Now come on. You know very well that there's a huge difference between what is happening in America today and what the Soviets did. I don't know about you, but I do not have any fear of being woken up in the middle of the night, thrown into a van, and being shipped off to some Siberian gulag just because I surfed the wrong website last night.
The Soviets lived in constant fear of Big Brother because unlike our government, who I'll admit has a penchant for snooping, they chose to act on the information they gathered to actively suppress any non-state sponsored viewpoints. And, unlike our government, the Russian people could not approach their Soviet leaders for the same fear of reprisal.
I've said it a million times before here -- if you don't like the way things are going, do one of two things -- vote, or run for office. But putting on the tin foil hats for things that are easily obtainable by the governemtn through other sources, just reeks of knee jerk reactions.
By far, I find the mouse gestures extension to be the greatest addition to Mozilla. This borrowed feature of Opera will certainly and permanently change the way you browse websites.
Perhaps we can also get Wil Wheaton to play the part of Sir Lancelot's page, Concorde, that gets shot by Prince Herbert's arrow. Call it poetic justice...
I don't think that it's too much to assume that if a public test of a technology is being offered that said technology should be able, even if only partially, to satisfy the testers requirements. It's not like this is an in-house alpha product that may have an incomplete interface or an unseeded database -- this is a public beta test aimed at user acceptance testing. Given that search results are sparse at best, I'd say their UAT is not going very well.
Re:yay for legalized bribery!
on
P2P Bits
·
· Score: 4, Interesting
Not a law yet, just a bill. And given Sen. Hatch's track record, I wouldn't worry too much about it becoming law.
I think people are misunderstanding what I'm saying. I'm not saying don't provide it to the poor (or the rich, or whatever). I'm saying that there are cultures out there that just don't have a need for this. Just because there's a supply doesn't necessarily mean there's a demand.
My point goes way beyond the digital divide issues that you're all bringing up and challenges the ethnocentric assumption being made that everyone wants to be connected. This is simply not true. There are cultures that still exist, such as Amazonian tribes and even the Amish, that are functioning just fine without the influence of Internet kiosks or wireless access.
Some cultures in the world are still in the Bronze Age -- period. What good is ubiquitous Internet connectivity to a people that are comparatively primitive?
The right to privacy, while not explicitly enumerated by the US Constitution, has at least been granted by the Supreme Court in decisions over the years. Specifically, it is protected by the 9th Amendment.
For further information, go check out the privacy section at the US Constitution Online website.
No, I never said that at all, and to imply that I did is pure speculation on your part. I'm well aware of the entertainment industry's liberal bias but also aware that Sen. Hatch does provide a podium for groups like the RIAA to to get their views heard in Congress.
Like you say, special interest groups frequently grease the palms of politicians on both sides of the floor, and Hatch represents an easy way for groups to get their viewpoints heard by Congress.
The only reason that Sen. Hatch can pitch the ideas that he does is due to the fact Utah is so overwhelmingly Republican that as long as he wants to be in office, he will never get voted out. The end result of this is that he can pretty much say anything at all publically without fear of reprisal from his constituents.
The GOP and other right-wing/corporate leaning organizations know this and use him to pitch ideas that other Senators can not safely propose without possibly drawing the ire of their constituencies and risk getting replaced in 2/6 years. By contrast, Democrats do not have this luxury in the Senate, as there is no state in the nation that is as heavily biased towards Dems as Utah is towards Republicans, therefore you rarely ever see bills in the Senate with as extreme a left-leaning slant as Hatch's right-leaning bills.
So even if Sen. Hatch's ideas seem completely crazy to everyone, including his own party members, they do serve a purpose, which is to make the moderate conservative bills seem less crazy and outlandish, and therefore to get more credence. Coupled with the lack of an extreme liberal counterbalance to make moderate liberal bills seem more plausible, what we're left with is a permanent tilt towards the right in the Senate.
Well, OK. If you're affected by this DNS problem, then yes, referentialy, this will look like a DNS outage. However, I am not affected by it, but my access to the server was slow, which given the time proximity to the release of a new/. article chalks it up as a classic slashdotting. Either way, getting to that site regardless of the problem is going to be problematic for everyone for a little while.
I'll tell ya what. I will have a Sun V240 with Veritas Volume Manager v3.5 installed at my disposal tomorrow, and will install and configure two volumes on two of the four internal disks -- one striped and one mirrored -- where I'll create two simple UFS filesystems. I will then do a 'mkfile 100M/tmp/foo' and iteratively copy that to each filesystem (the write test) and then copy it back to/tmp/foo-stripe and/tmp/foo-mirror (the read test). I'll post the results tomorrow.
But let's assume that I'm out in left field. How do you disprove the results of the article? Can you point to data (a website would be sufficient) to back up your claims?
(Honestly, I'm asking -- let's assume that my 10 years of experience setting up HW and SW RAID are faulty, and that there's something out there that can change my mind. I understand what you're getting at, but it's counterintuitive and in my experience, I've never seen it occur where a well tuned and configured RAID 1 setup can ever outperform a similarly tuned RAID 0 setup.)
The controller has to be able to multiplex the reads to get the kind of performance you're talking about here, and given the tests in this article, this does not seem to be the case. With the exception of the old Intel ICH5R controller, none of the other controllers showed anything more than a statistically insignificant difference in single disk vs. RAID 1 read performance, and in fact, most of the RAID 1 read results were less than single disk, which is expected of simple controllers.
With respect to comparing reads from mirrors vs. striped disks, you are way out in left field again here and the article definitely backs this up. IIRC, the transfer rate for striped reads was about 33% faster than mirrored reads. This included the Intel controller that seemed to have some intelligence built in to multiplex the reads. The third to last graph in this page from the article should remove all doubt that RAID 0 reads are faster than RAID 1 reads, and that RAID 1 reads are nearly identical to single disk reads. And to hammer this point home further, why would a RAID 1 setup that could multiplex the disks for reads be any faster than a RAID 0 setup? Unless your RAID 0 stripe size is so small as to make overall RAID 0 slow (say a 64kb stripe vs. 2MB stripes), a multiplexed RAID 1 read from several disks simultaneously essentially is a RAID 0 read.
And with respect to write performance, I stand by my statement entirely. Striping, while more risky, is far faster than mirroring for writes. Mirrored writes get a little bit better with a duplex setup where the two drives are on separate controllers, but it's still slower than a striped write to two disks on the same controller.
Research done, conclusions drawn, still saying the same thing I was before....any other questions?
When reading a large file, for instance, if each disk can transfer, say, 10 MB/second and the file is 20 MB in size, the file can be loaded in one second with mirroring and two seconds without.
No, you're referring to striping, not mirroring. In fact, mirroring shouldn't give you any speedup in transfer rate at all -- since the data is mirrored to both drives, you're writing 20MB each to both drives. This will at a minimum cause a slight delay over writing to a single drive since the controller must push that 20MB over both channels to both drives, not split it 10MB on one and 10MB on the other.
Striping enhances performance at the cost of reliability, while mirroring is vice versa. Striping w/ mirroring (RAID 0+1) is better but requires at least 4 disks, while mirroring w/ striping (RAID 1+0) is better yet for those extra disks.
Slashdot Mirror
What's the current state of donations and what is the future of Wikipedia if fund raising without advertisements does not increase?
Now come on. You know very well that there's a huge difference between what is happening in America today and what the Soviets did. I don't know about you, but I do not have any fear of being woken up in the middle of the night, thrown into a van, and being shipped off to some Siberian gulag just because I surfed the wrong website last night.
The Soviets lived in constant fear of Big Brother because unlike our government, who I'll admit has a penchant for snooping, they chose to act on the information they gathered to actively suppress any non-state sponsored viewpoints. And, unlike our government, the Russian people could not approach their Soviet leaders for the same fear of reprisal.
I've said it a million times before here -- if you don't like the way things are going, do one of two things -- vote, or run for office. But putting on the tin foil hats for things that are easily obtainable by the governemtn through other sources, just reeks of knee jerk reactions.
By far, I find the mouse gestures extension to be the greatest addition to Mozilla. This borrowed feature of Opera will certainly and permanently change the way you browse websites.
Perhaps we can also get Wil Wheaton to play the part of Sir Lancelot's page, Concorde, that gets shot by Prince Herbert's arrow. Call it poetic justice...
Tooth fairy and Santa Claus
I don't think that it's too much to assume that if a public test of a technology is being offered that said technology should be able, even if only partially, to satisfy the testers requirements. It's not like this is an in-house alpha product that may have an incomplete interface or an unseeded database -- this is a public beta test aimed at user acceptance testing. Given that search results are sparse at best, I'd say their UAT is not going very well.
Not a law yet, just a bill. And given Sen. Hatch's track record, I wouldn't worry too much about it becoming law.
"I tried to patent patent barratry, but there was prior art."
Apologies to whomever's sig I have just trampled on...
I think people are misunderstanding what I'm saying. I'm not saying don't provide it to the poor (or the rich, or whatever). I'm saying that there are cultures out there that just don't have a need for this. Just because there's a supply doesn't necessarily mean there's a demand.
My point goes way beyond the digital divide issues that you're all bringing up and challenges the ethnocentric assumption being made that everyone wants to be connected. This is simply not true. There are cultures that still exist, such as Amazonian tribes and even the Amish, that are functioning just fine without the influence of Internet kiosks or wireless access.
Some cultures in the world are still in the Bronze Age -- period. What good is ubiquitous Internet connectivity to a people that are comparatively primitive?
Mass migration looms.
Big business senses problems,
Sticks finger in dike.
free access to information
TANSTAAFL. Mark my words, this connection will not go unpaid for -- otherwise why do it in the first place?
The right to privacy, while not explicitly enumerated by the US Constitution, has at least been granted by the Supreme Court in decisions over the years. Specifically, it is protected by the 9th Amendment.
For further information, go check out the privacy section at the US Constitution Online website.
No, I never said that at all, and to imply that I did is pure speculation on your part. I'm well aware of the entertainment industry's liberal bias but also aware that Sen. Hatch does provide a podium for groups like the RIAA to to get their views heard in Congress.
Like you say, special interest groups frequently grease the palms of politicians on both sides of the floor, and Hatch represents an easy way for groups to get their viewpoints heard by Congress.
The only reason that Sen. Hatch can pitch the ideas that he does is due to the fact Utah is so overwhelmingly Republican that as long as he wants to be in office, he will never get voted out. The end result of this is that he can pretty much say anything at all publically without fear of reprisal from his constituents.
The GOP and other right-wing/corporate leaning organizations know this and use him to pitch ideas that other Senators can not safely propose without possibly drawing the ire of their constituencies and risk getting replaced in 2/6 years. By contrast, Democrats do not have this luxury in the Senate, as there is no state in the nation that is as heavily biased towards Dems as Utah is towards Republicans, therefore you rarely ever see bills in the Senate with as extreme a left-leaning slant as Hatch's right-leaning bills.
So even if Sen. Hatch's ideas seem completely crazy to everyone, including his own party members, they do serve a purpose, which is to make the moderate conservative bills seem less crazy and outlandish, and therefore to get more credence. Coupled with the lack of an extreme liberal counterbalance to make moderate liberal bills seem more plausible, what we're left with is a permanent tilt towards the right in the Senate.
The first rule about US media coverage of US Gov't is you do not talk about the US media's coverage of the US Gov't. The second rule....
Carry on citizen...Big Brother is appeased.
Want to bet that the cost of the firmware upgrade would be, oh, I dunno, around $99?
Well, OK. If you're affected by this DNS problem, then yes, referentialy, this will look like a DNS outage. However, I am not affected by it, but my access to the server was slow, which given the time proximity to the release of a new /. article chalks it up as a classic slashdotting. Either way, getting to that site regardless of the problem is going to be problematic for everyone for a little while.
I am unable to access the server listed above from various server locations spread across the country & using different ISP's.
That's not the DNS outage problem -- the site is simply slashdotted.
And the results are in.
/dev/dsk/c1t1d0 and disk02 is /dev/dsk/c1t2d0, /tmp is swapfs, resident in memory.
/dev/vx/dsk/stripevol /dev/vx/dsk/mirrorvol /stripe /mirror /dev/vx/dsk/stripevol /stripe /dev/vx/dsk/mirrorvol /mirror /tmp/foo
/tmp/foo /stripe; sleep 1; done /tmp/foo /mirror; sleep 1; done
/stripe/foo /tmp/foo-stripe; sleep 1; done /mirror/foo /tmp/foo-mirror; sleep 1; done
Background -- root disk encapsulated and left alone, disk01 is
Setup
# vxassist -b make stripevol 1G disk01 disk02 layout=stripe
# vxassist -b make mirrorvol 1B disk01 disk02 layout=mirror
# newfs
# newfs
# mkdir
# mount
# mount
# mkfile 100M
Test
# while true; do time cp
# while true; do time cp
Write to stripe takes an average of 2.25 seconds.
Write to stripe takes an average of 3.24 seconds.
So this is what is expected - a penalty of about 44% for the striped write.
# while true; do time cp
# while true; do time cp
Read from stripe takes an average of 0.54 seconds.
Read from mirror takes an average of 0.54 seconds.
This is what I've been saying. There is no appreciable difference at all in performance here. Nothing, 0%, nada.
I'll tell ya what. I will have a Sun V240 with Veritas Volume Manager v3.5 installed at my disposal tomorrow, and will install and configure two volumes on two of the four internal disks -- one striped and one mirrored -- where I'll create two simple UFS filesystems. I will then do a 'mkfile 100M /tmp/foo' and iteratively copy that to each filesystem (the write test) and then copy it back to /tmp/foo-stripe and /tmp/foo-mirror (the read test). I'll post the results tomorrow.
But let's assume that I'm out in left field. How do you disprove the results of the article? Can you point to data (a website would be sufficient) to back up your claims?
(Honestly, I'm asking -- let's assume that my 10 years of experience setting up HW and SW RAID are faulty, and that there's something out there that can change my mind. I understand what you're getting at, but it's counterintuitive and in my experience, I've never seen it occur where a well tuned and configured RAID 1 setup can ever outperform a similarly tuned RAID 0 setup.)
The controller has to be able to multiplex the reads to get the kind of performance you're talking about here, and given the tests in this article, this does not seem to be the case. With the exception of the old Intel ICH5R controller, none of the other controllers showed anything more than a statistically insignificant difference in single disk vs. RAID 1 read performance, and in fact, most of the RAID 1 read results were less than single disk, which is expected of simple controllers.
With respect to comparing reads from mirrors vs. striped disks, you are way out in left field again here and the article definitely backs this up. IIRC, the transfer rate for striped reads was about 33% faster than mirrored reads. This included the Intel controller that seemed to have some intelligence built in to multiplex the reads. The third to last graph in this page from the article should remove all doubt that RAID 0 reads are faster than RAID 1 reads, and that RAID 1 reads are nearly identical to single disk reads. And to hammer this point home further, why would a RAID 1 setup that could multiplex the disks for reads be any faster than a RAID 0 setup? Unless your RAID 0 stripe size is so small as to make overall RAID 0 slow (say a 64kb stripe vs. 2MB stripes), a multiplexed RAID 1 read from several disks simultaneously essentially is a RAID 0 read.
And with respect to write performance, I stand by my statement entirely. Striping, while more risky, is far faster than mirroring for writes. Mirrored writes get a little bit better with a duplex setup where the two drives are on separate controllers, but it's still slower than a striped write to two disks on the same controller.
Research done, conclusions drawn, still saying the same thing I was before....any other questions?
When reading a large file, for instance, if each disk can transfer, say, 10 MB/second and the file is 20 MB in size, the file can be loaded in one second with mirroring and two seconds without.
No, you're referring to striping, not mirroring. In fact, mirroring shouldn't give you any speedup in transfer rate at all -- since the data is mirrored to both drives, you're writing 20MB each to both drives. This will at a minimum cause a slight delay over writing to a single drive since the controller must push that 20MB over both channels to both drives, not split it 10MB on one and 10MB on the other.
Striping enhances performance at the cost of reliability, while mirroring is vice versa. Striping w/ mirroring (RAID 0+1) is better but requires at least 4 disks, while mirroring w/ striping (RAID 1+0) is better yet for those extra disks.