In sweden there is at least one powercompany (sydkraft) that offers broadband via the powerlines. You can get 0.5, 1 or 2 Mbps depending on what you're willing to pay and depending on where you live you get connected either via adsl, cable-tv, fibre or powerlines.:)
And non-US ISP's would start charging for data carried into US networks, noone would be willing to pay for all the data in DDoS attacks and the Slashdot effect, and the internet would fall apart. =) Not to mention that the infrastructure for charging every little ISP out there for the amount of data sent to them just isn't there and would be one hell of a project to set up. What would you do if a certain ISP won't pay you? Set up a firewall to stop packets from being sent onto the internet towards their subnet? Your customers would probably switch to a ISP who doesn't cut off parts of the internet. =) If you stop sending data to a certain subnet, nothing on that subnet can send data to you either. Well... Some data, but most protocols require two way traffic. And if a whole country refuses to implement this system? Would you cut off that entire country? You would start having segments of the internet that wouldn't be accessible from the US, and vice versa.
On the other hand, many ISP's *do* set a cap for how much data you can send/recieve in a month. If you exceed that amount, your connection will drop in speed or may even be cut of completely for the rest of the month.
And what will they call the product? They can't use the word "window" in the name, they'll get sued by microsoft! Or is it just "windows", with an s that ms got a trademark for? =/
But it would be great to have a "saver" station to switch to when the regular stations get too lousy, wouldn't it? =)
Where I live I've got two such stations. (Actually, I rarely switch *from* them at all, unless I'm listening to my mp3player.) One is a station funded by the state and covers the entire country. No commercials and therefore not *only* lousy programs playing *only* list music. The other is a student radiochannel, funded with commercials and sponsoring from local companies or companies who wants to hire the students once they become engineers, sent from the university in the neighbouring city. It's *very* rare for all of them to suck at the same time, which I can't say for the big commercial "music" channels that exist here.
But I guess things like a student-run radiostation would have a hard time getting air time on a countrywide broadcastsystem with only 100 channels in a country the size of the us.:-/ Especially since most of the students who go on the radio, at least here, are people with alternative tastes that hate the stupid commercial stations. Like people listening to goth, black rock, synth, ebm, trance, hardcore, etc. =) I don't really care about locality, even though it *is* nice to get tips about conserts and such. I want a station that plays mostly good music and small unknown bands that I'd never hear about otherwise, instead of just playing what's high on the charts right now.
The difference would be that you *double* the amount of raw proccessing power. Then you'd probably have ht on them too, so that you'd have one physical cpu, seen as two by any smp os, and then you'd hyperthread them so that you get *four* virtual cpu's if you activate ht too. You'll probably see a much bigger performance boost by putting two separate cores in one package than ht'ing one core. And if you put two or more of these in one box, well do the math. =)
I haven't actually read the entire articles so I don't know what resources the two cores would share, but it must be less than what is shared when doing hyperthreading. The major drawback is probably that the two cores would need to share the same data and adress bus to the main memory of your machine.
Well free is a bit of an overstatement, I'd say =) Don't you pay for every sms you send? And a high price at that!? For a service that never was ment to be reliable to start with and only uses surplus bandwidth in the mobile networks you'd expect to pay a very low price per message. That's why your messages sometimes takes ages to deliver. If there is no surplus bandwith in the controll protocol when your message gets to the top of the queue, it gets placed at the bottom of the queue again. =) If you're in an area with lots of mobiles changeing cells all the time, like in a city, there might be very little bandwidth left over for sms. At least, as far as I know, it used to be this way. Maybe they've changed the way things are handled, since reliability and speed seem to have increased lately.
But I always activate the "delivery reciept" funktion. That way I *know* when my message has been delivered. I started doing this after having an important message being deliverd more than 12 hours after me sending it. If I don't get a delivery reciept within a reasonable time, I can always make a phonecall instead if it's important and timecritical. =)
Well, the *article* says that you get more performance if you patch the kernel to be optimized for ht. =) It also says that you get a performance boost ever by using the standard smp kernel.
I wonder... Would a NDA be legally binding for something as illegal as creating a worm that "hacks" itself onto peoples computers? Wouldn't the one approched with a deal like that be obliged by law to report it to the police? If someone asks me to do something illegal in exchange for money, am I breaking the law if I don't report it? Even if I turn the offer down? =/
Well... As has been posted here a thousand time allready: The boosters are really, *really* dirty. (That's what they're talking about. Booster replacements.) You do *not* want to get that cloud blown towards you when they launch the shuttle. And they can not be stopped! Once lit, they burn until the fuel is gone.
This new booster would give of something as "clean" as a couple of thousand cars running for a while, would be possible to stop in case of an emergency and would cost a less per launch. (At least if the article is to be believed)
I think those things would be more important then to bounce around on the lunar surface for a couple of hours or to increase the IIS budget with a few percent.
But do you saturate the line with pings? You probably send your 10000 ping at a more moderate speed. Otherwise your test would end in a few milliseconds. What I'm talking about is of course to stop routing pings when they exceed a certain ping/sec for a period of time. Noone *needs* to send 100Mbit/sec pings for several hours.
And I don't assume all DDoS'es are pings. But some are. And stopping some are better than stopping none. What is really needed though is a new set of protocols that's designed for a new where idiots and bad guys are present. Allmost every aspect of the internet structure is designed from the bottom up to be used only by nice people.
Most trojans used in DOS attacks starts *flooding* pings (if it's a ping attack that is) at a given adress. If you get your trojan to 8 million boxes, you'll have 8 million boxes flooding their closest router with pings with that specified destination adress. If that router can see that "Oh, this user is sending me 1Mbit/second pings" and after half a second dicides to stop forwarding those pings, you'll stop those packages from even *starting* their journey. The attacked box will be DOS'ed for a second instead for several hours...
This has actually happened to a company i used to work for. We got complaints that the internet was slow. I got a sniffer and checked the traffic on our internet link and found out that one of our servers where sending 100Mbit/second pings at the router. It turned out that someone had hacked the server and installed a program that at a given signal would start DOS'ing a server... All those packages, or at least the 4Mbit/second that our ISP's router was capped at, got onto the internet and probably did some damaged at the recieving end. If our ISP's router had done what I suggested, that wouln't have been the case.
And there is no ID system in sight. Just a small extra check to see if the package the router just checked the recieving IP of is a ping or not.
Well... Since the router allready looks at the ip before decideing where to send the package, there is no extra lookup to be done. Yes, you *would* need to check into the package to see if it's a ping and store a table of pings per second per ip. But you could use a simple counter. You wouldn't need to actually store every package. Look at the package, "oh, it's a ping", check if ip is in list, increase counter, if the counter show a abnormal amount of pings in a short period, don't send it towards it's destination. Then you'd go through the table at regular intervals and remove ip's that haven't increased since the last check... The crucial part here is cpu power. You need to be able to look into packages without slowing down the routing. But many of todays more powerful routers are allready capable of looking into packages at line speed. Especially if you do it close to the source. You could settle for doing it at the ISP connection routers, the one closest to the subscriber. Those routers rarely need to deal with a multitude of 1 or 10Gbit lines, or at least most are capped at the bandwith the customer subscribes for. That way you wouldn't need to burden the more central routers with this.
Other DOS attacks might be harder to fight though... A really busy web-server configured to reverse dns-lookups of every connection might actually produce a frightening amount of dns-lookups per second. =/ But it should be possible to recognise abnormal traffic of dns, ssh and other protocols too...
What one *should* do is to configure backbone routers to not allow more than a cerain amount of ping per second... Noone has a legitimate need for streaming several hundereds or thousands pings per second... Or at least put a lid on it when someone starts sending lots of pings for more than a couple of seconds...
I thought an undefinded human being would be refered to as "she".:/ As in "If a human drinks too much, she will likely experience a hangover". But then, I'm no expert in the English langugage. =)
You buy two, or more, drives and put them in a raid. I'll double the cost for you storage, but loosing 160GB in one shot is *damn* painful, even if most of it can be found again on your audio-cd's and the internet. Think about how much time you spent getting all that data and I think you'll start thinking the extra $/MB is worth it.:)
Database? =) Who'd want to run a database on such drive, indeed? I want *four* of them to put in a raid5 for my Anime! I've allready filled 320 GB... Need more space! I don't need the small performance edge of scsi vs ide. I don't even need 7600rpm drives! I just need it to be able to read 8 - 9 MB/s so that I can saturate my 100Mbit network at home.
But they're still too expensive.:-/ I want my whole raid setup to cost about 2 - 300 Euro.
Yeah, and 1985 noone would *never ever* need a 1 GB drive. What could you possibly use 1 *GIG* for? Even 100MB would be ridiculously large for a home user!
I'll bet that when you try to rip your superhighresolution movie, Terminator 7, you'll be glad you had those 500GB free on you 10TB drive.
The thing is this: Would they be able to sell this licence at such a low price if they didn't have the proffits from Windows and Office? If I'm not misstaken, and I might be since I've only read it on the net, these two products are the only ones that MS makes money on. All other products, are sold at lower than competition prices because they've got a virtual monopoly in those two areas. And if I've got the US law correctly, doing that is illegal, right?:-/
I run a small TINI-webserver at home that mail it's logs to me when they grow larger than 16kb. (It's a *very* small computer) I get av few log's per hour, mostly filled with attempts to exploit known IIS exploits.:-) And *most* originate from my ISP's range of ip's:-/ Guess I've got quite a lot of lame scriptskiddies on my network. My guess this is the picture where ever you connect to the internet via broadband.
And my firewall log's at least as many portscans per hour. Haven't these idiots got anything better to do? =/
I use linux on my fileserver because I can do things I wouldn't be able to do with a MS system, like having a webinterface for my burner and getting good performance out of my old pentium.
I have swithed to linux on my workstation 3 or 4 times, with a dual boot into windows for things that just won't work otherwise. Every time I've ended up using windows 99.9% of the time. Why? I *can't* sync my calendar and adressbook from my phone in *any* other way than using Windows and Outlook. I *can't* use the software for my synths in any other OS. I probably can run my sequencer in Linux using Wine, but that would require me to do some serious configuring. (No, I do *not* want to use another sequencer. It's an interface issue. I like my sequencer. I hate the interface of Jazz++) And Linux doesn't have drivers for my MIDI interfaces. I've checked around. Nothing.
Since I usually do some sequencing, then surf the web for a bit, then do some administration of some server, continue on my song, check my mail, create some new patches for my synths, yadda yadda, I would have to spend hours every day just booting up OS's if I were to reboot into linux for the stuff that I can do in both OS's. For the time being, there isn't anyting I'm doing on my workstation that I can do in linux but can't do in windows. After a few months of never using my linux installation, I usually remove it due to lack of diskspace.
I agree with you that you *can* do a lot of things on a Gnu/Linux/XWindows system that you can't do on MS. The catch is that the same is true the other way around. And usually you have to be a Linux wizard with plenty of time on your hands to be able to set it up. Most people aren't, so they'll stick to MS or Mac where they can install the os, then their apps and then start to work. Yes, you may have to reinstall the whole system once or twice every year to keep it running but that's acceptable if I can run the apps I want.
Features are useless if it's not the features I want.
Well... In the "encounters with women" and "fewest days in gym" categories 0 must be the lowest possible times. There should be plenty of people worldwide at the number 0 and thus noone can be said to hold the record. To hold a record you should be the only one, otherwise it isn't really a record, is it?:-)
Slashdot Mirror
How long before security-patches for american software aren't allowed to be exported outside the US? =)
In sweden there is at least one powercompany (sydkraft) that offers broadband via the powerlines. :)
You can get 0.5, 1 or 2 Mbps depending on what you're willing to pay and depending on where you live you get connected either via adsl, cable-tv, fibre or powerlines.
And non-US ISP's would start charging for data carried into US networks, noone would be willing to pay for all the data in DDoS attacks and the Slashdot effect, and the internet would fall apart. =)
Not to mention that the infrastructure for charging every little ISP out there for the amount of data sent to them just isn't there and would be one hell of a project to set up.
What would you do if a certain ISP won't pay you?
Set up a firewall to stop packets from being sent onto the internet towards their subnet?
Your customers would probably switch to a ISP who doesn't cut off parts of the internet. =)
If you stop sending data to a certain subnet, nothing on that subnet can send data to you either.
Well... Some data, but most protocols require two way traffic.
And if a whole country refuses to implement this system? Would you cut off that entire country?
You would start having segments of the internet that wouldn't be accessible from the US, and vice versa.
On the other hand, many ISP's *do* set a cap for how much data you can send/recieve in a month.
If you exceed that amount, your connection will drop in speed or may even be cut of completely for the rest of the month.
And what will they call the product?
They can't use the word "window" in the name, they'll get sued by microsoft!
Or is it just "windows", with an s that ms got a trademark for? =/
But it would be great to have a "saver" station to switch to when the regular stations get too lousy, wouldn't it? =)
:-/
Where I live I've got two such stations. (Actually, I rarely switch *from* them at all, unless I'm listening to my mp3player.)
One is a station funded by the state and covers the entire country. No commercials and therefore not *only* lousy programs playing *only* list music.
The other is a student radiochannel, funded with commercials and sponsoring from local companies or companies who wants to hire the students once they become engineers, sent from the university in the neighbouring city.
It's *very* rare for all of them to suck at the same time, which I can't say for the big commercial "music" channels that exist here.
But I guess things like a student-run radiostation would have a hard time getting air time on a countrywide broadcastsystem with only 100 channels in a country the size of the us.
Especially since most of the students who go on the radio, at least here, are people with alternative tastes that hate the stupid commercial stations. Like people listening to goth, black rock, synth, ebm, trance, hardcore, etc. =)
I don't really care about locality, even though it *is* nice to get tips about conserts and such.
I want a station that plays mostly good music and small unknown bands that I'd never hear about otherwise, instead of just playing what's high on the charts right now.
The difference would be that you *double* the amount of raw proccessing power.
Then you'd probably have ht on them too, so that you'd have one physical cpu, seen as two by any smp os, and then you'd hyperthread them so that you get *four* virtual cpu's if you activate ht too.
You'll probably see a much bigger performance boost by putting two separate cores in one package than ht'ing one core.
And if you put two or more of these in one box, well do the math. =)
I haven't actually read the entire articles so I don't know what resources the two cores would share, but it must be less than what is shared when doing hyperthreading.
The major drawback is probably that the two cores would need to share the same data and adress bus to the main memory of your machine.
Well free is a bit of an overstatement, I'd say =)
Don't you pay for every sms you send? And a high price at that!?
For a service that never was ment to be reliable to start with and only uses surplus bandwidth in the mobile networks you'd expect to pay a very low price per message.
That's why your messages sometimes takes ages to deliver.
If there is no surplus bandwith in the controll protocol when your message gets to the top of the queue, it gets placed at the bottom of the queue again. =)
If you're in an area with lots of mobiles changeing cells all the time, like in a city, there might be very little bandwidth left over for sms.
At least, as far as I know, it used to be this way. Maybe they've changed the way things are handled, since reliability and speed seem to have increased lately.
But I always activate the "delivery reciept" funktion.
That way I *know* when my message has been delivered.
I started doing this after having an important message being deliverd more than 12 hours after me sending it.
If I don't get a delivery reciept within a reasonable time, I can always make a phonecall instead if it's important and timecritical. =)
You really need to remove that lobbying system of yours...
You shouldn't be able to legaly buy politicians oppinions in a democracy.
Well, the *article* says that you get more performance if you patch the kernel to be optimized for ht. =)
It also says that you get a performance boost ever by using the standard smp kernel.
I wonder...
Would a NDA be legally binding for something as illegal as creating a worm that "hacks" itself onto peoples computers?
Wouldn't the one approched with a deal like that be obliged by law to report it to the police?
If someone asks me to do something illegal in exchange for money, am I breaking the law if I don't report it? Even if I turn the offer down? =/
Well... As has been posted here a thousand time allready:
The boosters are really, *really* dirty. (That's what they're talking about. Booster replacements.)
You do *not* want to get that cloud blown towards you when they launch the shuttle.
And they can not be stopped!
Once lit, they burn until the fuel is gone.
This new booster would give of something as "clean" as a couple of thousand cars running for a while, would be possible to stop in case of an emergency and would cost a less per launch.
(At least if the article is to be believed)
I think those things would be more important then to bounce around on the lunar surface for a couple of hours or to increase the IIS budget with a few percent.
But do you saturate the line with pings?
You probably send your 10000 ping at a more moderate speed.
Otherwise your test would end in a few milliseconds.
What I'm talking about is of course to stop routing pings when they exceed a certain ping/sec for a period of time.
Noone *needs* to send 100Mbit/sec pings for several hours.
And I don't assume all DDoS'es are pings. But some are. And stopping some are better than stopping none.
What is really needed though is a new set of protocols that's designed for a new where idiots and bad guys are present.
Allmost every aspect of the internet structure is designed from the bottom up to be used only by nice people.
Most trojans used in DOS attacks starts *flooding* pings (if it's a ping attack that is) at a given adress.
If you get your trojan to 8 million boxes, you'll have 8 million boxes flooding their closest router with pings with that specified destination adress.
If that router can see that "Oh, this user is sending me 1Mbit/second pings" and after half a second dicides to stop forwarding those pings, you'll stop those packages from even *starting* their journey.
The attacked box will be DOS'ed for a second instead for several hours...
This has actually happened to a company i used to work for.
We got complaints that the internet was slow.
I got a sniffer and checked the traffic on our internet link and found out that one of our servers where sending 100Mbit/second pings at the router.
It turned out that someone had hacked the server and installed a program that at a given signal would start DOS'ing a server...
All those packages, or at least the 4Mbit/second that our ISP's router was capped at, got onto the internet and probably did some damaged at the recieving end.
If our ISP's router had done what I suggested, that wouln't have been the case.
And there is no ID system in sight. Just a small extra check to see if the package the router just checked the recieving IP of is a ping or not.
Well... Since the router allready looks at the ip before decideing where to send the package, there is no extra lookup to be done.
Yes, you *would* need to check into the package to see if it's a ping and store a table of pings per second per ip.
But you could use a simple counter. You wouldn't need to actually store every package.
Look at the package, "oh, it's a ping", check if ip is in list, increase counter, if the counter show a abnormal amount of pings in a short period, don't send it towards it's destination.
Then you'd go through the table at regular intervals and remove ip's that haven't increased since the last check...
The crucial part here is cpu power.
You need to be able to look into packages without slowing down the routing.
But many of todays more powerful routers are allready capable of looking into packages at line speed.
Especially if you do it close to the source.
You could settle for doing it at the ISP connection routers, the one closest to the subscriber.
Those routers rarely need to deal with a multitude of 1 or 10Gbit lines, or at least most are capped at the bandwith the customer subscribes for.
That way you wouldn't need to burden the more central routers with this.
Other DOS attacks might be harder to fight though...
A really busy web-server configured to reverse dns-lookups of every connection might actually produce a frightening amount of dns-lookups per second. =/
But it should be possible to recognise abnormal traffic of dns, ssh and other protocols too...
Hihi... ;)
Sorry. My bad. =)
Well... I failed my last math exam, that's probably the reason.
What one *should* do is to configure backbone routers to not allow more than a cerain amount of ping per second...
Noone has a legitimate need for streaming several hundereds or thousands pings per second...
Or at least put a lid on it when someone starts sending lots of pings for more than a couple of seconds...
I thought an undefinded human being would be refered to as "she". :/
As in "If a human drinks too much, she will likely experience a hangover".
But then, I'm no expert in the English langugage. =)
You buy two, or more, drives and put them in a raid. :)
I'll double the cost for you storage, but loosing 160GB in one shot is *damn* painful, even if most of it can be found again on your audio-cd's and the internet.
Think about how much time you spent getting all that data and I think you'll start thinking the extra $/MB is worth it.
Database? =)
:-/
Who'd want to run a database on such drive, indeed?
I want *four* of them to put in a raid5 for my Anime! I've allready filled 320 GB... Need more space!
I don't need the small performance edge of scsi vs ide. I don't even need 7600rpm drives! I just need it to be able to read 8 - 9 MB/s so that I can saturate my 100Mbit network at home.
But they're still too expensive.
I want my whole raid setup to cost about 2 - 300 Euro.
Well... I'll bet my post i redundant, but wth...
Yeah, and 1985 noone would *never ever* need a 1 GB drive.
What could you possibly use 1 *GIG* for?
Even 100MB would be ridiculously large for a home user!
I'll bet that when you try to rip your superhighresolution movie, Terminator 7, you'll be glad you had those 500GB free on you 10TB drive.
The thing is this: Would they be able to sell this licence at such a low price if they didn't have the proffits from Windows and Office? :-/
If I'm not misstaken, and I might be since I've only read it on the net, these two products are the only ones that MS makes money on. All other products, are sold at lower than competition prices because they've got a virtual monopoly in those two areas.
And if I've got the US law correctly, doing that is illegal, right?
Does the cost include the ingredients? =)
I run a small TINI-webserver at home that mail it's logs to me when they grow larger than 16kb. :-) :-/
(It's a *very* small computer)
I get av few log's per hour, mostly filled with attempts to exploit known IIS exploits.
And *most* originate from my ISP's range of ip's
Guess I've got quite a lot of lame scriptskiddies on my network.
My guess this is the picture where ever you connect to the internet via broadband.
And my firewall log's at least as many portscans per hour.
Haven't these idiots got anything better to do? =/
I use linux on my fileserver because I can do things I wouldn't be able to do with a MS system, like having a webinterface for my burner and getting good performance out of my old pentium.
I have swithed to linux on my workstation 3 or 4 times, with a dual boot into windows for things that just won't work otherwise.
Every time I've ended up using windows 99.9% of the time.
Why?
I *can't* sync my calendar and adressbook from my phone in *any* other way than using Windows and Outlook.
I *can't* use the software for my synths in any other OS.
I probably can run my sequencer in Linux using Wine, but that would require me to do some serious configuring. (No, I do *not* want to use another sequencer. It's an interface issue. I like my sequencer. I hate the interface of Jazz++)
And Linux doesn't have drivers for my MIDI interfaces. I've checked around. Nothing.
Since I usually do some sequencing, then surf the web for a bit, then do some administration of some server, continue on my song, check my mail, create some new patches for my synths, yadda yadda, I would have to spend hours every day just booting up OS's if I were to reboot into linux for the stuff that I can do in both OS's.
For the time being, there isn't anyting I'm doing on my workstation that I can do in linux but can't do in windows.
After a few months of never using my linux installation, I usually remove it due to lack of diskspace.
I agree with you that you *can* do a lot of things on a Gnu/Linux/XWindows system that you can't do on MS. The catch is that the same is true the other way around.
And usually you have to be a Linux wizard with plenty of time on your hands to be able to set it up.
Most people aren't, so they'll stick to MS or Mac where they can install the os, then their apps and then start to work.
Yes, you may have to reinstall the whole system once or twice every year to keep it running but that's acceptable if I can run the apps I want.
Features are useless if it's not the features I want.
Well... In the "encounters with women" and "fewest days in gym" categories 0 must be the lowest possible times. :-)
There should be plenty of people worldwide at the number 0 and thus noone can be said to hold the record.
To hold a record you should be the only one, otherwise it isn't really a record, is it?