George N. Capalbo, VP of backbone networks corp., happened to be nearby with his Leica still camera. which also takes short videos. He hit the "video" button by accident, and got this clip of the rousing billow of grey smoke. "Now I know what that guy felt like at the burning of the Hindenberg back in the 30s," George said.
Yes. Just like the guy who filmed the Hinderberg. 70 years from now, this crappy video of smoke coming out of a little white tower will be as etched on the human consciousness as the Hindenberg tragedy is now. Give me a break.
Stay away from CDNs if you're looking for performance (if it's pricing you're after, that's a seperate issue) - they are likely dealing with all the same issues that you are. Except that when something goes wrong, there's nothing you can do except shrug and say "it's the CDN's fault". My organization went to a CDN to save on ridiculous colo bills, and we heard all the claims about performance with skepticism, but we were shocked to find that despite all the "second generation edge servers" our CDN has, thruput actually went DOWN significantly when we switched to them. We didn't much care since it was saving us a lot of money and was usually unnoticeable or at least bareable...
But to return to your original question - since you're dealing with not-for-profits, clearly the answer is a live streaming P2P "edge network". Let me know when you release that code:)
If you don't have the development resources for that, the best solution is probably to scale down your investment in your current colo and pop a few servers in a different geographic locale, a kind of DIY edge network. With 3 or 4 colos (say, Europe, US-East, US-West), and smart determination of which host to route to using Geo::IP or the like, you could probably save a lot of traffic and latency.
Pics are nice, but what about battery life?
on
New iMac disassembled
·
· Score: -1, Offtopic
One of the things I loved about the G4 powerbook I briefly had was that it could do useful work for 5 hours on a routine basis. My wintel laptop knocks out after about 3-4 hours (and if I'm running my favorite Free OS where power management isn't a priority, I'm lucky to get half that much).
Going from the low power G4 with circa 6 hours of battery life to a dual core Intel, regardless of all the good reviews of the Duo, makes me nervous. Apple has also been consistently mum about battery life on these units.
Battery life is listed right on the front page of the powerbook page in the apple store - up to 5.5 hours. The MacBook page says nothing about battery life - nor does any other page I can find on apple's website.
I would probably spring for one of these guys if I knew battery life was comparable to the lowly G4 laptops apple sells - hell even if they told me it would only get 3-4 hours of battery life. But the fact that they omit battery life after raving so much about performance per watt leaves me wondering if folks who are ordering these puppies are going to be severely disappointed when they find their shiny new toy conks out after playing a full length movie.
Can someone with one of these beasts tell us how long the thing runs unplugged?
If you are worried about load, take the time to think about what really requires a round trip to the server, and what can more easily be done by populating some data on the webpage and then use that directly with javascript. My organization recently paid a certain overblown web design consultanting firm which botched a certain popular humor news site's website a lot of money, and they wanted us to use AJAX to autocomplete all of our forms - without making the simple connection that we only have about 100-200 elements in each autosuggest, and we could just put that data on the page and scrape it out with javascript, thus avoiding the extra server load and the lag time in the web browser. Make sure you see through the hype of AJAX, and don't use it when you don't need to.
If you use AJAX for everything, yes, your load will go through the roof. However, the good news is you probably don't need full blown AJAX to get most of the functionality you want. And, as other folks have pointed out, caching can make things a lot less painful where you really need it.
If you do something like this, be *VERY* careful how you cut partitions up on disk.
Essentially, you are doing RAID-0 (striping, no redundancy) of RAID-5 blocks. This means that if a single Raid-5 block goes out (not partity rebuild, but failed), so does all the rest of your data.
What I do is simple do software Raid5, and when I need to expand, bring my computer and a new disk in to work, dump my data onto a big frickin disk array we have there, reformat the Raid5 with an extra disk, restore, bring home, viola.
What's better than either of our solutions would be to simply get a RAID5 card that supports array expansion - they're not hard to find, and even though expanding is slow as molasses, you're guaranteed to have a recoverable state at each point in the expansion process.
Heh. Nothing is part of Moore's law: http://en.wikipedia.org/wiki/Moore's_Law . It's just an observation that's generally held true - clock speed tends to double every 1.5 years. What is "more of a law" is that there has to be a point when Moore's observation stops being the case, as we hit the hard limits implied by a single atom, electron, etc.
I didn't say that clock speed = performance, however, an increase in clock speed on the same architecture does indicate a probable increase in performance. I'm also sure you know that it is arguable at best that hyperthreading increases performance.
I'm also not sure where you got the impression that the article equated speed with clock rate - if they did, why did they bother with all those benchmarks?
As an example of the stagnation of the past few years, I have some code whose critical loop is unparallelizable (each instruction relies on the result of the previous instruction). My dual core Opteron running in 64 bit mode performs only about 15% better than my dual Athlon MP system from about 3 years ago. Synthetic benchmarks show a little more improvement, but largely only when memory access becomes a factor.
I'm not talking about servers, where parallelism is a necessity, or even general computing, I'm talking about unparallelizable, single threaded code. In this area, progress has been very slow. I'll grant you that the market is not as important in the scheme of things, but it is still there. Given how obsessed Intel in particular has been with clock speed to this point, it makes me wonder if they have gone to dual cores and such because they couldn't get more clock speed, which raises the question of whether we are hitting the physical limits of miniaturization.
Processor speeds haven't increased much in the past 2-3 years... Are we hitting the end of Moore's law, or just taking a detour as CPU makers decide dual-core is more important? I've been wondering about this for awhile, but haven't seen much discussion of this. Do we have to wait for quantum computers before we can get more single-thread performance?
'It doesn't help the economy produce more goods or services. It creates nothing of beauty or pleasure,' he writes. 'It simply helps someone get a bigger slice of the pie.'
You just described 99% of modern "brand-awareness" marketing. When will ads for Pepsi featuring the most recent just-overage female singer be recognized for the waste of resources that they are?
1) There is no significant protection here. A and B have to agree on a proxy. All the MPAA/RIAA has to do is a have a list of acceptable intermediate nodes C that are owned by them and not easily traceable to them, and push them out on the network. Now there is end to end encryption, but the MPAA knows who is talking to whom.
Combine this with periodic searches as a client for restricted content, and you've got a list of people offering probable restricted content.
They can even get trickier and start advertising content with filenames that sound right (but of course really just say "you're busted, neener neener"). In this case they act as B, the machine with the content, and they can have a very selective list of intermediate nodes (C) also controlled by them. In short, with a small farm, maybe 30 boxes, the MPAA is right where they are with Kazaa and other P2P applications.
2) There's nothing new here. This is just a stripped down version of anonymous remailers/onion routing, sans encryption.
3) The latency overhead of hopping to a node in between will be significant (as seen with tor), and probably kill the app. Not to mention the assymetric encryption overhead.
In short, it raises the bar a little, and for that is a good thing, but I'm afraid it's not raising the bar enough to make a difference for people who want to download copyrighted content (sorry).
With the EFF backing tor, you don't really have to worry about this scenario - they'd likely take the brunt of any legal suit, and I would imagine would provide legal assistance to any user of the tor network that got sued.
Before you all start griping, just remember... This is almost 1/2 a trillion dollars that won't be spent on "smart" bombs. Although I am not for the us military in general, I am all about it wasting it's money. Each $ wasted could be a life saved.
For me, the two pages were combined (Mozilla 6.0, WINNT 5.0). In any event, the Got Milk story sounds a lot more interesting to me. That's some disturbing s**t.
The nice thing about statistics, benchmarks, or anything else that tries to assign a number with some sort of value to something that's much more complicated is there are so many factors in the complicated thing, and thus ways to fsck up. Therefore, there are equally many ways to fudge.
Dude, don't mod this guy down, mod him up, he's right on.
The article is talking about "faster than the speed of electrons", whilst the title claims "faster than the speed of light". Two totally different things. This is clear even from the short description (ie w/o reading the article), and alot of the responses here make it clear the person only read the title (or didn't understand the rest).
Slashdot Mirror
Yes. Just like the guy who filmed the Hinderberg. 70 years from now, this crappy video of smoke coming out of a little white tower will be as etched on the human consciousness as the Hindenberg tragedy is now. Give me a break.
Guy 1: "Are you thinking what I'm thinking?"
Guy 2: "Yes!"
Guy 1: "I'll see you in court, asshole."
Stay away from CDNs if you're looking for performance (if it's pricing you're after, that's a seperate issue) - they are likely dealing with all the same issues that you are. Except that when something goes wrong, there's nothing you can do except shrug and say "it's the CDN's fault". My organization went to a CDN to save on ridiculous colo bills, and we heard all the claims about performance with skepticism, but we were shocked to find that despite all the "second generation edge servers" our CDN has, thruput actually went DOWN significantly when we switched to them. We didn't much care since it was saving us a lot of money and was usually unnoticeable or at least bareable...
:)
But to return to your original question - since you're dealing with not-for-profits, clearly the answer is a live streaming P2P "edge network". Let me know when you release that code
If you don't have the development resources for that, the best solution is probably to scale down your investment in your current colo and pop a few servers in a different geographic locale, a kind of DIY edge network. With 3 or 4 colos (say, Europe, US-East, US-West), and smart determination of which host to route to using Geo::IP or the like, you could probably save a lot of traffic and latency.
One of the things I loved about the G4 powerbook I briefly had was that it could do useful work for 5 hours on a routine basis. My wintel laptop knocks out after about 3-4 hours (and if I'm running my favorite Free OS where power management isn't a priority, I'm lucky to get half that much).
Going from the low power G4 with circa 6 hours of battery life to a dual core Intel, regardless of all the good reviews of the Duo, makes me nervous. Apple has also been consistently mum about battery life on these units.
Battery life is listed right on the front page of the powerbook page in the apple store - up to 5.5 hours. The MacBook page says nothing about battery life - nor does any other page I can find on apple's website.
I would probably spring for one of these guys if I knew battery life was comparable to the lowly G4 laptops apple sells - hell even if they told me it would only get 3-4 hours of battery life. But the fact that they omit battery life after raving so much about performance per watt leaves me wondering if folks who are ordering these puppies are going to be severely disappointed when they find their shiny new toy conks out after playing a full length movie.
Can someone with one of these beasts tell us how long the thing runs unplugged?
And since when is a tropical storm a hurricane?
:P
Sorry to be a snit, but you reap what you sow
Parent is racist troll, please mod
If you are worried about load, take the time to think about what really requires a round trip to the server, and what can more easily be done by populating some data on the webpage and then use that directly with javascript. My organization recently paid a certain overblown web design consultanting firm which botched a certain popular humor news site's website a lot of money, and they wanted us to use AJAX to autocomplete all of our forms - without making the simple connection that we only have about 100-200 elements in each autosuggest, and we could just put that data on the page and scrape it out with javascript, thus avoiding the extra server load and the lag time in the web browser. Make sure you see through the hype of AJAX, and don't use it when you don't need to.
If you use AJAX for everything, yes, your load will go through the roof. However, the good news is you probably don't need full blown AJAX to get most of the functionality you want. And, as other folks have pointed out, caching can make things a lot less painful where you really need it.
Good luck...
If you do something like this, be *VERY* careful how you cut partitions up on disk.
Essentially, you are doing RAID-0 (striping, no redundancy) of RAID-5 blocks. This means that if a single Raid-5 block goes out (not partity rebuild, but failed), so does all the rest of your data.
What I do is simple do software Raid5, and when I need to expand, bring my computer and a new disk in to work, dump my data onto a big frickin disk array we have there, reformat the Raid5 with an extra disk, restore, bring home, viola.
What's better than either of our solutions would be to simply get a RAID5 card that supports array expansion - they're not hard to find, and even though expanding is slow as molasses, you're guaranteed to have a recoverable state at each point in the expansion process.
Simple: Dual core processors are clocked slower so that heat dissipation is manageable due to having two cores in the same physical space.
Heh. Nothing is part of Moore's law: http://en.wikipedia.org/wiki/Moore's_Law . It's just an observation that's generally held true - clock speed tends to double every 1.5 years. What is "more of a law" is that there has to be a point when Moore's observation stops being the case, as we hit the hard limits implied by a single atom, electron, etc.
I didn't say that clock speed = performance, however, an increase in clock speed on the same architecture does indicate a probable increase in performance. I'm also sure you know that it is arguable at best that hyperthreading increases performance.
I'm also not sure where you got the impression that the article equated speed with clock rate - if they did, why did they bother with all those benchmarks?
As an example of the stagnation of the past few years, I have some code whose critical loop is unparallelizable (each instruction relies on the result of the previous instruction). My dual core Opteron running in 64 bit mode performs only about 15% better than my dual Athlon MP system from about 3 years ago. Synthetic benchmarks show a little more improvement, but largely only when memory access becomes a factor.
I'm not talking about servers, where parallelism is a necessity, or even general computing, I'm talking about unparallelizable, single threaded code. In this area, progress has been very slow. I'll grant you that the market is not as important in the scheme of things, but it is still there. Given how obsessed Intel in particular has been with clock speed to this point, it makes me wonder if they have gone to dual cores and such because they couldn't get more clock speed, which raises the question of whether we are hitting the physical limits of miniaturization.
Processor speeds haven't increased much in the past 2-3 years... Are we hitting the end of Moore's law, or just taking a detour as CPU makers decide dual-core is more important? I've been wondering about this for awhile, but haven't seen much discussion of this. Do we have to wait for quantum computers before we can get more single-thread performance?
Am I the only one that doesn't care? There's been at least 3 slashdot articles on this. Enough already.
'It doesn't help the economy produce more goods or services. It creates nothing of beauty or pleasure,' he writes. 'It simply helps someone get a bigger slice of the pie.'
You just described 99% of modern "brand-awareness" marketing. When will ads for Pepsi featuring the most recent just-overage female singer be recognized for the waste of resources that they are?
Ever heard of a tracking cookie?
I'd bet if you're running windows, you have an alexa cookie on your computer right now.
1) There is no significant protection here. A and B have to agree on a proxy. All the MPAA/RIAA has to do is a have a list of acceptable intermediate nodes C that are owned by them and not easily traceable to them, and push them out on the network. Now there is end to end encryption, but the MPAA knows who is talking to whom.
Combine this with periodic searches as a client for restricted content, and you've got a list of people offering probable restricted content.
They can even get trickier and start advertising content with filenames that sound right (but of course really just say "you're busted, neener neener"). In this case they act as B, the machine with the content, and they can have a very selective list of intermediate nodes (C) also controlled by them. In short, with a small farm, maybe 30 boxes, the MPAA is right where they are with Kazaa and other P2P applications.
2) There's nothing new here. This is just a stripped down version of anonymous remailers/onion routing, sans encryption.
3) The latency overhead of hopping to a node in between will be significant (as seen with tor), and probably kill the app. Not to mention the assymetric encryption overhead.
In short, it raises the bar a little, and for that is a good thing, but I'm afraid it's not raising the bar enough to make a difference for people who want to download copyrighted content (sorry).
With the EFF backing tor, you don't really have to worry about this scenario - they'd likely take the brunt of any legal suit, and I would imagine would provide legal assistance to any user of the tor network that got sued.
A verified node is not the same as a node, and we now have 100 VERIFIED nodes. RTFFAQ
My node, lemonmirangue, is within the past month, so was probably in the 90s. Someday, I'll get to brag about that.
Err, make that a half billion... :-)
Before you all start griping, just remember... This is almost 1/2 a trillion dollars that won't be spent on "smart" bombs. Although I am not for the us military in general, I am all about it wasting it's money. Each $ wasted could be a life saved.
For me, the two pages were combined (Mozilla 6.0, WINNT 5.0). In any event, the Got Milk story sounds a lot more interesting to me. That's some disturbing s**t.
for the love of god, mod nazi parent post down. thank you.
It is a benchmark, after all ;-)
The nice thing about statistics, benchmarks, or anything else that tries to assign a number with some sort of value to something that's much more complicated is there are so many factors in the complicated thing, and thus ways to fsck up. Therefore, there are equally many ways to fudge.
Dude, don't mod this guy down, mod him up, he's right on.
The article is talking about "faster than the speed of electrons", whilst the title claims "faster than the speed of light". Two totally different things. This is clear even from the short description (ie w/o reading the article), and alot of the responses here make it clear the person only read the title (or didn't understand the rest).