, the first packet into the buffer, will be the first packet to leave. So in a 10 packet queue, packet 1 (being at the front of the line) is sent first, then packet 2, 3, 4, and so on. The ordering completely ignores any data in the packet, including QOS tags, port number, IP... it ignores it all. This is typically how routers on the internet behave.
However, when links are unsaturated, packets are processed immediately. So, when packet 1 is received by the router, it is transmitted with delays that are in the microsecond range. There is "room" on the line, and so the packet is sent without queuing alongside other packets. So, with a queue of 0 or 1 packet, the most complex, super-sophisticated QOS algorithm behave exactly the same way as FIFO: They will deliver the single packet immediately.
To come back to the point, you can provide 100% quality on a best-effort network so long as no packets are dropped due to noise, saturated links, or saturated routers.
I apologize for not having any refernces to cite, but i assure you this is true.
Notice how the article ends on the tired "it'll be good to the consumer" strawman:
underinvestment by some operators may "drive quality traffic to quality networks.
I think you may have misunderstood that quote. He's not saying it's good for the consumer, he's saying that competition in the ISP industry is the answer to capacity problems.
That quote is from a Level(3) excutive. Level(3) is a competitor to Qwest, and is saying that Qwest is out of money, and can't upgrade capacity due to financial reasons, not technical ones. He says, "With appropriate continuing investment, the Internet is capable of handling any application. What we're starting to see is a distinction between those operators who have the capital to fund expansion and those that don't."
He's saying that Qwest customers will move to Level(3) because Level(3) will have a superior, non-overloaded network, which will "drive quality traffic to quality networks." So Youtube won't purchase bandwidth from the old, slow ISP, but the modern, fast one. Same with the savvy home or business customer.
In that sense, i thought the article ended well, but started poorly, especially the headline.
But it also reminds me of those people in the fun-fun situation of not having a competitive ISP to choose. This happens in the US and Canada quite a lot, from what i hear. If your default cable/DSL ISP is good, then you're in luck. But if they are poor, well... You could pick the other, slower DSL/cable company, but that's not not really competitive choice, which is what's required.
I believe that if the goverment should be interfering with the ISP market in any way, it would be to foster competition. Technology is improving just fine, thanks, it's not technology's fault that Qwest is having capacity problems.
IDEA: City government builds fibre to your home, but offers no services. Instead, it allows free-market, competitive companies to offer whatever service they like (telephone, TV, Internet, radio) over the city-owned fibre from the CO (the competition point) to your house.
Thank you for the excellent explanation of how things work from the ISP side. However, i think you have betrayed the ISP by citing what you did when AOL started to suck:
because AOL woefully misforecast those ratios, it became next to impossible to connect to AOL for quite a while until they caught up with modem provisioning (That's when I got rid of my AOL account and got my first real ISP acccount, yay!). Looks like everything old is new again.
(emphasis mine)
This is exactly the point. If qwest is starting to offer shitty service, it's proposterous to blame the customer, and then talk as if the internet itself is breaking cause of these damnable users.
If company A is not capable of delivering a good product, i'm sure company B will have something you'd be more interested in.
Following this logic, you come back to the situation that many of us in Canada and the US are faced with: Lack of competive choices for an ISP, resulting, in this case, with shitty service being blamed on the customer. I hope enourmously that no one in goverment is buying this tortured logic, and making policy decisions based on it.
Do any of them not use Gecko?
Yes, KHTML Seriously, man, he pre-answered your question, and yet you were still too lazy to look it up. How much hand-holding do you need?
Some that have only 1 "high-speed" option, that isn't really even high-speed.
So when SBC is the only provider of fibre connections, what will you say of cites with only one high-speed option?
What i'm saying is that you're just headed for another monopoly situation, aren't you? SBC doesn't love you.
Hey dude, i can see you like Yahoo's IM. You tell us, over and over, how great it is, how nothing will ever beat it, and, if i can quote you, "And did I mention free?" It's free, you boast. Free!
;)
Free, eh? Tell me, what software besides Yahoo IM are you running? Is it Windows? Is Windows free? If so, please be quiet, and listen to was the Free Software guys are talking about, when they talk about open protocols, and Free (really) software.
Additionally, "...you may as well return your DVD collection too." DVDs are not sponsored by advertising, which is why they aren't free. Compare with your TV example. (Strangely, most people pay for TV, though...
It should be noted that $5k is for companies with more than $1 million revenue from the product with "Linux" in the name, not from the entire operation of the company. This is even more reasonable than the parent's assumption.
Also, remember that Linus didn't write copyright law, he must work within the legal framework he is given.
I see no issue with Govt. providing that infrastructure, but I do take issue with govt. providing the services passed with the infrastructure.
I agree with this, and i'd like to share how i think this could be done. This isn't my own idea, but it's the most appealing i've heard so far.
Like the phone company, you establish a central office in an urban area. Now, depending on the size of your city, you will probably need more than one. Let's guess each CO could service 10,000 customers.
This CO is owned and operated by the Fibre Utility, which is government owned. From the CO, the utility runs dedicated a fibre pair to each house (business, apartment, etc). Maybe even 2 pairs, while they're at it. So the result is 1 building with 10,000 fibre terminations.
In the CO, the fibre utility doesn't really have any equipment, they are not capable of providing a service to anyone, really. However, at fair and equitable rates, any company can place any equipment they like in the racks. And then, they can sell the service provided by that equipment to the customer. It could be television. It could be pure IP, nothing more. It could be the next future type service that no one's even thought of.
When you get your bill, you see something like:
FooBar Full Service
Internet: $25
TV bundle 2: $32
Fibre Utility
2 Active pairs, 2x$5: $10
The fibre utility could bill this to the service provider, i suppose, but i prefer the transparency of billing me directly for used fibre.
Now, the primary reason why i like this idea is: It fosters competition.
Over and over it has been shown that copetition is good for the consumer. And the more competition, the better. In this proposed scenario, a new competitor would only have to put equipment in 1 place to have the potential of servicing 10,000 customers. If you imagine a city of 500,000, that's 50 installs for the service provider, to cover an entire city. The idea here is the the infrastructure is common, but not the service.
If any of you remember the first dial-up days, (i'm thinking about 1992-2000) there were TONS of Dial-up ISPs, all competing based on price, online time, and various other criteria. It was excellent for the consumer, giving us plenty of choice, and driving prices down. Common infrastructure, competing service.
I don't know about the rest of you, but when was the last time your broadband connection price went down? Mine hasn't. And yet we can be sure that all the equipment costs are falling just like all computer hardware. Funny, isn't it?
I hear Verizon and Bellsouth are maybe rolling out fibre. This is cool, but that just entrenches their monopoly EVEN MORE, which is, of course, terrible for the consumer. This concerns me greatly. Our goal should be fantastic public infrastructure, not private ownership of all-important communications systems.
Comments?
You're correct in stating that bandwidth does not equal latency. That's exactly correct. However, they are both important in determining the throughput of your transfer, and also they are interrelated, not orthogonal.
and the lag is pretty constant from 56k dialup to 100mbit lan...
Now this is very wrong, and I'll prove it.
First, let us decide we have 1000 bytes of data to send. (we'll forget additional headers, etc, for now) This is 8000 bits.
At 56,000 bits per second, 8000 bits will take 143ms to be serialized on the line. That is, even if the line is 2 metres long, between your computer and the next, at 56kbps, it will always take a minimum of 143ms to send. And remember, this is one way, RTT (round trip time) is double, at 286ms.
Let us now contrast that with 100Mbps, or 100,000,000 bits per second. The serialization delay is now 0.08ms, or 80us (microseconds). The point, then, is that serialzation delay is now not really much of a factor in the overall latency of the connection, and instead propagation delay is the main factor. That is, how quickly the signal propagates through the medium. For fibre, this is about 2/3 the speed of light, or about 200,000 km per second. in 10ms, then, it will travel 2000 km.
What that means then, is that using remote X on 56k will give you at LEAST 300ms latency (which is noticable, and poor) wheras using 100Mbps fibre (given good conditions; no slow routers, overloaded links, etc) will provide you with a sub 10ms latency anywhere within the metro area of a city. The office to your house should have latency nearly as low as a LAN in your own home.
4 sources of latency:
A) Propagation delay
B) Serialization delay
C) Routing/switching processing
D) Buffering
In a good network, A is your main source of latency, which cannot be overcome, since it's so hard to go faster than the speed of light, you know.
My point, then, is that the network admins are wrong. They are not using the network as designed, and as such, things are breaking, ceasing to work... such as p2p apps, FTP, and a whole host of other things that are done on the internet.
The network is not aware of the structure of the data crossing it. This is by design, not accidental, not some kind of error.
Blaming the protocol because the network is absurdly implemented isn't fair, nor wise.
See also: active vs. passive FTP. Any protocol that requires remote hosts to connect back to your client is going to make your network admins hate you.
What should the network care? There's data going in, out, and all directions, all the time. Why does the network care about protocol?
(keep in mind that active FTP has, and continues to work, just fine for thousands of users, for like 20 years now)
first, your post is a bit incoherent, it's hard to tell what you're saying. However, the reality is this:
As we updated the User-Agent, we considered application-compatibility issues, historical precedent, and feedback from the community. We arrived at a very simple string.
IE7 running on Longhorn will send the following User-Agent header: Mozilla/4.0 (compatible; MSIE 7.0b; Windows NT 6.0)
And that doesn't seem like they've fixed things at all. They're just doing the same old, same old.
... and you have a 4 digit UID? How long has this been puzzling you, more than 5 years? Amazing, man. :P
Counterpoint: Windows drivers are AWFUL; there are a myriad of problem, many differing, maybe better, maybe worse drivers, for the exact same device.
Proper Linux drivers are very, very good. They always work, are amazingly reliable, and are continually maintianed, fixed, and often vastly improved, with specific performance tuning for Linux, without the vendor even lifting a finger. All this available to you with no effort, just the upgrade of your kernel.
It boils down to the fact that linux drivers are just better, because of the way in which they're developed. Just like the Linux kernel itself, in fact. Only if you believe that Windows is actually superior to a GNU/Linux system can you hold the opinion that Linus (and all the other developers along with him) is wrong.
If 2- doesn't happen, then it's not theft.
:)
If, for example, I go to your house and break all your windows, then, 1- follows, you lost your property (the windows), but 2- doesn't, because I gained nothing. Then I would not be a thieve, I would be a mad man that breaks windows, a window destroyer, an aggresor, or I don't know what.
A vandal, i believe. Destruction of property is known as vandalism. See: http://en.wikipedia.org/wiki/Vandalism
(BTW, great post. To talk about theft, we must define what it is, and what it is not. Just thought i'd fill in the part you were not sure of.
why don't you PRINT THE NAME of this company? They are a public company, are they not? Their name is known. Mentioning their name in their post, as an ISP that you should not choose, is a service in the public good.
Hiding them, keeping their evil ways secret is HELPING THEM, not us. Please, when you have a comment about a corporation, TELL US THE NAME so that we might avoid them. Concepts like freedom of speech protect you, and all of us, from persecution for telling the truth. Since you are telling the truth, no harm will come to you for speaking the whole truth, instead of aiding and abbetting these villians.
Please, slashbots, be BOLD in calling these corporations on their shennanigans.
But, at the end of the day, a LOT of people out there only care if it works or not.
Yes, i agree with you. However, people are short-sighted in their wants. They do not realize that in wanting things to 'just work', and thus using IE, they have actually made things worse for the rest of the world that doesn't want to pay money to microsoft. And that same user may be surprised that when they go to their public library which uses a Linux in order to save money, that the video doesn't work, because of their selfish decision to use IE just like everyone else.
A chief objection amongst the Open Source users to proprietary software is that proprietary software eventually leads to the day when you get screwed because the corporation you were relying on is greedy, and they don't actually like you. Their interest is always to take more money from you, not help you out. So despite the fact that their effort has made it 'just work', you might not be getting the best bargain in the long run.
I'm saying that people are responsible for the outcomes of their actions. And thus, people using microsoft software are responsible for the sucess of MS, and all that MS has brought with them, like horrible business practises, support of DRM, and software patents like the one that killed email reform, which would probably have been the most important development in email in the last 20 years.
So despite the logic "the user just wants things to work" supporting using IE, in the long run, supporting open standards will help you and your fellow human EVEN MORE, despite any short term pain.
Ah i see, you're actually using DNS as a fail-over procedure. This is valid, i guess, and in my other comment (just above, it should appear) i actually use this in my own setup, but only because i have a shoestring budget.
For failure, using DNS like this is poor, (your response time is still 5 minutes) but using it as a maintenence method to move traffic to a different cluster should work just fine.
I suppose i didn't expect your fail-over procedures to involve DNS, because i do not feel DNS is a very good solution. But it makes sense. Thanks for the response.
You are aware that TTLs are set on a per-record basis, right?
So the process goes like this. You realize that some server will be moving to a new IP (for example). The TTL is currently 24 hours. Change the TTL to 5 minutes. Wait 24 hours for all old cached records to expire. Then, move your server, change the record, and voila, no one will remember the old record for more than 5 minutes. The updated record with the new IP can have it's TTL set back to some high number, like 24 hours.
The only reason you would want to put a low TTL on a record and leave it there is if you expect that record to change, and thus the clients to need new information. As an example, a small IRC network i run has servers that are prone to failing. The TTL on the round-robin DNS record is 10 mins, cause i want people to have a minimum amount of time (less than 10 mins) where they could attempt to connect to a downed server. However, the IN A record of the actual server (who's IP isn't changing, they're just down) is much higher, i think 6 hours.
In additional response to the parent: DNS is not instant. You can chose the length of others caching your info, so you should chose judiciously. Nothing in your comment contradics this.
Why do you use 5 minute TTLs? Do you really have machines that change IPs with less that 5 minutes of lead time?
That being said, I have a PC, and I just extracted the AIFF files... never saw the license ;)
Wait, do you mean you have windows?
When did personal computer, aka PC, come to mean windows? WTF is going on here? A Mac is a PC, a Dell is a PC, an my Linux box is a PC. Can i make use of the data on my Linux box?
I would also like to add that even if the scenario is playing out as those who made the scenario would have us believe--This is, artists are fairly compensated for lost sales--that the reality is still divorced from this.
It is important to note that the tariff is delivered to COPYRIGHT HOLDERS. Pick up a CD, any CD, and examine who the copyright is owned by. Then pick up a book and do the same thing. Chances are that with the CD, the copyright holder is a corporation, a record company (after all, they are in the business of selling records, which they buy from the artists, then own themselves) and in the book, it's probably the author.
So now, in reality, my tariffs are not really ever going to Celine Dion, Nickelback, The Hip, or Avril, even though they are rich and probably their music is downloaded more than anyone. The reason is because they don't own the copyright, the record company does. Therefore, it's BMI, and Sony, and Vivendi that are getting the tariff. You may note that these companies have never authored a creative work, or enriched your lives in any way through the creation or performance of music. Humans create music, not corporations. (You may argue that corps are now considered people, but that's a seperate discussion)
Directly, legally, and with moral justfication, large record companies have found a way to convinced us to allow them to tax the public. I'm amazed that any goverment empolyee ever though this was a good idea. But corporations have become very good at lying, and i think we're really starting to see the effects of it.
(btw, i've based my argument in this post not on any personal or aquired knowledge of the process, (in fact, i'm not sure anyone has clear knowledge of the process, disturbing in itself) but merely by observing the facts in front of me. The rule says 'copyright holder' and the copyright holder is not the band, not the artist, but the record company, as is clear from the markings on the CD. So i assume that if they are doing what they say they are doing, the money goes to record companies, not artists. But i really wish i could find out if this is true, with, you know, evidence)
As a businessman, you should understand that businesses are not nice, they are not generous, and they cannot be trusted at their word.
Linux and the OSS community have given given given given given things away, and asked for nothing in return but that you continue to share, as they shared with you. BitMover's goal is to profit, the same as any business. Somehow you figure that the OSS community is acting unfairly?
You say the OSDL went back on it's "word". Not like all those noble and honourable corporations who keep their word and come through on promises, right? Oh wait, they DID just backtrack and take back their product, something the OSS community will NEVER do, for in fact they have stated from the start that NO ONE is allowed to take it back.
It seems that businesses play hardball, they play tough, they play to win. There is no time to feel bad for a corp who's not doing well, it's just business. BitMover has decided to pull the product from being free. That's just business.
Oh, and please don't try to tell me that corporations have morals, or could hold the moral high ground on any issue. *Humans* have morals, principles, and integrity. Not corporations. Corporations are merely a profit machine, and it's only the fact that humans run them that they are kept in check.
This comment may be late, and my get buried, but i just wanted to correct the slashdot title for this article. (Which is strange cause /. is so reliable for facts)
it is: X Window system
it is not: X Windows system
Can you see the difference? There is no s on 'window'. I know that MS has taught us all to use the word 'windows', but we should keep our heads and use the correct names for technology.
As a reference, i will cite the X.org Website where they make reference to the "X Window System" extensively.
Thanks Zonk. You couldn't even copy from the submitter's words, who got it correct.
What about 'ready for the internet'?
I would suggest that it's very hard to argue that Windows is ready for the internet. In fact, when it comes to the internet, Windows it the worst OS, bar none. It gets viruses, worms, is remotely rooted, becomes a horrible piece of junk that spews spam, and all this after merely an hour connected to the internet. You know, the real internet, without firewalls and anti-virus, and anti-spyware and popup blockers.
I do not contest that Windows has a pretty good UI. It's just that all the technical stuff is so bad, i cannot use windows without embarassing myself.