jesus. no matter how stupid the topic is, you can find an apologist for it. Here're the rebuttal points:
1) USA is the most lawsuit happy country in the world. If the company can do business in USA, and the price covers the cost of lawsuits, then the same price will cover the price of lawsuits in Australia.
2) WTF are you smoking on development taxes? *NOBODY* pays fucking taxes on writing software, only on selling it. If you are talking about salaries of developers and so on, then shouldn't there be a *DISCOUNT* since it's so much cheaper to develop it elsewhere?!
3) I guarantee you Adobe has already figured out the taxing regulations previously, and it's fucking sunk cost. The cost of figuring out the tax regulations (as a delta against US tax regulations) do not recur yearly. If there are changes, it's the same kind of changes that happen in USA, and obviously Adobe USA can handle it, so why can't Adobe Australia?
You understand that BGP is used to manage routes right? Someone like Google would have tons of peers. That means they advertise their network to those peers.
That means Comcast gets those route advertisements.
So do TimeWarner
So do Verizon.
And so on and so forth.
Now, you're in some dumbfucktown, but you're cool and hip and have redundant links to a couple of local ISPs. Guess where those ISPs get their links from? Time Warner, Verizon, etc. All of whom have a bunch of peering points...
Following how the route advertisement comes to you is like reading one of those "find-your-adventure" books that used to be popular in the 80s where you have to follow the story by jumping from page to page, based on the decision you made.
Seriously? Go look up how DNSSEC works. You understand DNSSEC also covers subdomains right? And the DNS server that covers apple.com may not cover *.security.apple.com
So you have a fucking botnet. All 10 million machines are going to try to login using one ID. That's called bruteforcing, and hotmail should have mechanisms in place to stop it. If they don't, they suck.
Cracking rzfitdv on a desktop means jackshit because you don't know what the hell is stored on hotmail servers. Whoopie doo. You still have to bruteforce the hotmail servers by trying to login with every combination known.
And your inability to comprehend the question at hand - is hardware raid better than software raid, is incomprehensible, given all the evidence I've provided.
And you still dodge a direct question being asked. And you talk about sleazy politicians? Oh, the irony.
Heh. He was at a strip bar. One of the dancers saw him working away at his laptop (yeah, I know!), and it turns out she was in his Intro to C class, and one thing led to another...:)
Remember, nothing is impossible in this universe. You're just jealous it had to happen, and it happened to him, and not you:)
Seriously? Not everything is all singing, all dancing. Look at how big bejeweled is on the iPhone vs on the palm pilot. Bejeweled worked fine on the old palm pilots as well as they do on the iphone.
In the 1970s, a UNIX server supporting multiple servers were *SLOW* by 1990s standards. But it worked well. Even across a 56k line.
I pointed out that fact, and I'm the one shifting goal post? You are the one defending hardware raid to the death. You are the one who disagree with Intel, Youtube, Google, and Linux kernel developers (last one is left as a search exercise for you).
And when you are asked specifically whether you agree with Intel/Youtube, you refuse to answer and weasel your way out. Such a dick move.
ANSWER THE GODDAMNED QUESTION: DO YOU AGREE WITH INTEL AND YOUTUBE ENGINEERS THAT SOFTWARE RAID IS AS FAST AS, OR FASTER THAN HARDWARE RAID, AT MINIMUM IMPACT (5% or less) TO THE CPU?
Seriously?! hardware raid cards? Which century are you stuck in?
Anyone with a reading comprehension will realize that I'm responding to the OP that software raid is a joke, and that hardware raid is a necessity. My point, all along, has been that software raid is now good enough to replace hardware raid.
And apparently you have reading comprehension problems. Why buy another fucking socket when you can just get a cpu that can handle 5% more load? I gave an example of the CPU already - the CPU only cost *FIVE FUCKING DOLLARS MORE* for 5% more performance. In the edge case where you are sucking up every cpu cycle for your application, then, yes, hardware raid can make sense, but I'd also say you have seriously fucked up on your sizing because any server that's running at 100% is not capable of handling more load.
You keep going over into edge cases and tiny little machines that run at 100% of the cpu. *FOR MOST PEOPLE, SOFTWARE RAID IS BETTER THAN HARDWARE RAID*.
But you're a know it all. Can't even answer straight questions. I'm done. Enjoy being the "expert".
Do you agree/admit that current generation CPUs can calculate RAID parity faster that a hardware RAID card (in most cases, with minimal, 5% overhead)?
If yes, great.
If no, are you saying the realities that the youtube engineers saw, that other people saw (linux kernel hackers, etc), that Intel publicly stated are all false?
At newegg, a Phenom II X4 965 at 3.4Ghz is $125. A Phenom II X4 975 at 3.6Ghz is $130. That 5% is $5.
At times, you cannot avoid buying cards without RAID - they simply don't make it. But as youtube discovered, even when you have a hardware RAID enabled card, going the software RAID route increased performance by 20% to 30%.
In a RAID environment, the IO bottleneck is at one of various places:
1) Disk mechanism (not under discussion) 2) Interface (not under discussion) 3) calculating RAID parity (hardware RAID vs software RAID) 4) higher level functionality (moving LUNs around, RAID hole, etc - not under discussion)
As of 2010, Intel has publicly declared that software RAID (good implementations, obviously) is faster than hardware RAID, and does not take up more than 5% of your CPU. Youtube's engineers have stated that they get 20% to 30% more performance out of software RAID than hardware RAID. Obviously *YOUR* expertise triumphs them all.
I am also worried that a guy who can't price a CPU that has 5% more performance without breaking the bank. You must be working on itty bitty machines.
Slashdot Mirror
You do realize Halo was originally a Macintosh exclusive game until Microsoft waved bundles of money under Bungie's owners right?
But what does blathering on about nothing gains the shill?
jesus. no matter how stupid the topic is, you can find an apologist for it. Here're the rebuttal points:
1) USA is the most lawsuit happy country in the world. If the company can do business in USA, and the price covers the cost of lawsuits, then the same price will cover the price of lawsuits in Australia.
2) WTF are you smoking on development taxes? *NOBODY* pays fucking taxes on writing software, only on selling it. If you are talking about salaries of developers and so on, then shouldn't there be a *DISCOUNT* since it's so much cheaper to develop it elsewhere?!
3) I guarantee you Adobe has already figured out the taxing regulations previously, and it's fucking sunk cost. The cost of figuring out the tax regulations (as a delta against US tax regulations) do not recur yearly. If there are changes, it's the same kind of changes that happen in USA, and obviously Adobe USA can handle it, so why can't Adobe Australia?
You're full of shit.
I think it's me explaining it poorly. I was trying to point out you can send more shit along than just what's behind the current router.
You understand that BGP is used to manage routes right? Someone like Google would have tons of peers. That means they advertise their network to those peers.
That means Comcast gets those route advertisements.
So do TimeWarner
So do Verizon.
And so on and so forth.
Now, you're in some dumbfucktown, but you're cool and hip and have redundant links to a couple of local ISPs. Guess where those ISPs get their links from? Time Warner, Verizon, etc. All of whom have a bunch of peering points...
Following how the route advertisement comes to you is like reading one of those "find-your-adventure" books that used to be popular in the 80s where you have to follow the story by jumping from page to page, based on the decision you made.
Seriously? Go look up how DNSSEC works. You understand DNSSEC also covers subdomains right? And the DNS server that covers apple.com may not cover *.security.apple.com
How the hell are you upvoted, I have no clue.
Do you know what the hell you're talking about?
So you have a fucking botnet. All 10 million machines are going to try to login using one ID. That's called bruteforcing, and hotmail should have mechanisms in place to stop it. If they don't, they suck.
Cracking rzfitdv on a desktop means jackshit because you don't know what the hell is stored on hotmail servers. Whoopie doo. You still have to bruteforce the hotmail servers by trying to login with every combination known.
Try to follow the conversation or stay quiet, ya
Does it matter if it is "weak" or not? Unless the hackers compromised hotmail's password file and is busily trying to crack it, it is irrelevant.
What is relevant is that hotmail is apparently open to being bruteforced. Now, *THAT* is a fail.
Initially NT was DEC Alpha only because Dave Cutler came from |D|I|G|I|T|A|L|
Yet another moron who thinks a forced upgrade = users liked or wanted the damned ribbon
I don't get it. Why does it have to heat the air up ("to produce steam") ??
Why can't it just take the air and cool it down, instead of wasting energy for heating?
Apparently you get easily confused as to who's copying whom.
Doh, how the hell did you get it on the phone in the first place? Are basic concepts really that hard to understand?
And your inability to comprehend the question at hand - is hardware raid better than software raid, is incomprehensible, given all the evidence I've provided.
And you still dodge a direct question being asked. And you talk about sleazy politicians? Oh, the irony.
Heh. He was at a strip bar. One of the dancers saw him working away at his laptop (yeah, I know!), and it turns out she was in his Intro to C class, and one thing led to another... :)
Remember, nothing is impossible in this universe. You're just jealous it had to happen, and it happened to him, and not you :)
WHAT THE FUCK ARE YOU ON? It's not the foreign country that is suing is it?
Lets make it more clear.
If he went to Thailand. Bought the books there. Brought it back to USA.
Anything illegal yet?
Once he is done with his class, he sold the book.
Now all of a sudden that is illegal?! YOU ARE FUCKING SHITTING ME RIGHT?
Hey, don't knock pointers. I had a friend who got a semester's worth of lap dances for extra tutoring in pointers... :)
Seriously? Not everything is all singing, all dancing. Look at how big bejeweled is on the iPhone vs on the palm pilot. Bejeweled worked fine on the old palm pilots as well as they do on the iphone.
In the 1970s, a UNIX server supporting multiple servers were *SLOW* by 1990s standards. But it worked well. Even across a 56k line.
I'm responding to someone else.
You responded to me.
I pointed out that fact, and I'm the one shifting goal post? You are the one defending hardware raid to the death. You are the one who disagree with Intel, Youtube, Google, and Linux kernel developers (last one is left as a search exercise for you).
And when you are asked specifically whether you agree with Intel/Youtube, you refuse to answer and weasel your way out. Such a dick move.
ANSWER THE GODDAMNED QUESTION: DO YOU AGREE WITH INTEL AND YOUTUBE ENGINEERS THAT SOFTWARE RAID IS AS FAST AS, OR FASTER THAN HARDWARE RAID, AT MINIMUM IMPACT (5% or less) TO THE CPU?
http://slashdot.org/comments.pl?sid=2763753&cid=39563051 stated:
Also: HARDWARE RAID CARDS.
I can't stress that enough. software and semi-software raid is a joke.
In http://slashdot.org/comments.pl?sid=2763753&cid=39566051 I said:
Seriously?! hardware raid cards? Which century are you stuck in?
Anyone with a reading comprehension will realize that I'm responding to the OP that software raid is a joke, and that hardware raid is a necessity. My point, all along, has been that software raid is now good enough to replace hardware raid.
And apparently you have reading comprehension problems. Why buy another fucking socket when you can just get a cpu that can handle 5% more load? I gave an example of the CPU already - the CPU only cost *FIVE FUCKING DOLLARS MORE* for 5% more performance. In the edge case where you are sucking up every cpu cycle for your application, then, yes, hardware raid can make sense, but I'd also say you have seriously fucked up on your sizing because any server that's running at 100% is not capable of handling more load.
You keep going over into edge cases and tiny little machines that run at 100% of the cpu. *FOR MOST PEOPLE, SOFTWARE RAID IS BETTER THAN HARDWARE RAID*.
But you're a know it all. Can't even answer straight questions. I'm done. Enjoy being the "expert".
damn, and I just ran out of mod points
I guess the real question is this:
Do you agree/admit that current generation CPUs can calculate RAID parity faster that a hardware RAID card (in most cases, with minimal, 5% overhead)?
If yes, great.
If no, are you saying the realities that the youtube engineers saw, that other people saw (linux kernel hackers, etc), that Intel publicly stated are all false?
And people still don't know how to use authorized_keys? WTF?
At newegg, a Phenom II X4 965 at 3.4Ghz is $125. A Phenom II X4 975 at 3.6Ghz is $130. That 5% is $5.
At times, you cannot avoid buying cards without RAID - they simply don't make it. But as youtube discovered, even when you have a hardware RAID enabled card, going the software RAID route increased performance by 20% to 30%.
Which part of that do you not understand?
In a RAID environment, the IO bottleneck is at one of various places:
1) Disk mechanism (not under discussion)
2) Interface (not under discussion)
3) calculating RAID parity (hardware RAID vs software RAID)
4) higher level functionality (moving LUNs around, RAID hole, etc - not under discussion)
As of 2010, Intel has publicly declared that software RAID (good implementations, obviously) is faster than hardware RAID, and does not take up more than 5% of your CPU. Youtube's engineers have stated that they get 20% to 30% more performance out of software RAID than hardware RAID. Obviously *YOUR* expertise triumphs them all.
I am also worried that a guy who can't price a CPU that has 5% more performance without breaking the bank. You must be working on itty bitty machines.