I'm not even going to look at the 2650 because it's a 2U and thus isn't even in the same class. As Apple rolls out more servers it'll make sense to compare like to like.
The Dell 1650 with similar hard disk size, rails, an actual OS with Media (RedHat 7.2) and comparable NICs is $2652. You're right. If you're running Linux, the Dell is $350 cheaper. Take out the Linux and put in a 5 user Windows license and Xserve wins by a little less than $300. Take a look at a 25 user scenario and the Xserve wins by almost $2800.
Clearly, if the shop deploying the servers has ease of use issues with Linux, Xserve is the cheaper solution compared to Windows.
Way back at the top of the thread, the allegation was made that XServe is $1000 more expensive. It isn't. Rails, OS (with media), like networking options, etc. are going to erase much of the difference. For a lot of people, it seems worth it. For many others it won't be but to say that it isn't a credible offering is zealotry of its own.
In case you want to cry foul over the comparisons, here's the Dell options I chose. Processor = 1.26Ghz PIII (standard, didn't change) Rails = 4 post non-dell rails Memory's left at the minimum 256Mb No Direct Line, standard warranty service standard cdrom, power supply, PCI riser. no keyboard, monitor, or mouse The closest storage option was the 72Gb HD used 1 with no fancy controllers. I changed the NICs to 2 broadcom gigabit nics.
I could easily have eaten up those last $350 to make the Dell more expensive even using Linux. I was fair and didn't.
Funny, that wasn't there before. Oh well. I wouldn't recommend giving this machine because of the lack of floppy (splurge the $20 and add it). Thanks for supporting my argument.
It really does depend on how much this guy's time is worth. If he's out there billing at $80 an hour, its only a few tech support issues with Win95 before it's worth it for him to buy the thing for them. Word processors are at least $100 with some 'electronic typewriters' at $200 so if he kicks in a couple of hundred, the problem's quite nicely solved.
Another thing, Lindows is working with an OEM partner to drop the price to $199. At a certain point, it doesn't make sense to give an old machine to relatives.
This isn't something for an individual band to do but creating a music distribution system via kiosks is certainly possible. Look at Walmart. They're distributing Lindows because they don't care about MS' threats. Similarly, indie music needs real world floor space in a setting that doesn't care about the RIAA. I suggest Starbucks. It's everywhere, it sells Internet connectivity so you just need to rent a square yard of floor space, put in an e-commerce kiosk that has a CD burner and printer in it (for CD and case jacket) and keep it stocked with consumables. This would be a great franchise operation for MP3.com or competitor. Starbucks gets to have increased traffic and steady rent payments so the Starbucks franchise owners are likely to go for it.
Sure the technical problems aren't solved but I'm sure somebody could make a quick open source project to fix all that...
In the end, you have a generalized real world kiosk that can be used as a distribution conduit for music.
Then again, it can be used for software too so you have a two-fer bonus.
Getting Samba to run is one thing, integrating a linux box into your directory service infrastructure so that HR can continue to add and remove computer rights seamlessly throughout the enterprise as they hire and fire is something slightly more complicated.
Supermicro doesn't seem to sell a comparable configuration. Their 1 unit racks seem to all have 10/100 NICs. The Xserve configs all have dual gigabit connectors. They also don't have two full length PCI slots. As you noted, they also don't have 4 bays.
Apple's got a nice offering that's got some unique features. For people who care about space and bandwidth, you're going to be able to stuff more bits and move them in and out faster on the XServe.
Supermicro doesn't sell direct just through distribution so you're very much at the mercy of your reseller. Apple's service options are much more under the control of Apple directly and they seem to do well in the serice quality surveys year after year.
Apple certainly hasn't filled out their server line but for the market that it's targetted at (and that doesn't mean you, you've made it clear) it's a pretty good value proposition.
I was thinking something similar but I would tell them to go to Walmart.com and get the bottom end Lindows machine. You get a new machine, it's faster, and plenty of people are using it so they should be able to get support.
By contrast, Windows 95 is in end of life so it (and anything older like the non y2k compliant Win 3.1 somebody else suggested) isn't going to get fixed for any bugs found later, it's not particularly safe to put on the net if they decide to expand their computing horizons later and it's running on a slow, old machine which could start having hardware failures at any time.
At today's prices ($299) a new machine just makes sense.
Here's a scenario, Microsoft house has been pirating CALs. They undergo an audit (perhaps the company is thinking of going public) and they find out they need $15k in licensing fees for file and print on two servers. There is no $15k in the budget to do this and going to periodic payments for those extra licenses leads the shop to move to Linux. The IT department can't stomach linux due to its raw edges and they migrate to Xserves for file and print and shift the file and print servers to some sort of number crunching app server or a dev server that requires no additional windows license.
Another scenario is in mixed shops that get an Xserve to satisfy the graphics guys and the IT departmant falls in love with it on cost (v. Windows, remember) and ease of use grounds. It spreads.
Congratulations, you're not the target market. AFAIK, Apple's trying to compete with Dell, IBM, HP, etc. not with Joe Blow white box vendor down the street. I believe that most of the hardware price difference goes away if you compare only national brands and in a few configurations, the X-Serve is cheaper.
I always thought that the largest segment of users would be Windows refugees who use it for file and print services in unlimited license mode. For that scenario (and there are lots of people who need this), the XServe is *much* cheaper while the administrative ease means you don't need to radically upgrade your IT department skills.
You should give MS a call only to get a written statement of their HIPAA compliance assertion. Anything else should be handled by your own lawyers and techs. I wouldn't trust MS's legal department for unbiased or even accurate opinion. I would only trust them to convey legally accurate information as strictly required by law.
When MS starts doing medical care, it will, indeed, be bound by HIPAA. But MS doesn't provide medical care so it is only trivially bound by it. It provides general purpose computing software to run on x86, PPC, and Sparc systems running Windows, Mac OS, and Solaris software in that order. The people who buy this stuff in a medical environment are bound by HIPAA and must cease using systems that are not compliant by certain dates.
MS may be compliant with HIPAA or it may not be. If it wants to remain in the medical computing field in the US, it can either create a HIPAA compliant EULA or it can exit the field. It's up to them.
What medical facilities can't do is purchase and continue to use MS (or any other vendor's) software that is non-compliant with the HIPAA rules.
Other posters have made it clear that the HIPAA rules are not very clear. This means that in the relatively near future there's going to be some serious people like Blue Cross/Blue Shield, Kaiser, and several insurance firms that are going to petition for a ruling and MS licensing is either going to be ruled compliant or not. The medical software field may have to do some heavy shifting if an impasse is reached but that's the breaks.
Anytime you see an MS station in a medical setting from now on, admire their bravery for bucking the HIPAA rules.
Doctors don't give a damn about MS. They certainly don't understand that they might be non-compliant and subject to lawsuit. They will drop MS in a second if they find out that this is true.
Btw: My wife is a doctor that's newly entered practice and she's been having me look up this sort of stuff to educate her.
Unless the early days of ISA and PnP ISA include last year, there are still ocassional problems. Pulling a card, restarting, and then reseating a card to get it recognized should have no place in a modern operating system environment. That was on a Win2K system which was a sealed box with an extended warranty so it had to go into the shop to do it.
They're trying to move all configuration files to XML so everything can get managed with one simple tool. Now *that's* something that advances the state of Unix and workalikes. I believe NetInfo is also available in Darwin so the source is available for reimplementation elsewhere.
It's a good thing that Apple will be able to get IBM Power4lite chips soon. Motorola is a dog, true but it's not the only possible supplier of PPC chips. IBM's got a very valuable Power4 franchise and the volume off of their own RS/6000 machines will make PPC a priority for them for the indefinite future.
The truth is that 'just working' is mostly a function of whether or not you use a PC-type BIOS or you go for Open Firmware. If Intel and Microsoft came out and said that their next PC hardware reference platform included open firmware, the IRQ hassle would go away and MS could actually make very good plug and play support for Windows.
Shame they don't do it but they're deathly afraid of innovating here because of the legacy problem.
I'm not going to prove the world is round. There are plenty of objective studies out there that are *not* done by parties with an axe to grind. If you are genuinely curious and are not coming out with an agenda, you'll spend a few minutes on google and come across them.
If you taught statistics, you should certainly know the fallacy of anecdotal evidence. The fact is that there is widespread dissatisfaction with the public school system and objectively that school system has soaked up a lot of resources. Private schools often do more with less resources (especially the religious ones). Home school children are showing up as winners in national academic contests far out of proportion to their presence in the relevant population (home schooling is still under 1% nationally).
The NEA and AFT have been putting out FUD for years regarding home schooling after their previous campaigns of legal intimidation and storm trooper tactics didn't pan out. One by one their excuses for poor performance in the public schools have been shot down. The socialization of home school children is pretty much the last one. Don't buy into it.
But is it a gift? If you give something, you have passed ownership and your entire transaction income is $0. Everybody knows that Microsoft claims to never pass ownership to anybody. They license their software and they're obligating New Orleans to future purchases so they will get income from this. So it sounds like they are printing out a massive volume license agreement with up front discounting down to zero. If New Orleans *has to* purchase MS software under the contract, money will change hands at a future date and it's just an extended credit line we're talking about. After all, if you offer net 60 credit on a series of purchases, it's not considered a gift either.
The big payoff is when the new IT infrastructure goes MS only. When you can do much of your civic duties/responsibilities via computer (as is the likely future), it's important to make sure that you're not forcing everybody onto one system. MS has an extensive history of using loss leaders to get market share and eventual lock-in at which point your freebies stop and they make all that money back and more. Using New Orleans as a loss leader to get 50 other cities immediately, and New Orleans itself later isn't a bad strategy for MS, just a bad strategy for New Orleans.
It's funny that when people actually do studies, the socialization scores of home schoolers end up right in the mainstream of the student population but their academic scores end up superior to the public school population. I guess you don't buy into those statistically valid samples and peer reviewed methodologies.
The problem of academically gifted, socially awkward students is something that plays out every day in schools with the 'jocks' mocking the 'geeks' being a recurring theme. The question isn't whether a student is awkward, that happens in all systems. The real question is whether the incidence of awkwardness is within normal paramaters.
As for curricula, yes, they do change. When I was in primary and secondary school we didn't have Gaia advocates and PETA shills coming in to the schools. My younger cousin did. Heather has two Mommies is also a recent elementary school innovation.
The point is that changing curricula isn't necessarily good or bad. The idea of creating an open curricula project that can be forked if it beaches on the shoals of controversy is a very salutory idea. Math is math and when you see declining test scores over several cycles of curricula 'improvement' (see the mid 60s through mid 80s) then something is severely wrong.
The problem is that amateurs are producing superior academic results to the professionals. Social interaction is something you can get by having a big building and funneling a lot of kids there. It takes precious little skill on the part of teachers and administrators. The academic skills are the metrics that are most influenced by teachers and in the areas where home schooling is breaking out, those teacher influenced metrics are atrocious.
Slashdot Mirror
I'm not even going to look at the 2650 because it's a 2U and thus isn't even in the same class. As Apple rolls out more servers it'll make sense to compare like to like.
The Dell 1650 with similar hard disk size, rails, an actual OS with Media (RedHat 7.2) and comparable NICs is $2652. You're right. If you're running Linux, the Dell is $350 cheaper. Take out the Linux and put in a 5 user Windows license and Xserve wins by a little less than $300. Take a look at a 25 user scenario and the Xserve wins by almost $2800.
Clearly, if the shop deploying the servers has ease of use issues with Linux, Xserve is the cheaper solution compared to Windows.
Way back at the top of the thread, the allegation was made that XServe is $1000 more expensive. It isn't. Rails, OS (with media), like networking options, etc. are going to erase much of the difference. For a lot of people, it seems worth it. For many others it won't be but to say that it isn't a credible offering is zealotry of its own.
In case you want to cry foul over the comparisons, here's the Dell options I chose.
Processor = 1.26Ghz PIII (standard, didn't change)
Rails = 4 post non-dell rails
Memory's left at the minimum 256Mb
No Direct Line, standard warranty service
standard cdrom, power supply, PCI riser.
no keyboard, monitor, or mouse
The closest storage option was the 72Gb HD used 1 with no fancy controllers.
I changed the NICs to 2 broadcom gigabit nics.
I could easily have eaten up those last $350 to make the Dell more expensive even using Linux. I was fair and didn't.
Funny, that wasn't there before. Oh well. I wouldn't recommend giving this machine because of the lack of floppy (splurge the $20 and add it). Thanks for supporting my argument.
It really does depend on how much this guy's time is worth. If he's out there billing at $80 an hour, its only a few tech support issues with Win95 before it's worth it for him to buy the thing for them. Word processors are at least $100 with some 'electronic typewriters' at $200 so if he kicks in a couple of hundred, the problem's quite nicely solved.
Another thing, Lindows is working with an OEM partner to drop the price to $199. At a certain point, it doesn't make sense to give an old machine to relatives.
This isn't something for an individual band to do but creating a music distribution system via kiosks is certainly possible. Look at Walmart. They're distributing Lindows because they don't care about MS' threats. Similarly, indie music needs real world floor space in a setting that doesn't care about the RIAA. I suggest Starbucks. It's everywhere, it sells Internet connectivity so you just need to rent a square yard of floor space, put in an e-commerce kiosk that has a CD burner and printer in it (for CD and case jacket) and keep it stocked with consumables. This would be a great franchise operation for MP3.com or competitor. Starbucks gets to have increased traffic and steady rent payments so the Starbucks franchise owners are likely to go for it.
Sure the technical problems aren't solved but I'm sure somebody could make a quick open source project to fix all that...
In the end, you have a generalized real world kiosk that can be used as a distribution conduit for music.
Then again, it can be used for software too so you have a two-fer bonus.
Getting Samba to run is one thing, integrating a linux box into your directory service infrastructure so that HR can continue to add and remove computer rights seamlessly throughout the enterprise as they hire and fire is something slightly more complicated.
Supermicro doesn't seem to sell a comparable configuration. Their 1 unit racks seem to all have 10/100 NICs. The Xserve configs all have dual gigabit connectors. They also don't have two full length PCI slots. As you noted, they also don't have 4 bays.
Apple's got a nice offering that's got some unique features. For people who care about space and bandwidth, you're going to be able to stuff more bits and move them in and out faster on the XServe.
Supermicro doesn't sell direct just through distribution so you're very much at the mercy of your reseller. Apple's service options are much more under the control of Apple directly and they seem to do well in the serice quality surveys year after year.
Apple certainly hasn't filled out their server line but for the market that it's targetted at (and that doesn't mean you, you've made it clear) it's a pretty good value proposition.
I was thinking something similar but I would tell them to go to Walmart.com and get the bottom end Lindows machine. You get a new machine, it's faster, and plenty of people are using it so they should be able to get support.
By contrast, Windows 95 is in end of life so it (and anything older like the non y2k compliant Win 3.1 somebody else suggested) isn't going to get fixed for any bugs found later, it's not particularly safe to put on the net if they decide to expand their computing horizons later and it's running on a slow, old machine which could start having hardware failures at any time.
At today's prices ($299) a new machine just makes sense.
Here's a scenario, Microsoft house has been pirating CALs. They undergo an audit (perhaps the company is thinking of going public) and they find out they need $15k in licensing fees for file and print on two servers. There is no $15k in the budget to do this and going to periodic payments for those extra licenses leads the shop to move to Linux. The IT department can't stomach linux due to its raw edges and they migrate to Xserves for file and print and shift the file and print servers to some sort of number crunching app server or a dev server that requires no additional windows license.
Another scenario is in mixed shops that get an Xserve to satisfy the graphics guys and the IT departmant falls in love with it on cost (v. Windows, remember) and ease of use grounds. It spreads.
I'm sure there are other scenarios out there.
Congratulations, you're not the target market. AFAIK, Apple's trying to compete with Dell, IBM, HP, etc. not with Joe Blow white box vendor down the street. I believe that most of the hardware price difference goes away if you compare only national brands and in a few configurations, the X-Serve is cheaper.
I always thought that the largest segment of users would be Windows refugees who use it for file and print services in unlimited license mode. For that scenario (and there are lots of people who need this), the XServe is *much* cheaper while the administrative ease means you don't need to radically upgrade your IT department skills.
You should give MS a call only to get a written statement of their HIPAA compliance assertion. Anything else should be handled by your own lawyers and techs. I wouldn't trust MS's legal department for unbiased or even accurate opinion. I would only trust them to convey legally accurate information as strictly required by law.
When MS starts doing medical care, it will, indeed, be bound by HIPAA. But MS doesn't provide medical care so it is only trivially bound by it. It provides general purpose computing software to run on x86, PPC, and Sparc systems running Windows, Mac OS, and Solaris software in that order. The people who buy this stuff in a medical environment are bound by HIPAA and must cease using systems that are not compliant by certain dates.
MS may be compliant with HIPAA or it may not be. If it wants to remain in the medical computing field in the US, it can either create a HIPAA compliant EULA or it can exit the field. It's up to them.
What medical facilities can't do is purchase and continue to use MS (or any other vendor's) software that is non-compliant with the HIPAA rules.
Other posters have made it clear that the HIPAA rules are not very clear. This means that in the relatively near future there's going to be some serious people like Blue Cross/Blue Shield, Kaiser, and several insurance firms that are going to petition for a ruling and MS licensing is either going to be ruled compliant or not. The medical software field may have to do some heavy shifting if an impasse is reached but that's the breaks.
Anytime you see an MS station in a medical setting from now on, admire their bravery for bucking the HIPAA rules.
Doctors don't give a damn about MS. They certainly don't understand that they might be non-compliant and subject to lawsuit. They will drop MS in a second if they find out that this is true.
Btw: My wife is a doctor that's newly entered practice and she's been having me look up this sort of stuff to educate her.
Great thread.
Since this effectively cripples your ability to maintain a secure box, when are you going to start migrating off MS in order to maintain your NDAs?
Unfortunately, I can't seem to get ATLAS, a required prerequisite up on OSX. Apparently, the code doesn't like G3 processors.
Unless the early days of ISA and PnP ISA include last year, there are still ocassional problems. Pulling a card, restarting, and then reseating a card to get it recognized should have no place in a modern operating system environment. That was on a Win2K system which was a sealed box with an extended warranty so it had to go into the shop to do it.
They're trying to move all configuration files to XML so everything can get managed with one simple tool. Now *that's* something that advances the state of Unix and workalikes. I believe NetInfo is also available in Darwin so the source is available for reimplementation elsewhere.
Here's a few more scenario's for Mac superiority
Biotech desktop running BLAST (Altivec makes BLAST much faster than x86 variants)
Getting a computer for mom, dad, and the grandparents
Running a file and print server cheaply with no not very savvy IT support and without pirating.
I'm sure there are lots of others.
It's a good thing that Apple will be able to get IBM Power4lite chips soon. Motorola is a dog, true but it's not the only possible supplier of PPC chips. IBM's got a very valuable Power4 franchise and the volume off of their own RS/6000 machines will make PPC a priority for them for the indefinite future.
The truth is that 'just working' is mostly a function of whether or not you use a PC-type BIOS or you go for Open Firmware. If Intel and Microsoft came out and said that their next PC hardware reference platform included open firmware, the IRQ hassle would go away and MS could actually make very good plug and play support for Windows.
Shame they don't do it but they're deathly afraid of innovating here because of the legacy problem.
10.3 is supposed to be code named puma.
Sorry you didn't get it. I'm on the same side as your point.
I'm not going to prove the world is round. There are plenty of objective studies out there that are *not* done by parties with an axe to grind. If you are genuinely curious and are not coming out with an agenda, you'll spend a few minutes on google and come across them.
If you taught statistics, you should certainly know the fallacy of anecdotal evidence. The fact is that there is widespread dissatisfaction with the public school system and objectively that school system has soaked up a lot of resources. Private schools often do more with less resources (especially the religious ones). Home school children are showing up as winners in national academic contests far out of proportion to their presence in the relevant population (home schooling is still under 1% nationally).
The NEA and AFT have been putting out FUD for years regarding home schooling after their previous campaigns of legal intimidation and storm trooper tactics didn't pan out. One by one their excuses for poor performance in the public schools have been shot down. The socialization of home school children is pretty much the last one. Don't buy into it.
But is it a gift? If you give something, you have passed ownership and your entire transaction income is $0. Everybody knows that Microsoft claims to never pass ownership to anybody. They license their software and they're obligating New Orleans to future purchases so they will get income from this. So it sounds like they are printing out a massive volume license agreement with up front discounting down to zero. If New Orleans *has to* purchase MS software under the contract, money will change hands at a future date and it's just an extended credit line we're talking about. After all, if you offer net 60 credit on a series of purchases, it's not considered a gift either.
The big payoff is when the new IT infrastructure goes MS only. When you can do much of your civic duties/responsibilities via computer (as is the likely future), it's important to make sure that you're not forcing everybody onto one system. MS has an extensive history of using loss leaders to get market share and eventual lock-in at which point your freebies stop and they make all that money back and more. Using New Orleans as a loss leader to get 50 other cities immediately, and New Orleans itself later isn't a bad strategy for MS, just a bad strategy for New Orleans.
It's funny that when people actually do studies, the socialization scores of home schoolers end up right in the mainstream of the student population but their academic scores end up superior to the public school population. I guess you don't buy into those statistically valid samples and peer reviewed methodologies.
The problem of academically gifted, socially awkward students is something that plays out every day in schools with the 'jocks' mocking the 'geeks' being a recurring theme. The question isn't whether a student is awkward, that happens in all systems. The real question is whether the incidence of awkwardness is within normal paramaters.
As for curricula, yes, they do change. When I was in primary and secondary school we didn't have Gaia advocates and PETA shills coming in to the schools. My younger cousin did. Heather has two Mommies is also a recent elementary school innovation.
The point is that changing curricula isn't necessarily good or bad. The idea of creating an open curricula project that can be forked if it beaches on the shoals of controversy is a very salutory idea. Math is math and when you see declining test scores over several cycles of curricula 'improvement' (see the mid 60s through mid 80s) then something is severely wrong.
The problem is that amateurs are producing superior academic results to the professionals. Social interaction is something you can get by having a big building and funneling a lot of kids there. It takes precious little skill on the part of teachers and administrators. The academic skills are the metrics that are most influenced by teachers and in the areas where home schooling is breaking out, those teacher influenced metrics are atrocious.