Yeah, I tried that before, doesn't work all that well with Windows. With Linux and practically any other OS you can just deploy a generic image with a modular kernel on the system and it will generally work. Network adapters work with a generic driver, video cards work with a generic driver, usb ports work with a generic driver.
With Windows, only the APIC or the boot drive hardware (PATA/SATA/SCSI) have to be different from the original host for the thing to give a blue screen even when sysprepped. Even when the drivers are included and you have an image of a system with the same APIC, the system has been sysprepped but the USB ports aren't the same as whatever machine you made the image off, the system won't be able to react to your input unless all USB hardware has been re-detected (which can take a while and sometimes requires a cold reboot as you can't click on the dialogs). Whenever an update (especially Service Packs) needs to be included in your image, all drivers have to be re-checked (manually) for all your different hardware to make sure none needs to be updated as well. Ideally you would have test-systems, replicas of each piece of hardware you have but even in small organizations this can add up to 10's or 100's of idle hardware that you have to acquire and justify.
I now know why large organizations standardize on a single vendor and can't offer their end-users any choices in hardware besides the amount of RAM and hard drive space. Windows is just plain bad to maintain even with experienced admins. I have virtualized practically all installations of it and even though it takes a slight performance hit, it's much easier to manage than trying to keep up with images for all the different hardware you can have in a single organization.
It doesn't make SSL less secure. It just makes it less trustworthy. You should never trust an SSL certificate to provide you with any identification. SSL's are there to encrypt the connection and so far, that level of protection provided is fairly immune to attacks.
If you want a website to identify themselves, they should provide you with something else, an identifier that is unique to your account with them that you might have established before (as many banks are starting to do - they provide a picture with a non-sense phrase that you established during the account creation procedure).
I would never hire anyone with an MS** certification as they are particularly braindead and responsible for some of the most stupid decisions in any IT organization and responsible for the bad name that most IT departments have when it comes down to costs and ROI.
I run my IT department with maybe 10% of my yearly budget spent on software licensing (most of which is MATLAB and Microsoft Office). There are other departments with very similar end-user requirements where compulsory Microsoft and Oracle software licensing accounts for 50% of their budgets (the other 50% being personnel and hardware) and their total budget at least 300% as large as mine. The difference I put back in hardware. Where most other end-users have to be content with Core2Duo's as their high-end machines, I can afford to give my engineers/scientists dual quad core machines and next week dual hex core workstations.
Yes, ZFS is very simple (the code is very clean and only a few thousand lines - a lot of them comment) but it desperately needs to be ported to other platforms. OpenSolaris is nice but the only reason I use it is ZFS. Imitation ZFS (Btrfs) might eventually become good enough for me to use it but the current feature set of ZFS is very attractive for just about anyone who needs to host multi-GB's of data. Other nice features of OpenSolaris includes their COMSTAR stack (make cheap hardware appear as very expensive iSCSI or FibreChannel targets), zones and dtrace.
Because Amazon is getting paid for their services. Amazon isn't making a loss when criminal syndicates use their services nor are they providing it for free to those organizations. They're probably still pumping cash into the whole EC2 thing since "cloud computing" isn't really as popular and world-changing in most businesses as was projected 5 years ago so they could probably use the $.50/GB at whatever rate these people are pumping out.
Technically, if you throw all of that data online, there is no need to keep the brick-and-mortar buildings open decreasing cost substantially to the already bankrupt state.
CUPS has the ability to support authenticated print jobs and you can lock it down as far as you want to, it's just that nobody uses it (or knows about it) because it's hardly necessary. CUPS is an admin interface, the end-user doesn't have to know about it so it could be at some points a bit arcane to use it if you don't know what you're doing.
The main problem with CUPS and just about any other printer management solution is the drivers. With the advent of inkjet printing and Windows 95, device makers didn't have to support standards anymore, they would just roll their own drivers and Microsoft made it easy for them (and later for more malicious agents) to automatically install crappy software. Olivetti was among the first inkjet printers that didn't have any processing power aboard, leaving all print processing to be done by the host. I think this was one of the reasons they don't exist anymore. Unless you got a laser printer, you just couldn't send a postscript file down to the printer anymore.
These days, people want their printers to connect to their networks so we're starting to see even the cheap printers have their own PostScript processing and support open standards so connecting to it is simple enough even for 'dumber' devices like routers and handheld devices that don't have place for custom printing drivers.
This is typical of short-term thinking on the side of management. Here they are thinking really, really short-term. The people that complain already purchased their copy of whatever game they were selling. There is no benefit in providing them with any more service as they won't pay to cover the costs of it, the profit has been made, the game has been sold in most places you can't return digital media once you opened the shrink wrap without a court order practically.
Well, I guess it depends on where you live then. Everything is technically lethal. However just because somebody can't see that you used a certain weapon does not mean that you should use it more frequently, however, sadly that is what happens - cops (threaten to) taser you when you argue with them in your car when they want to give you a ticket. Ever heard of a cop clubbing somebody because they didn't want to sign their ticket? Or pepper-spray them?
And off-topic, not my quote, it's a shortened SCOTUS-case quote and in that case pertained to the first amendment, not the others (although you could argue that the other amendments should also not be submitted to vote):
"The very purpose of a Bill of Rights was to withdraw certain subjects from the vicissitudes of political controversy, to place them beyond the reach of majorities and officials and to establish them as legal principles to be applied by the courts. One's right to life, liberty, and property, to free speech, a free press, freedom of worship and assembly, and other fundamental rights may not be submitted to vote; they depend on the outcome of no elections." -- Supreme Court Justice Robert Jackson
The problem with the taser replacing the club is that when you hit somebody with a club, everybody knows if you hit hard enough, you can kill somebody, it also leaves bruises. The problem with the taser is that everybody knows 'the company that makes them says they are not lethal at all under any circumstance' and you can't see whether somebody's been hit with a taser. In the beginning of the taser-era, officers would call an ambulance before or after tasering somebody. These days it seems they don't even bother anymore (depending on the type of tasers they use).
The taser has not been tested in any viable study I know off against either human targets or human replica's. As MythBusters and many electrical engineers will tell you, a shock across the heart of just 1mA can kill you, 100mA is lethal. As every geek knows I = V/R and tasers bolt out about 50-100kV which means your skin-to-heart resistance needs to put up a resistance of 500k (if my calculations are correct). Your body resistances ranges anywhere from 300ohms-6Mohms. When your skin is moist (sweating,...) as is common with drug-addicts and people running from the law your body's resistance will drop. If a taser hits you near the heart in those conditions, they can be theoretically very lethal.
That's because of the nature of an Unidentified Flying Object. It flies through the sky but is too far away or obscured by other things (weather, buildings) to be identified. UFO's will always remain UFO's no matter whether they be nature-made, man-made or alien-made phenomenon.
That's why we should legalize all drugs and have them sold in drugstores (also called pharmacies). Make it illegal to sell to underage children and make it illegal to drive/work under the influence of any drug. This would take care of violent drug cartels whom are only fueled by the high prices the underground trading commands. Making it legal takes the risk out of it, drops the prices and leaves those drug dealers out of a job.
People will get high somehow, people will kill themselves with it just as they do with alcohol right now, nobody cares that somebody drinks themselves to death or that they can't keep a job because they're drunk all the time - they made that decision, they will bear the consequences, they will eventually die.
Everybody "peers" at some point. As an end-user (and most website owners too) we don't have enough money to peer with any of the big players. That's why we peer with our ISP's (or host our servers at somebody else's datacenter) to provide this peering. We don't necessarily pay for bandwidth (as those agreements may vary from dedicated bandwidth pipes to pay-per-bit bandwidth) but we all pay to peer to other networks.
Google is big enough that they can peer themselves. They chip in at the point of the core routing infrastructure just like all the other ISP's, datacenter etc. Google is it's own private ISP. When ISP's want Google to pay for bandwidth usage on their networks, they would (if they were honest) also have to start charging/paying each other to use each other's bandwidth whether it be their customers BitTorrenting between each other or connecting to a server in a datacenter.
Do you know what a web app firewall actually does? It's practically a proxy server that sanitizes the inputs, meaning for each application you have to make custom rules which are more involved than just "allow 0.0.0.0/0 to 192.168.1.1 port 80". Especially if you have a very custom-built application, for each data input (each field) you will have to specify what form of data it is and what the valid ranges are. This SHOULD be done within the application already by the developers, there is no reason to go out of your way to buy software and practically rebuild the whole front-end and validation of each of those pages, you might as well build the back-end as well and just do the whole application yourself.
(Bad) sequels being written or story lines being hijacked has nothing to do with copyright and probably can't even be protected by it (since it's not the same story). After all, the original author is dead. The only way you can show your disappointment is by not buying those books, eventually the author will have to move on to fresh content or change jobs where imagination is not required as much.
You will also notice that these 'pledges' don't do very much in the long run. IBM, Google, Microsoft, Novell, Red Hat and Apple all have 'pledged' some type of protection for their open source ancestors but those things are not legally binding no matter what they might say about it (I think it's MSFT that has such a claim).
That's also why you should avoid implementing any of their proprietary crap in your Open Source project (or any project that's being made public or sold in any shape or form) because if for any reason they want to leverage their arbitrary licenses on it, they can no matter what they have promised whether it's uncanny legal speak or so-called patent pools.
The main problem with most compliance protocols (HIPAA or PCI) is that at best they do nothing at all, at worst it's actually counterproductive as it opens the company up to more breaches (due to human nature, laziness or conflicting policies).
I am involved in both HIPAA and PCI compliance and in the past I have been involved with Sarbanes-Oxley as well. For example with PCI as well as Federal wiretapping compliance, you need to have your respectively wireless and public networks (if you're a de-facto wireless internet provider to random strangers - eg. libraries, universities,...) run through a separate (3rd party) provider and needs to be either logically or physically divided from the main network. Therefore, anyone on your public or wireless network will have to tunnel a VPN through a 3rd party provider, route it out to the internet and back into your primary provider to get work done which makes the whole system inherently less secure because your data goes outside your network.
PCI requires a firewall before your internet facing servers but also a perimeter firewall (if you have a really large institution) before all your edges even though you may have separate departmental firewalls. This does not make sense as you get to have 2 or 3 layers of firewalls - the first 2 layers being the ones that were historically built-up and the 3rd layer, a concentrated firewall and internet provider hub which becomes 1) easier to attack because it's all in one point, 2) easier to fail for the same reason, 3) more difficult to maintain because you still need the hierarchy of departmental firewalls to prevent attacks from other departments or other points in the network.
UPS/FedEx? That's their business. The USPS (and a lot of other postal systems throughout the world) has been putting more emphasis on their 'added services' like package shipping and money transfers to adjust for the diminishing returns of snail mail delivery. I actually get mad when people send me snail mail letters on dead tree - it's bad for the environment and e-mail is a lot simpler and easier.
Most people do get their licenses taken away after the first or second one. I don't know in other states but I know in NY if you get your license taken away, it's quite difficult to get it back. Off course, then you get the problem of people driving around WITHOUT licenses.
Sun died, not because it gave away stuff but because it overpriced stuff and made some really bad decisions. A Thumper costs about 50-100k and all it is, is an AMD Opteron-based machine with some SATA controllers that fits lots of disks. I can get the same hardware for 10-20k elsewhere. Same goes for their 1U servers - I believe it's 10k for an entry-level model. They have some really nice CPU's though (Niagara now) but they never marketed it right, never priced it right and management killed some really nice projects.
The same with ZFS. They never marketed it as a company but the developers did market it to other developers and the open source community which made it very popular. ZFS is simply awesome and if upper management would've seen what this project does, they could've easily been taking over NetApp (where their buyers don't seem to mind spending 20k/TB) and other high-end proprietary storage vendors while staying safe from a hostile takeover.
I am building a dual-parity, redundant 30TB storage array with 320GB read cache and 64GB write cache (POSIX-correct) for under 30k with ZFS. Sun could have easily charge 300k for something like this and most customers would be grateful not to pay NetApp or EMC for it.
But when they do get CPU-bound, you have serious problems. I don't know if things changed since 2000/2003 but I remember very well when 6 geographically separated, load-balanced MSSQL boxes (with 8 cores each back then - very costly setup) hit the CPU bound, they all became unresponsive for a couple of seconds (dropped down to 0% CPU), came back, did a few thousand queries and then repeated the cycle every 5 seconds.
The report fails to mention that across a large enough population, the results will be more-or-less correct within a certain percentage point because as he mentions, some people will test with a lot of bandwidth available at a certain point but others will test with their available bandwidth constricted. Overall, out of a large enough population the outliers are washed away.
comScore needs to realize that correlation != causation. It's not because your bandwidth correlates with other users' high-bandwidth plans, that it is caused by you actually buying the plan. But even then, even in the report the statistics show that it evens out pretty good with only a small percentage error.
Off course this brief report reeks more like paid research. Off course comScore measures the users' experience connecting to large-bandwidth centers like Akamai which has a lot of large sites on it and it doesn't accurately measures what the provider offers in the last mile. I don't care that I actually get my 10Mbps connecting to my neighborhood (unless a bunch of my neighbors actually host the Linux-ISO torrent I want) I care about getting on average getting maybe 50% of what I pay for which I usually don't get (I get closer to 1-10% depending on what I'm doing). comScore accurately reflects the poor status of broadband in this metropolitan area - dual-ISDN speeds (early 90's) on the best high-tier packages money can buy in this area. The only alternative is DSL which is horribly outdated.
I could get a decent quad core computer/dual core 64-bit laptop with decent hardware for the price of your license (http://www.microsoft.com/windowsserver2008/en/us/pricing.aspx) and I could even get an Apple machine if I add your hardware, your Server version and whatever you plunked down for your desktop license together.
Slashdot Mirror
Yeah, I tried that before, doesn't work all that well with Windows. With Linux and practically any other OS you can just deploy a generic image with a modular kernel on the system and it will generally work. Network adapters work with a generic driver, video cards work with a generic driver, usb ports work with a generic driver.
With Windows, only the APIC or the boot drive hardware (PATA/SATA/SCSI) have to be different from the original host for the thing to give a blue screen even when sysprepped. Even when the drivers are included and you have an image of a system with the same APIC, the system has been sysprepped but the USB ports aren't the same as whatever machine you made the image off, the system won't be able to react to your input unless all USB hardware has been re-detected (which can take a while and sometimes requires a cold reboot as you can't click on the dialogs). Whenever an update (especially Service Packs) needs to be included in your image, all drivers have to be re-checked (manually) for all your different hardware to make sure none needs to be updated as well. Ideally you would have test-systems, replicas of each piece of hardware you have but even in small organizations this can add up to 10's or 100's of idle hardware that you have to acquire and justify.
I now know why large organizations standardize on a single vendor and can't offer their end-users any choices in hardware besides the amount of RAM and hard drive space. Windows is just plain bad to maintain even with experienced admins. I have virtualized practically all installations of it and even though it takes a slight performance hit, it's much easier to manage than trying to keep up with images for all the different hardware you can have in a single organization.
It should also support native X11, Blackbox, Enlightenment, perlwm, PLWM and WindowMaker.
It doesn't make SSL less secure. It just makes it less trustworthy. You should never trust an SSL certificate to provide you with any identification. SSL's are there to encrypt the connection and so far, that level of protection provided is fairly immune to attacks.
If you want a website to identify themselves, they should provide you with something else, an identifier that is unique to your account with them that you might have established before (as many banks are starting to do - they provide a picture with a non-sense phrase that you established during the account creation procedure).
I would never hire anyone with an MS** certification as they are particularly braindead and responsible for some of the most stupid decisions in any IT organization and responsible for the bad name that most IT departments have when it comes down to costs and ROI.
I run my IT department with maybe 10% of my yearly budget spent on software licensing (most of which is MATLAB and Microsoft Office). There are other departments with very similar end-user requirements where compulsory Microsoft and Oracle software licensing accounts for 50% of their budgets (the other 50% being personnel and hardware) and their total budget at least 300% as large as mine. The difference I put back in hardware. Where most other end-users have to be content with Core2Duo's as their high-end machines, I can afford to give my engineers/scientists dual quad core machines and next week dual hex core workstations.
Yes, ZFS is very simple (the code is very clean and only a few thousand lines - a lot of them comment) but it desperately needs to be ported to other platforms. OpenSolaris is nice but the only reason I use it is ZFS. Imitation ZFS (Btrfs) might eventually become good enough for me to use it but the current feature set of ZFS is very attractive for just about anyone who needs to host multi-GB's of data. Other nice features of OpenSolaris includes their COMSTAR stack (make cheap hardware appear as very expensive iSCSI or FibreChannel targets), zones and dtrace.
Because Amazon is getting paid for their services. Amazon isn't making a loss when criminal syndicates use their services nor are they providing it for free to those organizations. They're probably still pumping cash into the whole EC2 thing since "cloud computing" isn't really as popular and world-changing in most businesses as was projected 5 years ago so they could probably use the $.50/GB at whatever rate these people are pumping out.
Technically, if you throw all of that data online, there is no need to keep the brick-and-mortar buildings open decreasing cost substantially to the already bankrupt state.
CUPS has the ability to support authenticated print jobs and you can lock it down as far as you want to, it's just that nobody uses it (or knows about it) because it's hardly necessary. CUPS is an admin interface, the end-user doesn't have to know about it so it could be at some points a bit arcane to use it if you don't know what you're doing.
The main problem with CUPS and just about any other printer management solution is the drivers. With the advent of inkjet printing and Windows 95, device makers didn't have to support standards anymore, they would just roll their own drivers and Microsoft made it easy for them (and later for more malicious agents) to automatically install crappy software. Olivetti was among the first inkjet printers that didn't have any processing power aboard, leaving all print processing to be done by the host. I think this was one of the reasons they don't exist anymore. Unless you got a laser printer, you just couldn't send a postscript file down to the printer anymore.
These days, people want their printers to connect to their networks so we're starting to see even the cheap printers have their own PostScript processing and support open standards so connecting to it is simple enough even for 'dumber' devices like routers and handheld devices that don't have place for custom printing drivers.
This is typical of short-term thinking on the side of management. Here they are thinking really, really short-term. The people that complain already purchased their copy of whatever game they were selling. There is no benefit in providing them with any more service as they won't pay to cover the costs of it, the profit has been made, the game has been sold in most places you can't return digital media once you opened the shrink wrap without a court order practically.
Depends on how hot she is (http://en.wikipedia.org/wiki/Debra_Lafave)
Well, I guess it depends on where you live then. Everything is technically lethal. However just because somebody can't see that you used a certain weapon does not mean that you should use it more frequently, however, sadly that is what happens - cops (threaten to) taser you when you argue with them in your car when they want to give you a ticket. Ever heard of a cop clubbing somebody because they didn't want to sign their ticket? Or pepper-spray them?
And off-topic, not my quote, it's a shortened SCOTUS-case quote and in that case pertained to the first amendment, not the others (although you could argue that the other amendments should also not be submitted to vote):
"The very purpose of a Bill of Rights was to withdraw certain subjects from the vicissitudes of political controversy, to place them beyond the reach of majorities and officials and to establish them as legal principles to be applied by the courts. One's right to life, liberty, and property, to free speech, a free press, freedom of worship and assembly, and other fundamental rights may not be submitted to vote; they depend on the outcome of no elections." -- Supreme Court Justice Robert Jackson
The problem with the taser replacing the club is that when you hit somebody with a club, everybody knows if you hit hard enough, you can kill somebody, it also leaves bruises. The problem with the taser is that everybody knows 'the company that makes them says they are not lethal at all under any circumstance' and you can't see whether somebody's been hit with a taser. In the beginning of the taser-era, officers would call an ambulance before or after tasering somebody. These days it seems they don't even bother anymore (depending on the type of tasers they use).
The taser has not been tested in any viable study I know off against either human targets or human replica's. As MythBusters and many electrical engineers will tell you, a shock across the heart of just 1mA can kill you, 100mA is lethal. As every geek knows I = V/R and tasers bolt out about 50-100kV which means your skin-to-heart resistance needs to put up a resistance of 500k (if my calculations are correct). Your body resistances ranges anywhere from 300ohms-6Mohms. When your skin is moist (sweating, ...) as is common with drug-addicts and people running from the law your body's resistance will drop. If a taser hits you near the heart in those conditions, they can be theoretically very lethal.
That's because of the nature of an Unidentified Flying Object. It flies through the sky but is too far away or obscured by other things (weather, buildings) to be identified. UFO's will always remain UFO's no matter whether they be nature-made, man-made or alien-made phenomenon.
That's why we should legalize all drugs and have them sold in drugstores (also called pharmacies). Make it illegal to sell to underage children and make it illegal to drive/work under the influence of any drug. This would take care of violent drug cartels whom are only fueled by the high prices the underground trading commands. Making it legal takes the risk out of it, drops the prices and leaves those drug dealers out of a job.
People will get high somehow, people will kill themselves with it just as they do with alcohol right now, nobody cares that somebody drinks themselves to death or that they can't keep a job because they're drunk all the time - they made that decision, they will bear the consequences, they will eventually die.
Everybody "peers" at some point. As an end-user (and most website owners too) we don't have enough money to peer with any of the big players. That's why we peer with our ISP's (or host our servers at somebody else's datacenter) to provide this peering. We don't necessarily pay for bandwidth (as those agreements may vary from dedicated bandwidth pipes to pay-per-bit bandwidth) but we all pay to peer to other networks.
Google is big enough that they can peer themselves. They chip in at the point of the core routing infrastructure just like all the other ISP's, datacenter etc. Google is it's own private ISP. When ISP's want Google to pay for bandwidth usage on their networks, they would (if they were honest) also have to start charging/paying each other to use each other's bandwidth whether it be their customers BitTorrenting between each other or connecting to a server in a datacenter.
Do you know what a web app firewall actually does? It's practically a proxy server that sanitizes the inputs, meaning for each application you have to make custom rules which are more involved than just "allow 0.0.0.0/0 to 192.168.1.1 port 80". Especially if you have a very custom-built application, for each data input (each field) you will have to specify what form of data it is and what the valid ranges are. This SHOULD be done within the application already by the developers, there is no reason to go out of your way to buy software and practically rebuild the whole front-end and validation of each of those pages, you might as well build the back-end as well and just do the whole application yourself.
(Bad) sequels being written or story lines being hijacked has nothing to do with copyright and probably can't even be protected by it (since it's not the same story). After all, the original author is dead. The only way you can show your disappointment is by not buying those books, eventually the author will have to move on to fresh content or change jobs where imagination is not required as much.
You will also notice that these 'pledges' don't do very much in the long run. IBM, Google, Microsoft, Novell, Red Hat and Apple all have 'pledged' some type of protection for their open source ancestors but those things are not legally binding no matter what they might say about it (I think it's MSFT that has such a claim).
That's also why you should avoid implementing any of their proprietary crap in your Open Source project (or any project that's being made public or sold in any shape or form) because if for any reason they want to leverage their arbitrary licenses on it, they can no matter what they have promised whether it's uncanny legal speak or so-called patent pools.
The main problem with most compliance protocols (HIPAA or PCI) is that at best they do nothing at all, at worst it's actually counterproductive as it opens the company up to more breaches (due to human nature, laziness or conflicting policies).
I am involved in both HIPAA and PCI compliance and in the past I have been involved with Sarbanes-Oxley as well. For example with PCI as well as Federal wiretapping compliance, you need to have your respectively wireless and public networks (if you're a de-facto wireless internet provider to random strangers - eg. libraries, universities, ...) run through a separate (3rd party) provider and needs to be either logically or physically divided from the main network. Therefore, anyone on your public or wireless network will have to tunnel a VPN through a 3rd party provider, route it out to the internet and back into your primary provider to get work done which makes the whole system inherently less secure because your data goes outside your network.
PCI requires a firewall before your internet facing servers but also a perimeter firewall (if you have a really large institution) before all your edges even though you may have separate departmental firewalls. This does not make sense as you get to have 2 or 3 layers of firewalls - the first 2 layers being the ones that were historically built-up and the 3rd layer, a concentrated firewall and internet provider hub which becomes 1) easier to attack because it's all in one point, 2) easier to fail for the same reason, 3) more difficult to maintain because you still need the hierarchy of departmental firewalls to prevent attacks from other departments or other points in the network.
UPS/FedEx? That's their business. The USPS (and a lot of other postal systems throughout the world) has been putting more emphasis on their 'added services' like package shipping and money transfers to adjust for the diminishing returns of snail mail delivery. I actually get mad when people send me snail mail letters on dead tree - it's bad for the environment and e-mail is a lot simpler and easier.
Most people do get their licenses taken away after the first or second one. I don't know in other states but I know in NY if you get your license taken away, it's quite difficult to get it back. Off course, then you get the problem of people driving around WITHOUT licenses.
Sun died, not because it gave away stuff but because it overpriced stuff and made some really bad decisions. A Thumper costs about 50-100k and all it is, is an AMD Opteron-based machine with some SATA controllers that fits lots of disks. I can get the same hardware for 10-20k elsewhere. Same goes for their 1U servers - I believe it's 10k for an entry-level model. They have some really nice CPU's though (Niagara now) but they never marketed it right, never priced it right and management killed some really nice projects.
The same with ZFS. They never marketed it as a company but the developers did market it to other developers and the open source community which made it very popular. ZFS is simply awesome and if upper management would've seen what this project does, they could've easily been taking over NetApp (where their buyers don't seem to mind spending 20k/TB) and other high-end proprietary storage vendors while staying safe from a hostile takeover.
I am building a dual-parity, redundant 30TB storage array with 320GB read cache and 64GB write cache (POSIX-correct) for under 30k with ZFS. Sun could have easily charge 300k for something like this and most customers would be grateful not to pay NetApp or EMC for it.
But when they do get CPU-bound, you have serious problems. I don't know if things changed since 2000/2003 but I remember very well when 6 geographically separated, load-balanced MSSQL boxes (with 8 cores each back then - very costly setup) hit the CPU bound, they all became unresponsive for a couple of seconds (dropped down to 0% CPU), came back, did a few thousand queries and then repeated the cycle every 5 seconds.
Both sides need to learn more about statistics.
The report fails to mention that across a large enough population, the results will be more-or-less correct within a certain percentage point because as he mentions, some people will test with a lot of bandwidth available at a certain point but others will test with their available bandwidth constricted. Overall, out of a large enough population the outliers are washed away.
comScore needs to realize that correlation != causation. It's not because your bandwidth correlates with other users' high-bandwidth plans, that it is caused by you actually buying the plan. But even then, even in the report the statistics show that it evens out pretty good with only a small percentage error.
Off course this brief report reeks more like paid research. Off course comScore measures the users' experience connecting to large-bandwidth centers like Akamai which has a lot of large sites on it and it doesn't accurately measures what the provider offers in the last mile. I don't care that I actually get my 10Mbps connecting to my neighborhood (unless a bunch of my neighbors actually host the Linux-ISO torrent I want) I care about getting on average getting maybe 50% of what I pay for which I usually don't get (I get closer to 1-10% depending on what I'm doing). comScore accurately reflects the poor status of broadband in this metropolitan area - dual-ISDN speeds (early 90's) on the best high-tier packages money can buy in this area. The only alternative is DSL which is horribly outdated.
I could get a decent quad core computer/dual core 64-bit laptop with decent hardware for the price of your license (http://www.microsoft.com/windowsserver2008/en/us/pricing.aspx) and I could even get an Apple machine if I add your hardware, your Server version and whatever you plunked down for your desktop license together.