To get rsync-like behavior using a multicast file transfer tool, one could use something like xdelta3 to distribute difference files repeatedly. And the file format is actually an IETF draft standard (which cannot be said of rsync or rdiff-backup).
I've never heard of any other organizational use besides Linux distribution and Blizzard's "critical business function". What are your other examples?
It would seem to me that IP multicast tools and efficient client-server tools like rsync are far more useful in an organization for content distribution than BitTorrent for "critical" processess. You get better traffic locality, deterministic behavior, logging, and a lot of other necessities that you can't get with BitTorrent clients.
Let me guess: you shop at Wal-Mart exclusively?
As a corporate buyer, I'm willing to spend a bit mroe for reliable service. CDW does actually have an accurate compatability database and most significantly accurate and exhaustive stock levels. Which is very important to corporate buyers, especially when you're buying something like a Cisco switch with a million options or a $200K SAN.
Having been burned many times by "oops, it's really out of stock" or "wrong SKU for description" or "price increased since your order" with Tiger, Insight, NewEgg, and others, we went back to sending most of our money CDW's way. The platinum discount is significant on software, and even without their corporate accounts you would pay maybe 1-2% more than the lowest price out there. More than worth it if it matters that your gear will arrive the next day or even same-day when necessary.
Vendors with good, professional service save me and my staff time, and that time is worthmoney. If you spend an hour comparison shopping to save $100, you're a fool.
Yea, I recently built an 8T server. It cost me about $5000, and has no raid controller, and uses linux software raid.
And you suffer from the famous "RAID-5 write hole". Hello silent data corruption. And good luck during a rebuild if when you have a latent sector error.
If I really wanted, I could buy a second $5000 server and do DRBD between them to have 2x redundancy.
DRBD is not instant-failover clustering and load balancing, it is simply block-level replication. The vast majority of applications - especially databases - would undoubtedly corrupt data during a node failure on your proposed setup. If all you're storing is MP3s, fine, but if you have lots of apps banging away at your storage, Linux software RAID and DRBD don't cut it.
We have some real "enterprise-class" clustered storage systems built on Linux from LeftHand. In testing we've ripped the plugs on one of the five nodes (plain HP DL320S's rebadged) and every application just kept on trucking with no corruption or downtime. And 10K random 8K IOPS, 1/3 of which are writes. Nothing you can build yourself with Linux can currently match those capabilites, so we paid $100K for it.
Google gets away with cheap storage using very sophisticated software, and by controlling the entire application stack to tolerate storage and node failures. You can't buy Google's software, and even if you could you couldn't run your app of choice on BigTable without a lot of coding. But you can get similar capability by paying money to LeftHand, IBRIX, NetApp and other "enterprise" clustered storage vendors.
Sure, JavaScript is pointless in a PDF viewer and should be disabled, but it is worth noting that PostScript itself is a programming language. It has conditionals, functions, loops, etc. I myslef once hand-coded a PostScript program to draw a high-res graph of a particular function for a class back in college. This 1K file basically owned the imagesetter in the print lab for about 45 minutes while it rendered at 1200 dpi.
If I recall correctly, there were even a couple of postscript exploits back in the 1990s that could "brick" Apple LaserWrtiers.
I'm sorry, but where did you get the idea that environmental costs show up explicitly and directly on balance sheets?
I never said they did. What I said is that the non-balance sheet environmental costs don't make a difference in decisionmaking, because everyone who matters ignores them. Do you really think China's Ministry of Transportation gives a shit about Braziallian rain forests? Or BP's stockholders? They don't, or at least not enough to matter when it comes down to money.
If the options are $X for this solution, and $1.5 * X for the eco-friendly solution, guess which one will be widely implemented, no matter what the Sierra Club has to say?
Yes, but let us hope that the best alternative fuel solution is implemented and not the most profitable.
Um.... best = "most profitable" in this case. That is, the best alternative fuel is the one with the lowest costs and highest return on investment. Those costs include cost of manufacture, distribution, and infrastructure upgrades needed for widescale use.
If you're going to try to redefine "best" to be "the one that kills the fewest four-toed sloths" or something, dream on. This is the real world.
It sounds like you're doing electronic check conversion, which is essentially a good old-fashioned wire transfer. Are you sure you can do that for your personal account? If so, that's the first I've heard of it, and I work in the banking industry.
In any case, paper check writers are just as protected in a ECC transaction as they are with a paper check. Different laws, but he liability is essentially on you as the payee and both banks to handle the trasnsaction securely, and accept the consequences of fraud.
But very few criminals will do check fraud for $500, beacuse the penalty is so high if she is caught. Also, doing lots of little $500 fradulent transactions makes it that much more likely to be caught. Not to mention that each instance adds 5 years or whatever to a prison sentence.
You know... I can't even recall seeing checks outside America since the 80ths. The rest of the world uses cash, bank transfers and credit/debit cards. And we survive, without the costs and problems associated with a ridiculously broken check system
Well, almost nobody in america uses paper checks for retail purchases either. Most retailers recognize the possibiltiy for fraud and don't accept them except from well-known customers. But checks are still widely used for payments made through the mail, either B2B, P2B or P2P. Going to a bank branch and doing a bank transfer simply isn't cost-effective or convenient, although Internet banking has changed that somewhat. Cross-bank transfers via a website are still pretty dicey and untrusted or not supoprted by banks at all.
As for chip-and-pin or smartcards, when I was last in France and the UK in 2006 neither was even close to ubiquitous. And both have had widely reported security issues. Credit and debit cards suck just as bad as paper checks from a security perspective, but are a lot more convenient. My point was there has been no technology that US Banks have been willing to push, because none pass the return-on-investment test. And the worst possible thing would be to spend hundreds of billions deploying some new smartcard based system that turned out to be insecure, and have to do it all over again. We like to spend that kind of money on Stealth bombers instead.
any piece of stationary with mag ink at the bottom with bank a.b.a., account number, check number, will be accepted as check
No, it most likely won't. What you say may have been true 10 or even 5 years ago, but is generally not true with modern check imaging systems. The "Check 21" legislation basically enabled all banks to move to electronic check image storage. Of course, they had to upgrade all of their imaging systems to recognize that cost savings, and these new systems are quite discerning, especially for higher-value checks. Manual inspection is required for most high-value checks, and even things like a changed paper stock or layout can be flagged for manual review.
Also, nearly every company of reasonable size is required to implement positive pay, meaning they send a list of check numbers, dollar amounts, and payees to the bank before the checks are actually cut. So when you go to cash a fake check, the bank knows it is fake immediately. There are of course ways to get around this, especially with personal accounts (which usually do not offer positive pay), but check fraud is no longer as simple as portrayed in Catch me if you Can.
That said, check still fraud remains a major cost for banks, and believe it or not they are working hard to make it less possible. But there is as yet no "magic bullet" technology to replace paper checks. Chip-and-PIN, smartcards, etc. all suffer from different security and operational issues. They also cost a lot to implement worldwide, even after including the costs of paper check fraud. A paper check is fairly easily validated, can be sent through the mail, and requires no "secure" hardware terminals at every merchant.
Actually, the hash function already includes the file size. The padding algorithm for all widely used cryptographic hashes uses the byte length of the file as part of the padding.
So, simply using a longer hash function is a much better way to reduce false-positives. SHA-256, for example, would make the chances of collision on the order of 2^(-128). SHA-512 or Whirlpool would result in collision probability of 2^(-256).
Yes, the big-iron SANs cost a lot. Too much. But you can build your own for not that much money. Yeah, you lose some of the cool management features. But so what? They're so cheap, that for the price of one commercial SAN, you can build TEN of the things.
Building ten OpenFilers makes the problems that need solving worse, not better. Your have islands of storage, 10x the failure probability, and 10x the mangement expense.
SANs are about provisioning and protecting storage in a mangable, policy-driven way. As well as enabling features like clustering, thin provisioning, differntial replication, zero-loss failover, VSS-aware snapshots, and a million other things no FOSS solution even comes close to doing. We paid >>$100K for ~10 TB of clustered iSCSI storage, and it has been worth every penny. We have had zero downtime and zero data loss since implementation four years ago, all the while performing upgrades to the SAN adding new features and increasing performance. In testing we yanked the power from one of the storage modules (really just a rebadged HP DL320S) and everything hummed along nicely, and re-synced in minutes when power was restored.
Contrary to Slashrone lore, IT purchasing managers are not idiots. We pay vendors for these expensive toys because they are actually better in every way and more valuable in every way than your home-brew Linux Frankenserver.
Courts have determined that a life has a monetary value, I think in the mid 6 figures.
This is certainly false, at least in the United States. Do you have a credible reference?
It may be that the average judgement in a wrongful death civil case might be in the mid six figues, but that is a very different thing than a legal precedent which says "one life = $X". There are cases where $x is tens of millions, and others where $x is near zero. It depends on the merits of the case, the skill of the lawyers involved, and the mood of the jury and judge. Which is as it should be.
Full 1920x1808 HD video requires >10x the bandwidth of YouTube-quality video per stream. And none of the mentioned HD sites has financial the infrastructure resources (private peering), of Google, Inc. Does that explain it enough for ya?
Aww, cmon. Less than 6 hours is "so long"? What time zone are you in?
Seriously, though, I had one hell of a time time convincing our CFO that the facilites were more important than the networking gear and servers, which could be changed rather cheaply and easily at any point in the project. I actually now wish that we hadn't built our own datacenter at all, as we made budget compromises that makes our shiny new DC riskier than leasing connectivity to leased "tier 1" colocation space. Our fire-suppression system is water based, for instance. And we have a single point of failure in the plumbing for chilled water.
Slashdot Mirror
To get rsync-like behavior using a multicast file transfer tool, one could use something like xdelta3 to distribute difference files repeatedly. And the file format is actually an IETF draft standard (which cannot be said of rsync or rdiff-backup).
I've never heard of any other organizational use besides Linux distribution and Blizzard's "critical business function". What are your other examples?
It would seem to me that IP multicast tools and efficient client-server tools like rsync are far more useful in an organization for content distribution than BitTorrent for "critical" processess. You get better traffic locality, deterministic behavior, logging, and a lot of other necessities that you can't get with BitTorrent clients.
Except we're talking about price differences of (typically) less than 5% here. So your argument holds no water.
Let me guess: you shop at Wal-Mart exclusively? As a corporate buyer, I'm willing to spend a bit mroe for reliable service. CDW does actually have an accurate compatability database and most significantly accurate and exhaustive stock levels. Which is very important to corporate buyers, especially when you're buying something like a Cisco switch with a million options or a $200K SAN. Having been burned many times by "oops, it's really out of stock" or "wrong SKU for description" or "price increased since your order" with Tiger, Insight, NewEgg, and others, we went back to sending most of our money CDW's way. The platinum discount is significant on software, and even without their corporate accounts you would pay maybe 1-2% more than the lowest price out there. More than worth it if it matters that your gear will arrive the next day or even same-day when necessary. Vendors with good, professional service save me and my staff time, and that time is worthmoney. If you spend an hour comparison shopping to save $100, you're a fool.
And you suffer from the famous "RAID-5 write hole". Hello silent data corruption. And good luck during a rebuild if when you have a latent sector error.
DRBD is not instant-failover clustering and load balancing, it is simply block-level replication. The vast majority of applications - especially databases - would undoubtedly corrupt data during a node failure on your proposed setup. If all you're storing is MP3s, fine, but if you have lots of apps banging away at your storage, Linux software RAID and DRBD don't cut it.
We have some real "enterprise-class" clustered storage systems built on Linux from LeftHand. In testing we've ripped the plugs on one of the five nodes (plain HP DL320S's rebadged) and every application just kept on trucking with no corruption or downtime. And 10K random 8K IOPS, 1/3 of which are writes. Nothing you can build yourself with Linux can currently match those capabilites, so we paid $100K for it.
Google gets away with cheap storage using very sophisticated software, and by controlling the entire application stack to tolerate storage and node failures. You can't buy Google's software, and even if you could you couldn't run your app of choice on BigTable without a lot of coding. But you can get similar capability by paying money to LeftHand, IBRIX, NetApp and other "enterprise" clustered storage vendors.
You, sir, are never allowed anywhere near my data centers!
Sure, JavaScript is pointless in a PDF viewer and should be disabled, but it is worth noting that PostScript itself is a programming language. It has conditionals, functions, loops, etc. I myslef once hand-coded a PostScript program to draw a high-res graph of a particular function for a class back in college. This 1K file basically owned the imagesetter in the print lab for about 45 minutes while it rendered at 1200 dpi.
If I recall correctly, there were even a couple of postscript exploits back in the 1990s that could "brick" Apple LaserWrtiers.
I never said they did. What I said is that the non-balance sheet environmental costs don't make a difference in decisionmaking, because everyone who matters ignores them. Do you really think China's Ministry of Transportation gives a shit about Braziallian rain forests? Or BP's stockholders? They don't, or at least not enough to matter when it comes down to money.
If the options are $X for this solution, and $1.5 * X for the eco-friendly solution, guess which one will be widely implemented, no matter what the Sierra Club has to say?
Um.... best = "most profitable" in this case. That is, the best alternative fuel is the one with the lowest costs and highest return on investment. Those costs include cost of manufacture, distribution, and infrastructure upgrades needed for widescale use.
If you're going to try to redefine "best" to be "the one that kills the fewest four-toed sloths" or something, dream on. This is the real world.
Are you sure? I haven't looked at the block diagrams in detail, but apparently this thing is the bee's knees.
But it defies the common notation. I think "Core 0+7i" would be unambiguous. What the hell is wrong with those marketing types?
Wouldn't that make it "Core 7i" instead of "Core i7"?
It sounds like you're doing electronic check conversion, which is essentially a good old-fashioned wire transfer. Are you sure you can do that for your personal account? If so, that's the first I've heard of it, and I work in the banking industry.
In any case, paper check writers are just as protected in a ECC transaction as they are with a paper check. Different laws, but he liability is essentially on you as the payee and both banks to handle the trasnsaction securely, and accept the consequences of fraud.
But very few criminals will do check fraud for $500, beacuse the penalty is so high if she is caught. Also, doing lots of little $500 fradulent transactions makes it that much more likely to be caught. Not to mention that each instance adds 5 years or whatever to a prison sentence.
Well, almost nobody in america uses paper checks for retail purchases either. Most retailers recognize the possibiltiy for fraud and don't accept them except from well-known customers. But checks are still widely used for payments made through the mail, either B2B, P2B or P2P. Going to a bank branch and doing a bank transfer simply isn't cost-effective or convenient, although Internet banking has changed that somewhat. Cross-bank transfers via a website are still pretty dicey and untrusted or not supoprted by banks at all.
As for chip-and-pin or smartcards, when I was last in France and the UK in 2006 neither was even close to ubiquitous. And both have had widely reported security issues. Credit and debit cards suck just as bad as paper checks from a security perspective, but are a lot more convenient. My point was there has been no technology that US Banks have been willing to push, because none pass the return-on-investment test. And the worst possible thing would be to spend hundreds of billions deploying some new smartcard based system that turned out to be insecure, and have to do it all over again. We like to spend that kind of money on Stealth bombers instead.
No, it most likely won't. What you say may have been true 10 or even 5 years ago, but is generally not true with modern check imaging systems. The "Check 21" legislation basically enabled all banks to move to electronic check image storage. Of course, they had to upgrade all of their imaging systems to recognize that cost savings, and these new systems are quite discerning, especially for higher-value checks. Manual inspection is required for most high-value checks, and even things like a changed paper stock or layout can be flagged for manual review.
Also, nearly every company of reasonable size is required to implement positive pay, meaning they send a list of check numbers, dollar amounts, and payees to the bank before the checks are actually cut. So when you go to cash a fake check, the bank knows it is fake immediately. There are of course ways to get around this, especially with personal accounts (which usually do not offer positive pay), but check fraud is no longer as simple as portrayed in Catch me if you Can.
That said, check still fraud remains a major cost for banks, and believe it or not they are working hard to make it less possible. But there is as yet no "magic bullet" technology to replace paper checks. Chip-and-PIN, smartcards, etc. all suffer from different security and operational issues. They also cost a lot to implement worldwide, even after including the costs of paper check fraud. A paper check is fairly easily validated, can be sent through the mail, and requires no "secure" hardware terminals at every merchant.
Actually, the hash function already includes the file size. The padding algorithm for all widely used cryptographic hashes uses the byte length of the file as part of the padding.
So, simply using a longer hash function is a much better way to reduce false-positives. SHA-256, for example, would make the chances of collision on the order of 2^(-128). SHA-512 or Whirlpool would result in collision probability of 2^(-256).
Building ten OpenFilers makes the problems that need solving worse, not better. Your have islands of storage, 10x the failure probability, and 10x the mangement expense.
SANs are about provisioning and protecting storage in a mangable, policy-driven way. As well as enabling features like clustering, thin provisioning, differntial replication, zero-loss failover, VSS-aware snapshots, and a million other things no FOSS solution even comes close to doing. We paid >>$100K for ~10 TB of clustered iSCSI storage, and it has been worth every penny. We have had zero downtime and zero data loss since implementation four years ago, all the while performing upgrades to the SAN adding new features and increasing performance. In testing we yanked the power from one of the storage modules (really just a rebadged HP DL320S) and everything hummed along nicely, and re-synced in minutes when power was restored.
Contrary to Slashrone lore, IT purchasing managers are not idiots. We pay vendors for these expensive toys because they are actually better in every way and more valuable in every way than your home-brew Linux Frankenserver.
This is certainly false, at least in the United States. Do you have a credible reference?
It may be that the average judgement in a wrongful death civil case might be in the mid six figues, but that is a very different thing than a legal precedent which says "one life = $X". There are cases where $x is tens of millions, and others where $x is near zero. It depends on the merits of the case, the skill of the lawyers involved, and the mood of the jury and judge. Which is as it should be.
Okay, so I challenge you: can you give an example of a centrally planned economy that has not failed miserably?
Free market capitalism is the worst form of economy, except for all the others.
Full 1920x1808 HD video requires >10x the bandwidth of YouTube-quality video per stream. And none of the mentioned HD sites has financial the infrastructure resources (private peering), of Google, Inc. Does that explain it enough for ya?
Because internet transit bandwidth costs money, and international transit bandwidth costs more money than US-only bandwidth?
Of course! I mean, after all, a centrally-planned, price-fixed economy worked so well in the Soviet Uniuon and Eastern Europe of the 1980s...
clearly your broken keyboard resulted in many bugs committed and also poorly written business documents, which is probably why you really got canned
Aww, cmon. Less than 6 hours is "so long"? What time zone are you in?
Seriously, though, I had one hell of a time time convincing our CFO that the facilites were more important than the networking gear and servers, which could be changed rather cheaply and easily at any point in the project. I actually now wish that we hadn't built our own datacenter at all, as we made budget compromises that makes our shiny new DC riskier than leasing connectivity to leased "tier 1" colocation space. Our fire-suppression system is water based, for instance. And we have a single point of failure in the plumbing for chilled water.