This is a crap post, sorry. There is no modern OS that can mitigate an application that is bad. The only thing that can do that is a good operator (which re-nices the forking process or better yet, fixes the bug that is causing it or adds new capacity).
Solaris won't ride its golden winged horse down out of the heavens to save you from this kind of problem, trust me.
Windows 7 easily has the most confusing, difficult to configure network properties of any Windows. Granted, I like how it differentiates between "new" network connections as far as work, public, home for the purposes of firewall config, but it's BRUTAL to actually configure the network properties otherwise. All the obfuscation gets in your way and makes your teeth grind.
You can't get thousands of users on a single base station. 200 is really, really stretching it. If you have 25Mhz of bandwidth with your license, that typically means a top of four base stations per tower, and you have to be very careful how you place your towers and aim your antennas so that you don't self-interfere with your other towers.
Redline claims 500 users per base station and Solectek claims 1024. Both total crap. 500 users @ 64Kbps or 1024 users at 28.8Kbps really is useless in any real-world application except perhaps SCADA. But given that the range is 10 miles, that would be a hell of a lot of municipal equipment to monitor.
Time spacing is different. However, don't confuse "AIR SPEED" with "ACTUAL SPEED." I'm talking actual speed.
AIR SPEED is what the manufacturers like to use to describe their product, and there is a roughly 50% overhead for error correction, signalling and encryption.
I own an ISP (www.kos.net) which does rural broadband, and we've been very successful at that. Sometimes performance isn't what people expect, but that's to be expected... just not by those people who want unlimited fibre to their cottage.;)
WiMax has been shouted from mountaintop to mountaintop for the past six years. And it isn't really being deployed much. Why?
Let me tell you.
First of all, WiMax has a range of 10 miles. That's not so bad but in rural installations, it's just barely enough.
Secondly in many places, WiMax has a max emitted power of 43dbm. In others it has up to 58dbm. Either way, it's not that much power. Further, the wattage is about 15 watts for the CPE equipment and about 200 watts for the base station - but the base station cannot emit 200 watts to a single transmitter by law, so that's just a red herring unless you plan to build out a base station with multiple antennas (good idea for the first build). Many (Redline, Alvarion, Aperto) use 2.3 watts.
Third, line of sight is a major requirement for WiMax. 802.16d or 802.16e. It is possible to get some non line of sight connections at close range (2 miles) if the conditions are right, but in the end it's a high speed wave (2.3Ghz for the Clearwire + Inukshuk early adopters, 3.4-3.7Ghz for the later adopters like my company). High speed waves have a much higher chance of hitting a particle and stopping than lower speed waves. 400Mhz used for voice and cell data has a much better NLoS capability and 900Mhz radios we use for rural broadband are also quite good, especially at low power levels they are allowed. WiMax does, however, have pretty good NEAR line of sight capabilities, we're finding.
Fourth, WiMax standards aren't. Rarely will one device interoperate with another vendor's equipment. So those WiMax chips in your Intel laptop? Junk.
Fifth, performance of WiMax isn't as good as 5.8Ghz access points. That's right. WiMax uses a 5, 7 or 10Mhz channel and while 10Mhz has slightly better throughput, you're not going to see much of that because of antenna spacing and distance characteristics. It's "54Mbps" rating per base antenna ($5,000 U.S. per + antenna) works out to 23Mbps aggregate at a 50% RX/TX spacing, which means 10.5Mbps in either direction.... although some companies are talking about a 75TX/25RX percentage split in upcoming firmware. In any case, it's not a panacea, and yes that bandwidth is shared between up to 200 people per base station.
Sixth, WiMax gear is really expensive. Everything about it is expensive, from base stations to subscriber modules, to tower placement, to purchasing licenses for transmitting.
Seventh, because it's licensed and it is a high-speed wave, it's mostly useless except to rural customers. Not to mention that nobody wants 15 watts of emitted power on their crotch.
Eigth, you will need to register your transmitter with Industry Canada or the FCC in the States. Maybe both. Not only is this fairly complex to do unless you're a service provider, you may find yourself having to bid on spectrum or with the 3.65Ghz band, you may be told that another transmitter is too near to you and you'll have to coordinate with that operator.
So. That's WiMax. It's not much good to most people, it has limited abilities to provide rural service (only better than 2.4Ghz WiFi because it's got more output power and has a licensed clean channel), it's really expensive, it's not fast enough and it's got complex licensing.
Why are we using it? We are desperately in need of another frequency, since we have filled many of our 900Mhz radios and the 5.8Ghz radios are not good for rural use at all (no near or non line of sight ability). 2.4Ghz is a dirty frequency with a lot of operators and power-company "smart meters" in it (that's a stupidity of a whole other level that needs a whole other discussion). So, it's WiMax.
BTW, those thinking city-wide WiFi is cool should do a little study on why it's not workable. I
With customers expecting to pay less and less for more and more, how do you expect to buy more bandwidth... for less money?
It's nowhere near so cut and dried. Pricing has been where it's at because everyone has been trying to kill or prevent competition. Now that the status quo has been established, they have to either cut costs or find more profitability, because up to now nobody has made money on broadband Internet.
Because you never pay for it, the price has a built-in over-subscription requirement. Dedicated bandwidth costs a lot more. Go price a DS-3 and see.
What you're saying is a little like saying you want to use the whole road for yourself at the maximum rate possible. After all, your taxes pay for your access to it.
Actually there are quite a few features found in Linux not found in FreeBSD or PC-BSD. Like write access to NTFS, inotify, many journalled file systems, user-space FS, multiple kernel virtualization methods and client optimizations, stable LVM, et al
And how about those forklift upgrades that FreeBSD seems to love... no "apt-get dist-upgrade" foolishness for BSD. Oh no... it's time to format!
Why most people don't standardize on APT, though... it's beyond me. IMHO they should all talk to the Debian people and build a hardened core OS that they can build their management systems etc. on top of in order to differentiate themselves.
ZFS is really not that impressive. Only if you are building systems with giant disk subsystems - which is exactly where you will be buying SAN at this point instead. We just bought a HP 5000 series SAN with 138 disk drives in it... does self-healing and provides up to 1.4TB/s throughput, doesn't use ZFS. ZOMG!
Yet, I have been using it nearly exclusively for 10 years and watched it progress as a result of some of these policies. Very seldom have I had an issue caused by the OS.
Re:Ummm...Karma to Burn.
on
Linux 2.6.26 Out
·
· Score: 3, Insightful
Yes, you're going to get flamed, and for good reason. Just how easy do you think it is to support chipsets from manufacturers who supply no documentation, who load their firmware from their drivers, and who threaten to sue anyone who tries to do it on their own?
And so, yeah, maybe YOU should BECOME a developer.
Slashdot Mirror
This is a crap post, sorry. There is no modern OS that can mitigate an application that is bad. The only thing that can do that is a good operator (which re-nices the forking process or better yet, fixes the bug that is causing it or adds new capacity).
Solaris won't ride its golden winged horse down out of the heavens to save you from this kind of problem, trust me.
Do you know how to tell a french Canadian from a distance?
I don't know either, but you can't tell one fuck-all from up close.
Yes, they are. Demonstrably. They would have long ago forced ISPs to filter if they weren't.
Windows 7 easily has the most confusing, difficult to configure network properties of any Windows. Granted, I like how it differentiates between "new" network connections as far as work, public, home for the purposes of firewall config, but it's BRUTAL to actually configure the network properties otherwise. All the obfuscation gets in your way and makes your teeth grind.
FYI, while they "own" the infrastructure, they didn't pay for it. Your tax dollars did.
And they use up BILLIONS of dollars per year worth of free right-of-way that only they have access to.
If I had upvotes, I'd upvote this to the moon. This is exactly correct.
You can't get thousands of users on a single base station. 200 is really, really stretching it. If you have 25Mhz of bandwidth with your license, that typically means a top of four base stations per tower, and you have to be very careful how you place your towers and aim your antennas so that you don't self-interfere with your other towers.
Redline claims 500 users per base station and Solectek claims 1024. Both total crap. 500 users @ 64Kbps or 1024 users at 28.8Kbps really is useless in any real-world application except perhaps SCADA. But given that the range is 10 miles, that would be a hell of a lot of municipal equipment to monitor.
Time spacing is different. However, don't confuse "AIR SPEED" with "ACTUAL SPEED." I'm talking actual speed.
AIR SPEED is what the manufacturers like to use to describe their product, and there is a roughly 50% overhead for error correction, signalling and encryption.
802.11g is much the same.
I own an ISP (www.kos.net) which does rural broadband, and we've been very successful at that. Sometimes performance isn't what people expect, but that's to be expected... just not by those people who want unlimited fibre to their cottage. ;)
WiMax has been shouted from mountaintop to mountaintop for the past six years. And it isn't really being deployed much. Why?
Let me tell you.
First of all, WiMax has a range of 10 miles. That's not so bad but in rural installations, it's just barely enough.
Secondly in many places, WiMax has a max emitted power of 43dbm. In others it has up to 58dbm. Either way, it's not that much power. Further, the wattage is about 15 watts for the CPE equipment and about 200 watts for the base station - but the base station cannot emit 200 watts to a single transmitter by law, so that's just a red herring unless you plan to build out a base station with multiple antennas (good idea for the first build). Many (Redline, Alvarion, Aperto) use 2.3 watts.
Third, line of sight is a major requirement for WiMax. 802.16d or 802.16e. It is possible to get some non line of sight connections at close range (2 miles) if the conditions are right, but in the end it's a high speed wave (2.3Ghz for the Clearwire + Inukshuk early adopters, 3.4-3.7Ghz for the later adopters like my company). High speed waves have a much higher chance of hitting a particle and stopping than lower speed waves. 400Mhz used for voice and cell data has a much better NLoS capability and 900Mhz radios we use for rural broadband are also quite good, especially at low power levels they are allowed. WiMax does, however, have pretty good NEAR line of sight capabilities, we're finding.
Fourth, WiMax standards aren't. Rarely will one device interoperate with another vendor's equipment. So those WiMax chips in your Intel laptop? Junk.
Fifth, performance of WiMax isn't as good as 5.8Ghz access points. That's right. WiMax uses a 5, 7 or 10Mhz channel and while 10Mhz has slightly better throughput, you're not going to see much of that because of antenna spacing and distance characteristics. It's "54Mbps" rating per base antenna ($5,000 U.S. per + antenna) works out to 23Mbps aggregate at a 50% RX/TX spacing, which means 10.5Mbps in either direction.... although some companies are talking about a 75TX/25RX percentage split in upcoming firmware. In any case, it's not a panacea, and yes that bandwidth is shared between up to 200 people per base station.
Sixth, WiMax gear is really expensive. Everything about it is expensive, from base stations to subscriber modules, to tower placement, to purchasing licenses for transmitting.
Seventh, because it's licensed and it is a high-speed wave, it's mostly useless except to rural customers. Not to mention that nobody wants 15 watts of emitted power on their crotch.
Eigth, you will need to register your transmitter with Industry Canada or the FCC in the States. Maybe both. Not only is this fairly complex to do unless you're a service provider, you may find yourself having to bid on spectrum or with the 3.65Ghz band, you may be told that another transmitter is too near to you and you'll have to coordinate with that operator.
So. That's WiMax. It's not much good to most people, it has limited abilities to provide rural service (only better than 2.4Ghz WiFi because it's got more output power and has a licensed clean channel), it's really expensive, it's not fast enough and it's got complex licensing.
Why are we using it? We are desperately in need of another frequency, since we have filled many of our 900Mhz radios and the 5.8Ghz radios are not good for rural use at all (no near or non line of sight ability). 2.4Ghz is a dirty frequency with a lot of operators and power-company "smart meters" in it (that's a stupidity of a whole other level that needs a whole other discussion). So, it's WiMax.
BTW, those thinking city-wide WiFi is cool should do a little study on why it's not workable. I
With customers expecting to pay less and less for more and more, how do you expect to buy more bandwidth... for less money?
It's nowhere near so cut and dried. Pricing has been where it's at because everyone has been trying to kill or prevent competition. Now that the status quo has been established, they have to either cut costs or find more profitability, because up to now nobody has made money on broadband Internet.
Use Foxit! Reader on Windows and something else on other operating systems, such as Okular.
Because you never pay for it, the price has a built-in over-subscription requirement. Dedicated bandwidth costs a lot more. Go price a DS-3 and see.
What you're saying is a little like saying you want to use the whole road for yourself at the maximum rate possible. After all, your taxes pay for your access to it.
This is probably the first sane statement in this article yet. :) Thanks for that.
So, you're basically a Solaris fanboi. Got it.
What kind of crap are you talking about here. GCC bugs? They're common to ALL operating systems using GCC - including OpenSolaris.
Security issues have been either few and far between - or in applications running on TOP of linux, which is a whole other subject.
If you had a clue, you'd eat it.
Actually there are quite a few features found in Linux not found in FreeBSD or PC-BSD. Like write access to NTFS, inotify, many journalled file systems, user-space FS, multiple kernel virtualization methods and client optimizations, stable LVM, et al
And how about those forklift upgrades that FreeBSD seems to love... no "apt-get dist-upgrade" foolishness for BSD. Oh no... it's time to format!
The LSB solves most of this issue.
Why most people don't standardize on APT, though... it's beyond me. IMHO they should all talk to the Debian people and build a hardened core OS that they can build their management systems etc. on top of in order to differentiate themselves.
But what do I know...
ZFS is really not that impressive. Only if you are building systems with giant disk subsystems - which is exactly where you will be buying SAN at this point instead. We just bought a HP 5000 series SAN with 138 disk drives in it... does self-healing and provides up to 1.4TB/s throughput, doesn't use ZFS. ZOMG!
Yet, I have been using it nearly exclusively for 10 years and watched it progress as a result of some of these policies. Very seldom have I had an issue caused by the OS.
Easy to say. Not so easy to prove.
Yes, you're going to get flamed, and for good reason. Just how easy do you think it is to support chipsets from manufacturers who supply no documentation, who load their firmware from their drivers, and who threaten to sue anyone who tries to do it on their own?
And so, yeah, maybe YOU should BECOME a developer.
It's anything but pointless. Allow Bush to set the precedent of what a president should be and it will only get worse.
Bullshit.
He's already working on it. See the stories about DNA testing and routine police & immigration roadstops going up all over.
Bell no longer claims to offer unlimited access, and never claimed to offer a dedicated amount of bandwidth. People assume too much.