Ditto. I've recommended it to two friends who've signed up for it--I only wish they had a referral program so I'd get a discount or something.:)
I've been using it for a few months and I'm quite pleased. It just runs and keeps things backed up. I have important stuff prioritized over unimportant stuff, so backups of changes to important stuff interrupt the unimportant backups. And it works well in Linux.
Same here. I've used Tb, Kmail, and offlineimap to do this, but always manually. I'm running offlineimap again now, but the last time was September of 2009 according to the last log file. Yeesh. I could cron it, but I might not always want it using my bandwidth to sync my email whenever it is set to run, so...urgh.
I used to do that. The problem is that catchall addresses catch all email, including spam sent to addresses that never existed. You're left to either blacklist addresses spammers send to (not from), or just let the spam filters catch all the spam. Blacklisting doesn't work since the spammers use random strings. You can whitelist addresses you give out, but that's not a catchall address--you have to make new addresses or whitelisting rules and keep track of what address you used on what site. So you end up relying on just the spam filters, which defeats the purpose of catchall addresses.
If your reason for using them is to notice when a site sells your address to spammers, ok, but that doesn't stop the spam from coming to the address. By that time, it's likely been years since you were on the site anyway. And you have to blacklist it manually. That's a losing game.
I used to use Sneakemail, but that boils down to whitelisting, and I'd have to lookup addresses to log in to sites. The time it saved from receiving spam was less than the time it wasted maintaining and using it.
The real solution is automated filters, and Gmail's do an excellent job. So now I just use one address for everything. It's nice to not have to lookup what address I used to register on what web site, and I actually get less spam than when I had a catchall address. I got more spam sent to non-existent addresses at my domain than to my real email address that can be found on Google.
For offsite backup I recommend CrashPlan: $50/year, unlimited space for one computer (or you can mount network drives so it thinks it's one computer), or $100/year for 10 computers. Files you delete are never deleted from CrashPlan's online backups. Works on Linux, too.
You can also just use their software to backup on local devices or to friends'/family's computers, and the software itself is free.
Another nice thing about it is that it does integrity checking all throughout the process, and over time, so that data degradation can be avoided. rsync and rdiff-backup, duplicity, etc. are great, but they don't protect against media degradation.
Originality is a myth. Every idea that ever has been had wouldn't have been had were it not for other ideas others had in the past. Anyone who claims he is owed for an idea he had is a hypocrite for he then also owes many others for the ideas that inspired his.
Also, people think of things all the time, everywhere, and have forever. There's no way to know who thought of an idea first, and it doesn't matter, since people can think of things independently at different times.
All these "controversies" are nothing more than a bunch of whining babies crying, "I thought of it first!" They think they're entitled to receive compensation for as long as they live because they "created" something, and that they should be able to control others' actions that have anything to do with anything related to anything they ever "created."
I'm getting so sick of the very idea of "intellectual property." Ideas cannot be--should not be able to be--owned. If there were a global referendum to abolish copyrights and patents, I'd vote in favor of it. Let those who want to keep secrets keep them as trade secrets, as many already do. Let others reverse engineer ideas and products, as they already do, but without legal problems. Let others use the ideas that flow into and out of societies to build upon and improve the world.
Let those who work be rewarded for their work, and let those who think of ideas also be rewarded for their work, not be eternally rewarded for a thought they once had, or a "work" they once did or recorded.
The world would be better off. (Can you imagine cavemen paying royalties to the guy who made the first wheel? "Hey, Grog, you can't cut that piece of wood into that shape and roll something on top of it! I thought of that! You have to give me your firstborn mastodon or else!")
As Jim Gettys and others commenting on this article have shown, many "real network engineers" don't or did not understand bufferbloat. Obviously that's so, otherwise the hardware with excessive buffers wouldn't have been produced and bufferbloat wouldn't have happened.
Your suggesting that end users throw away 40% of their last-mile bandwidth is a completely different matter than ISPs saving bandwidth for redundant paths. End-users have a single path: their cable or DSL connection. Those are two completely different problems with completely different appropriate solutions. One is irrelevant to the other. Why are you being misleading and dishonest by suggesting they're related? Or do you just not understand the difference?
I have the right to defend myself and my family against bodily harm. I should be able to exercise that right, and equip myself with the necessary tools, without being harassed by those who are so overcome by the fear of Evil that they are unable to face the Truth that Evil exists in the world and Good people must be prepared to face it, sometimes by responding with equal force. It's really none of your business if I want to carry a gun, as long as I don't misuse it and harm anyone or anyone else's property. It's not alive; it's not going to jump out of my holster and shoot anyone on its own, just like my car isn't going to run anyone over without someone driving it. Guns do not have cooties.
In the end, I don't care whether or not statistics support concealed-carry--though I believe they do. What matters is that I have the right to defend myself and my family, and that I value the lives of myself and my family enough to protect them. What's truly sad--almost as sad as there being people who commit acts of Evil--is that there are people who would sooner see Good, innocent people die at the hands of an Evil person, than to see Good people saved by killing an Evil person who's in the act of committing Evil. All human life is sacred--but all humans have Free Will and responsibility for their actions, and justice is no more served by allowing Evil people to commit their Evil acts than by allowing Good people to defend themselves against Evil people, even if it results in the deaths of Evil people.
Refusal to face the Truth is simply cowardice, and cowards' counsel should not be heeded.
Your defeatist attitude is sad. If there's nothing you can do but hope that some bad guy doesn't pick you to randomly kill and steal from someday, you might as well go live in a bunker.
It's not based in reality either. Few muggers would shoot before demanding what they want of a random person. They don't just go around killing people and taking their stuff--that's called pillaging. Our struggle is not against Vikings.
So you're either being foolish or dishonest. You're so afraid of having to face the reality that there is Evil in the world and that Good people must stand against it, sometimes with force, that you can't stand to consider the Truth. Maybe that's what Gore's documentary should have been about--the title fits.
I think you really should go live in a bunker so the rest of us can exercise our right to defend ourselves against bodily harm without being harassed for valuing our own lives.
Haha, wow, more assumptions. I said I spent some time working at a help desk, not that it is my career.
I think if you talked to real network engineers and real network programmers, like folks who worked on the protocols, they'd say that throwing away 40% of your bandwidth to fix latency is a ludicrous "solution."
You sure do make a lot of assumptions. I've been running my own systems and networks for over 20 years. I've done my own time as help desk support. I think I'm capable of figuring out if the single computer connected to this DSL connection, assembled and installed and maintained by myself, is running P2P apps. Guess what? It's not.
You expect me to rate-limit my upstream and downstream bandwidth to 60% of rated? That's ludicrous and totally unnecessary. Even if it were to fix the latency, that's an incredibly stupid "solution." I guess instead of fixing my car's unbalanced wheels, I'll just drive 30mph on the interstate.
Get off your high horse. Stop assuming you know me and my equipment better than I do. Either be helpful or be quiet.
Why wouldn't they store the data? What incentive do they have to delete it? Do you have any idea how much storage capacity Google has? The fact is that there shouldn't have been any privacy-sensitive data in there, because it was captured from in-the-clear transmissions.
Do you suggest that it ought to be illegal to stand on the sidewalk and look through a window in a building, or listen to the sounds coming out of a building? Because that's no different than receiving in-the-clear wireless network transmissions: it's just another kind of wireless transmission.
If Google had cracked encrypted data, that'd be a completely different matter.
If you're arguing that Google should have been more proactive to avoid capturing anything but the SSIDs, in the off chance that some other data might have been captured which wasn't intended for in-the-clear transmission, I can agree with that. But there's no way it should be required by law, and there's no way it should be a crime to not do so. The onus should be on those transmitting to ensure that they aren't transmitting sensitive data in the clear. If you say Google should be more responsible for their actions, the same principle applies to those running wireless networks: they also should exercise due diligence, and Google shouldn't be faulted for the transmitters' negligence.
Maybe you are allowed to do that. Have you checked any laws in your jurisdiction? If you're remaining on your own property or on public property, I'd be surprised if it were illegal for you to point a microphone or an antenna or a camera anywhere.
It shouldn't be illegal, because if it were, it'd always be a judgement call as to whether someone was being malicious, and innocent people could be hauled in by people who didn't like them for doing something that "looked bad." Oh, wait, that happens to people with cameras all the time since 9/11...
IMO it boils down to this: if you don't want people seeing into your house, close the curtains. If you don't want people hearing what goes in on your house, close the windows and doors, install soundproofing if necessary, or just don't talk so loudly. If you don't want people being able to listen in to your wireless network, encrypt it and change the key every so often. If a person is broadcasting his wireless network in the clear, yelling out the front door, or leaving his curtains open at night with the interior lights on, he shouldn't expect privacy.
Google didn't crack anyone's encryption or pick any locks, so I don't think there's a crime here.
Now if a government agency had done it, that'd be different...oh wait, I forgot about ECHELON, et al.
This whole argument against Google is like telling a neighbor to close their eyes when they look out their window which happens to line up with your window so that you can leave your curtains open and they won't see into your house. What you should do is close your curtains when you want privacy. Now if your neighbor sneaks over and puts a fiber-optic camera through the crack around the window frame, then you have a valid complaint.
As devices become "smarter" and more connected, these kinds of flaws and vulnerabilities will only increase in number and severity. It's highly unlikely that there will ever be enough economic incentive for manufacturers to keep the embedded software in their consumer devices secure and up-to-date, not to mention the lack of software update mechanisms.
This is why we need Free Software. Standard platforms running Free Software can be patched and updated simply and easily, and maintained by community efforts. Once the build system is in place for a device or a platform of devices, simple patches can be pushed out with little effort, regardless of a manufacturer's continued interest.
No consumer should be forced to choose between having a device that he can't trust or buying a newer device which has fixed software. That's not a valid reason to buy a new device. But, of course, the manufacturers and vendors would love for folks to buy new TVs every few years, even if the old ones are fine. "Security bugs? It's obsolete, friend. You need to buy a new set." To them it may be like another form of planned obsolescence.
That's my point exactly: the problem is about TCP, which does not "request packets." The "request" is made at a higher level, in HTTP.
You can throw around terminology like "OSI model" all you want, but you have demonstrated that you still don't understand how buffering or TCP connections work, and why bufferbloat is a real problem.
I disagree. If a connection's buffers were 50-100ms in size (perhaps even less than 50ms would be enough sometimes), the latency could never exceed 50-100ms, which is plenty for most interactive protocols. Maybe you're still thinking of buffers in terms of bytes, but it's much more helpful in this case to consider them in terms of the time it takes to empty them.
Did you read my comment? I said that downloading a single file via HTTP ends up filling the buffers and causing the latency. The way you "remove" what's in the buffer is to stop the download. That's the whole point: the latency only becomes a problem when the connection is being used. Your solution is to not use the connection. That's like a doctor telling a patient with a broken arm to just not use his arm anymore, instead of fixing the broken arm. I guess I'll just cancel my DSL and not use the Internet anymore.
Oh, you're right, it was my imaginary little brother using Kazaa. Why didn't I think of that? How silly of me....
I don't understand why you think he made up "bufferbloat" (the problem, that is, not the term). He clearly explained his methods and his work with other parties who are indeed experts. I've used Netalyzr on my own connection and it reports very large buffers on my DSL connection, and I've observed the exact same symptoms that Mr. Gettys has observed. Other people commenting on this article (even some who work in ISPs) have concurred.
If you want to convince me that bufferbloat is not a real problem, and not the cause of the observed symptoms, you'll need to provide some evidence of alternative explanations, because Mr. Gettys has done an excellent job of proving his case. He's also not the only person who's written about the problem, only the most recent.
If you read the series of articles and understand enough about how TCP works, it's clear that it is indeed a real problem.
I used to have a problem with DSL when it rained at another place I lived. But AT&T tried to charge me for the technician's visit. Any advice for getting them to fix their lines without charging me?:)
Slashdot Mirror
Ditto. I've recommended it to two friends who've signed up for it--I only wish they had a referral program so I'd get a discount or something. :)
I've been using it for a few months and I'm quite pleased. It just runs and keeps things backed up. I have important stuff prioritized over unimportant stuff, so backups of changes to important stuff interrupt the unimportant backups. And it works well in Linux.
Same here. I've used Tb, Kmail, and offlineimap to do this, but always manually. I'm running offlineimap again now, but the last time was September of 2009 according to the last log file. Yeesh. I could cron it, but I might not always want it using my bandwidth to sync my email whenever it is set to run, so...urgh.
I recommend CrashPlan.
I used to do that. The problem is that catchall addresses catch all email, including spam sent to addresses that never existed. You're left to either blacklist addresses spammers send to (not from), or just let the spam filters catch all the spam. Blacklisting doesn't work since the spammers use random strings. You can whitelist addresses you give out, but that's not a catchall address--you have to make new addresses or whitelisting rules and keep track of what address you used on what site. So you end up relying on just the spam filters, which defeats the purpose of catchall addresses.
If your reason for using them is to notice when a site sells your address to spammers, ok, but that doesn't stop the spam from coming to the address. By that time, it's likely been years since you were on the site anyway. And you have to blacklist it manually. That's a losing game.
I used to use Sneakemail, but that boils down to whitelisting, and I'd have to lookup addresses to log in to sites. The time it saved from receiving spam was less than the time it wasted maintaining and using it.
The real solution is automated filters, and Gmail's do an excellent job. So now I just use one address for everything. It's nice to not have to lookup what address I used to register on what web site, and I actually get less spam than when I had a catchall address. I got more spam sent to non-existent addresses at my domain than to my real email address that can be found on Google.
For offsite backup I recommend CrashPlan: $50/year, unlimited space for one computer (or you can mount network drives so it thinks it's one computer), or $100/year for 10 computers. Files you delete are never deleted from CrashPlan's online backups. Works on Linux, too.
You can also just use their software to backup on local devices or to friends'/family's computers, and the software itself is free.
Another nice thing about it is that it does integrity checking all throughout the process, and over time, so that data degradation can be avoided. rsync and rdiff-backup, duplicity, etc. are great, but they don't protect against media degradation.
Originality is a myth. Every idea that ever has been had wouldn't have been had were it not for other ideas others had in the past. Anyone who claims he is owed for an idea he had is a hypocrite for he then also owes many others for the ideas that inspired his.
Also, people think of things all the time, everywhere, and have forever. There's no way to know who thought of an idea first, and it doesn't matter, since people can think of things independently at different times.
All these "controversies" are nothing more than a bunch of whining babies crying, "I thought of it first!" They think they're entitled to receive compensation for as long as they live because they "created" something, and that they should be able to control others' actions that have anything to do with anything related to anything they ever "created."
I'm getting so sick of the very idea of "intellectual property." Ideas cannot be--should not be able to be--owned. If there were a global referendum to abolish copyrights and patents, I'd vote in favor of it. Let those who want to keep secrets keep them as trade secrets, as many already do. Let others reverse engineer ideas and products, as they already do, but without legal problems. Let others use the ideas that flow into and out of societies to build upon and improve the world.
Let those who work be rewarded for their work, and let those who think of ideas also be rewarded for their work, not be eternally rewarded for a thought they once had, or a "work" they once did or recorded.
The world would be better off. (Can you imagine cavemen paying royalties to the guy who made the first wheel? "Hey, Grog, you can't cut that piece of wood into that shape and roll something on top of it! I thought of that! You have to give me your firstborn mastodon or else!")
As Jim Gettys and others commenting on this article have shown, many "real network engineers" don't or did not understand bufferbloat. Obviously that's so, otherwise the hardware with excessive buffers wouldn't have been produced and bufferbloat wouldn't have happened.
Your suggesting that end users throw away 40% of their last-mile bandwidth is a completely different matter than ISPs saving bandwidth for redundant paths. End-users have a single path: their cable or DSL connection. Those are two completely different problems with completely different appropriate solutions. One is irrelevant to the other. Why are you being misleading and dishonest by suggesting they're related? Or do you just not understand the difference?
I think you meant "anti-materiel." :)
I have the right to defend myself and my family against bodily harm. I should be able to exercise that right, and equip myself with the necessary tools, without being harassed by those who are so overcome by the fear of Evil that they are unable to face the Truth that Evil exists in the world and Good people must be prepared to face it, sometimes by responding with equal force. It's really none of your business if I want to carry a gun, as long as I don't misuse it and harm anyone or anyone else's property. It's not alive; it's not going to jump out of my holster and shoot anyone on its own, just like my car isn't going to run anyone over without someone driving it. Guns do not have cooties.
In the end, I don't care whether or not statistics support concealed-carry--though I believe they do. What matters is that I have the right to defend myself and my family, and that I value the lives of myself and my family enough to protect them. What's truly sad--almost as sad as there being people who commit acts of Evil--is that there are people who would sooner see Good, innocent people die at the hands of an Evil person, than to see Good people saved by killing an Evil person who's in the act of committing Evil. All human life is sacred--but all humans have Free Will and responsibility for their actions, and justice is no more served by allowing Evil people to commit their Evil acts than by allowing Good people to defend themselves against Evil people, even if it results in the deaths of Evil people.
Refusal to face the Truth is simply cowardice, and cowards' counsel should not be heeded.
Your defeatist attitude is sad. If there's nothing you can do but hope that some bad guy doesn't pick you to randomly kill and steal from someday, you might as well go live in a bunker.
It's not based in reality either. Few muggers would shoot before demanding what they want of a random person. They don't just go around killing people and taking their stuff--that's called pillaging. Our struggle is not against Vikings.
So you're either being foolish or dishonest. You're so afraid of having to face the reality that there is Evil in the world and that Good people must stand against it, sometimes with force, that you can't stand to consider the Truth. Maybe that's what Gore's documentary should have been about--the title fits.
I think you really should go live in a bunker so the rest of us can exercise our right to defend ourselves against bodily harm without being harassed for valuing our own lives.
Haha, wow, more assumptions. I said I spent some time working at a help desk, not that it is my career.
I think if you talked to real network engineers and real network programmers, like folks who worked on the protocols, they'd say that throwing away 40% of your bandwidth to fix latency is a ludicrous "solution."
Isn't that where retransmitting comes in?
I'm still interested in how the HTTP process can request and[sic] entire file at once
GET /big.file HTTP/1.1
Host: example.com
Connection: keep-alive
Referer: http://slashdot.org/comments.pl?sid=1939940&cid=34826476
Accept: application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US) AppleWebKit/534.13 (KHTML, like Gecko) Chrome/9.0.597.45 Safari/534.13
Accept-Encoding: gzip,deflate,sdch
Accept-Language: en-US,en;q=0.8
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.3
and how that file gets sent back unattended
What do you mean by "unattended"?
I never said HTTP controls the flow of packets. You're seriously misunderstanding me.
Thanks for replying so politely. I'm glad you know what the word is...I try not to comment on grammar and spelling but that one just bugged me. :)
n/t
You sure do make a lot of assumptions. I've been running my own systems and networks for over 20 years. I've done my own time as help desk support. I think I'm capable of figuring out if the single computer connected to this DSL connection, assembled and installed and maintained by myself, is running P2P apps. Guess what? It's not.
You expect me to rate-limit my upstream and downstream bandwidth to 60% of rated? That's ludicrous and totally unnecessary. Even if it were to fix the latency, that's an incredibly stupid "solution." I guess instead of fixing my car's unbalanced wheels, I'll just drive 30mph on the interstate.
Get off your high horse. Stop assuming you know me and my equipment better than I do. Either be helpful or be quiet.
As a dual-booter, I'm curious, what's the performance like under CXG compared to running in Windows? I'd like to not have to dual-boot.
Why wouldn't they store the data? What incentive do they have to delete it? Do you have any idea how much storage capacity Google has? The fact is that there shouldn't have been any privacy-sensitive data in there, because it was captured from in-the-clear transmissions.
Do you suggest that it ought to be illegal to stand on the sidewalk and look through a window in a building, or listen to the sounds coming out of a building? Because that's no different than receiving in-the-clear wireless network transmissions: it's just another kind of wireless transmission.
If Google had cracked encrypted data, that'd be a completely different matter.
If you're arguing that Google should have been more proactive to avoid capturing anything but the SSIDs, in the off chance that some other data might have been captured which wasn't intended for in-the-clear transmission, I can agree with that. But there's no way it should be required by law, and there's no way it should be a crime to not do so. The onus should be on those transmitting to ensure that they aren't transmitting sensitive data in the clear. If you say Google should be more responsible for their actions, the same principle applies to those running wireless networks: they also should exercise due diligence, and Google shouldn't be faulted for the transmitters' negligence.
Maybe you are allowed to do that. Have you checked any laws in your jurisdiction? If you're remaining on your own property or on public property, I'd be surprised if it were illegal for you to point a microphone or an antenna or a camera anywhere.
It shouldn't be illegal, because if it were, it'd always be a judgement call as to whether someone was being malicious, and innocent people could be hauled in by people who didn't like them for doing something that "looked bad." Oh, wait, that happens to people with cameras all the time since 9/11...
IMO it boils down to this: if you don't want people seeing into your house, close the curtains. If you don't want people hearing what goes in on your house, close the windows and doors, install soundproofing if necessary, or just don't talk so loudly. If you don't want people being able to listen in to your wireless network, encrypt it and change the key every so often. If a person is broadcasting his wireless network in the clear, yelling out the front door, or leaving his curtains open at night with the interior lights on, he shouldn't expect privacy.
Google didn't crack anyone's encryption or pick any locks, so I don't think there's a crime here.
Now if a government agency had done it, that'd be different...oh wait, I forgot about ECHELON, et al.
This whole argument against Google is like telling a neighbor to close their eyes when they look out their window which happens to line up with your window so that you can leave your curtains open and they won't see into your house. What you should do is close your curtains when you want privacy. Now if your neighbor sneaks over and puts a fiber-optic camera through the crack around the window frame, then you have a valid complaint.
As devices become "smarter" and more connected, these kinds of flaws and vulnerabilities will only increase in number and severity. It's highly unlikely that there will ever be enough economic incentive for manufacturers to keep the embedded software in their consumer devices secure and up-to-date, not to mention the lack of software update mechanisms.
This is why we need Free Software. Standard platforms running Free Software can be patched and updated simply and easily, and maintained by community efforts. Once the build system is in place for a device or a platform of devices, simple patches can be pushed out with little effort, regardless of a manufacturer's continued interest.
No consumer should be forced to choose between having a device that he can't trust or buying a newer device which has fixed software. That's not a valid reason to buy a new device. But, of course, the manufacturers and vendors would love for folks to buy new TVs every few years, even if the old ones are fine. "Security bugs? It's obsolete, friend. You need to buy a new set." To them it may be like another form of planned obsolescence.
Still an AC I see.
That's my point exactly: the problem is about TCP, which does not "request packets." The "request" is made at a higher level, in HTTP.
You can throw around terminology like "OSI model" all you want, but you have demonstrated that you still don't understand how buffering or TCP connections work, and why bufferbloat is a real problem.
I disagree. If a connection's buffers were 50-100ms in size (perhaps even less than 50ms would be enough sometimes), the latency could never exceed 50-100ms, which is plenty for most interactive protocols. Maybe you're still thinking of buffers in terms of bytes, but it's much more helpful in this case to consider them in terms of the time it takes to empty them.
Did you read my comment? I said that downloading a single file via HTTP ends up filling the buffers and causing the latency. The way you "remove" what's in the buffer is to stop the download. That's the whole point: the latency only becomes a problem when the connection is being used. Your solution is to not use the connection. That's like a doctor telling a patient with a broken arm to just not use his arm anymore, instead of fixing the broken arm. I guess I'll just cancel my DSL and not use the Internet anymore.
Oh, you're right, it was my imaginary little brother using Kazaa. Why didn't I think of that? How silly of me....
I don't understand why you think he made up "bufferbloat" (the problem, that is, not the term). He clearly explained his methods and his work with other parties who are indeed experts. I've used Netalyzr on my own connection and it reports very large buffers on my DSL connection, and I've observed the exact same symptoms that Mr. Gettys has observed. Other people commenting on this article (even some who work in ISPs) have concurred.
If you want to convince me that bufferbloat is not a real problem, and not the cause of the observed symptoms, you'll need to provide some evidence of alternative explanations, because Mr. Gettys has done an excellent job of proving his case. He's also not the only person who's written about the problem, only the most recent.
If you read the series of articles and understand enough about how TCP works, it's clear that it is indeed a real problem.
I used to have a problem with DSL when it rained at another place I lived. But AT&T tried to charge me for the technician's visit. Any advice for getting them to fix their lines without charging me? :)