Note also that it's highly likely the Dell prices would be discounted further after speaking with a sales rep, particularly if multiple units are being purchased. The chances of getting any discount from Apple - let alone a significant one - is basically zero.
Tell that to O.J.'s lawyers and anybody else who wants to claim, or actually believes, that the police might want to tamper with evidence to frame them. Or tell that to all the commentators on fark who are certain that the police would delete evidence to help protect one of their own. You should understand that these kind of enterprise storage systems that protect data from any kind of tampering (and "losing" data is a kind of tampering) even by system administrators are intended to be used in lieu of mountains of WORM media. When you've properly designed for their use they offer a worthwhile benefit.
I'm well aware of how the systems work. My first point was that "modification" and "deletion" are different operations, and allowing one by disallowing the other should be a minor technical issue. My second point was that with proper backups and auditing, it should be possible to implement deletion functionality without any need to compromise the trust of the system.
And don't say it was the application developers' fault; it's very hard to know the requirements of an operating system that will come out years in the future when you're writing your application.
It *is* the developers' fault, completely and utterly. They ran out of excuses not to write "UAC-friendly" software back around 1998 or so.
They are now the running Server 2008 kernel inside Win 7, and that is not a light upgrade, or even a service pack upgrade. That is a complete redesign of the core architecture.
It's nothing of the sort.
Vista and Server 2008 had the same major kernel revision.
Windows XP x64 and Windows 2003 are the same major kernel revision.
Windows 2000 (and earlier) Workstation and Server are the same major kernel revision.
Finally, they're *all* just different revisions of the same core architecture - Windows NT.
And yet they could have used the Internet for feedback well more than a decade ago.
*They* could have. Their customers, on the other hand, probably couldn't. In 1995, less than 10% of people would have had internet connections. Even ca. 1997-98, it was probably only around 30%.
Absolutely. Intel looked like they were going to start producing "decent" integrated graphics when, in 2004, they announced the GMA 900. It looked to Microsoft like the world's largest GPU maker would finally have something capable of desktop compositing, so they figured they could finally add this capability to Windows without a huge performance hit.
"Desktop compositing" was in the plan long before that. It first appeared in Longhorn in 2002.
Once the data is written, it can not be modified or deleted. Now, the reasoning behind this is so the police can not digitally manipulate the timestamps or data in any way. This is to protect the integrity of the data so it can withstand legal challenges.
Deleting and modifying are two distinctly separate things. It shouldn't be a problem for a system to allow deletions but not modifications.
Combined with proper backups and auditing, that should be sufficient to retain the legal value of any digital data.
That's a reason to make each respective piece of software _workable_. But there is no incentive to make it actually work as well on the foreign platform as it does on the native one...
There aren't many ways one could define Office for Mac as "foreign".
The discussion was comparing blades to minis on the basis of cost, not density, but if you want to talk density...
Well, the cost argument works just as well for rackmount machines. An appropriately specced 1U rackmount is about $6.5k. 8 Mac Minis is about $8k.
I've just measured the rack in my garage - it's 30" deep by 19" across, and a mac mini fits handily into a 7" square by 4 holes in the rack with space to spare. 6U is 18 holes, so I can get 4 slide-out racks of minis into 6U, and 8 minis (including PSUs and wiring) per rack in a 2-across by 4 deep configuration. That's a total of 8x4x2 CPUs (because they're core-2 duos in the mini), or 64 CPUs.
Yes. I estimated you'd get about 8 Minis per 2U. You've measured it out at about 8 Minis per 1.5U - though I suspect you'd have trouble keeping the density you're talking about properly cooled, given the lack of high-volume forced cooling in the Mini. The ones at the upper rear of the rack, in particular, would suffer.
The blades pack 8 blades into 6U. The ones we started off discussing therefore put 8 CPUs into 6U. Whoosh. If we jump to the later offering, that's 8x4=32 cores in 6U, or 32 less than the minis. Only if we go to the 8-core blades (for a total of 4299+8*1610 for the E5405 CPUs, or $17179 before you add OS and disk storage) do you get the same density of 64 CPUs in 6U.
I didn't see the blades you were referring to earlier. A dell M10000e, or HP's c-Class, will put 16 blades into 10U, slightly higher density than the ones you are talking about.
The largest number of cores I'm aware of in standard 1U is dual quad-core (soon to be dual 6-core I suppose), so in 6U you get 6x8 = 48 CPUs. That doesn't work out too well compared to 6U of minis (64 CPUs).
Actually, it works out extremely well. Those rackmount servers have Nehalem based CPUs, which are roughly twice as fast as Core 2 CPUs at the same clock speed. Even being conservative, the 48 cores in 6U of rackmount servers will be quite a bit faster than the 64 cores in the Minis. The same is true of using Xserves, though the price comparison probably won't be quite as favourable.
I'm not advocating that you build your next computing cluster out of mac minis. I'm saying they scale (in price) reasonably well, that's all. I assume that there's a design difference between a blade server and a more consumer-orientated server (though I've been using a mini as an entertainment hub for the last 2 years without a single problem).
The thing is, a Mac Mini is not designed to be a 24/7 server. It also lacks numerous "standard" server features like power supply redundancy, remote management capabilities, and hot-swap hard disks. A Mini comes with a 1 year warranty standard, while even a low-end Dell server will come with a 3 year warranty. A Mini is difficult and time-consuming to service. A typical rackmount is trivial to break down and nearly any arbitrary component can be replaced in a matter of minutes.
If you need a low-end Mac server (or two), then a Mac Mini is certainly a reasonable option. If you just need "a server", though, and particularly if you need a lot of bang for your buck, then the Mini just doesn't match up. 1U rackmounts are a better deal up until about 8-10U, and after that the ~13 cores/RU (and Nehalem cores at that) blades provide is untouchable - and that's even before getting into the reliability and manageability aspects. There's no way those "Mac Mini farms" can be competitive in the face of Nehalem-powered blades and virtualisation, except for the handful of people who genuinely need a Mac.
Yes, and an 8-core XServe is 1U as well, but that's also out of scope of the discussion...
Why ? The context of the discussion is density. An example given was 8 Mac Minis. A 1U server will do as much as 8 Minis, more reliably, in less space, for less money.
Heck, your last post was comparing *blade servers*. How can blade servers be relevant and 1U rack servers not ?
I was using 8's because that's the unit of the enclosure. If you want to start using more blades, the 16-way enclosure is even more expensive, relatively, with prices for just the enclosure starting at $17k. Just multiply up to suit the numbers. I think you'll easily fit 4 mini's within a 6U rack-space though - you could probably fit them on a 2U tray...
However, a single 1RU server will fit in less space than 8 Minis and provide similar performance with all the benefits of the blades.
Overall, the mini comes out looking pretty good:)
Only if your requirements are low. If you need lots of capacity, 1U servers will comfortably provide ca. 1.5x the density, and blades will comfortably provide 2x the density, and either will provide a _vastly_ superior physical package.
You can cramp LOADS of those things in a small space and have massive storage and crunching power without needing an entire room.
Standard 1U rackmount servers have density at least as good (and generally better). Blades have (depending on vendor) up to 60% better density than that.
Not to mention real server hardware having vastly better cooling dynamics, substantially longer warranty, remote access and management tools and actually being designed to run 24/7.
20% cheaper, half the physical footprint (the Minis + a shelf to hold them will take 2 RU), more redundancy, vastly better cooling dynamics, actually designed to run 24/7 and 3 times the warranty. I know where I'd be spending my money.
I was quoting your remark about how "unusual" registry editing was in windows and you start calling me names. I guess that says it all about you and your 'arguments'.
No, you were being facetious. The discussion is (or was) about editing the Registry directly.
I will repeat it (you just try to read it slowly so that your synaptic pathways do not get damaged in the meantime): if you use windows, you edit the registry all the time (even if you do not know that you do so).
Great, but that's not what the discussion was about. Ergo, trying to equate it to what the discussion *was* about, is disingenuous.
Windows 7 was brought forward by 2 years at least, MS usually expects their OSs to last 5-7 years between updates.
Where do you get this number from ? The average time between major Windows NT releases is ~2.5 years, as you can see here. Windows 7 was pretty much bang on schedule. Even Vista wasn't really "late", when you consider it was mostly late in a marketing, rather than development, sense (much of the early work into "Vista" - then Longhorn - appeared in the market as Windows 2003, XP SP2 and 2003 R2).
NT was a cleanr OS, but they did graft the Win95 UI on top of it, and move the graphics drivers into the kernel to support it with enough speed.
They reimplemented the Windows 95 UI on top of it, yes. Why is this a bad thing ? Why is it any more "grafted" than, say, OS X's GUI ?
Also, most OSes run video drivers of some description in kernel space for performance.
I can't say that they broke more of it in Vista, but I wouldn't be surprised given the amount of changes they made to everything for Vista.
You haven't even managed to describe anywhere that they started "breaking" it, let alone did "more".
.NET doesn't intercept Win32 calls, it is mostly a layer on top of Win32 to provide a 'cleaner' (read different) API.
.NET is Win32's replacement. It's the "new API". The API "akin to WINE" is Win32 itself.
There was never, ever, even the slightest possibility that Microsoft would do a dump-and-replace of Win32 in Windows 7, now is there in Windows 8, or any other forseeable release. It's simply not how they operate - way too disruptive to customers. Win32 will be phased out over time - you *might* see the first version of Windows without any win32 support ca. 2020, though it will probably be more like 2025-2030.
You mention OS 9 -> OS X move, but it's worth remembering that OS 9 wasn't itself a proper OS by modern standards, and was largely in the same boat as Win9x in terms of OS design.
Win9x ? Classic MacOS was much more like Windows 3.1 than it was 95.
But their biggest change was grafting their current broken OS on top of NT rather than reinvent properly.
Er, what ? NT's lineage from its first relase today is pretty clear and doesn't involve any "grafting".
It needs to be secure, flexible, low-latency, scalable, modular and customizable. It needs to be their Unix. Only, Steve Jobs beat them to the punch with OS X.
The only one of those that OS X is, is "secure" - and that's mostly because no-one is really targeting it yet.
Even as they implemented terminal services, they still worked around a broken multi-user model.
Broken how ?
And even when they saw their kernel was behind the curve on performance, they instead decided to bloat it even more.
For example ?
There were some claims that after Vista, they'd throw out their current API and start Windows anew, using an emulation layer (akin to Wine) to intercept old API calls.
Yes. It's called.NET.
7 really isn't the savior press make it out to be. Most of the Vista UI regressions remain.
What UI regressions ?
Vista's failure made 7 a necessity.
Eh ? Windows 7 proceeded pretty much as expected. What "necessity" are you referring to ?
Microsoft continually shoots itself in the foot by completely changing the user interface with each new release of software, resulting in massive productivity losses as everyone has to relearn how to do their job.
Rubbish. Office 2007 was the first "complete change" in the interface since, oh, at least Office 4.0 back in 1994. Windows hasn't had a "complete change" of its UI since Windows 95.
Not having to throw out your OS and apps, replace them all, and retrain everyone every few years on Microsoft's schedule is one of the real, tangible economic benefits of using open source.
The changes between different major releases of OSS software are *at least* as significant as those between Microsoft software.
Slashdot Mirror
Thanks for forcing me to wade through the suckiest site on the Web. Why the hell do they offer the CPUs with the exact same description for 2 prices?
As far as I can see they don't.
And when I actually come to the harddrives, there is no fucking 1 TB drive for the R610, because it only uses 2.5" drives. Beg your pardon?
Ah, sorry, I meant the R410, it uses 3.5" drives.
So with twice the HD capacity, the XServe costs 12% more. Way to go, Dell. A loooong way.
Er, no. Using 6x500GB drives in R610 gives you the same capacity, and also roughly twice the performance.
Xserve w/2x2.66Ghz, 24GB RAM (6x4), 3x1TB, RAID, redundant PSUs and Applecare: $10,349
Dell R410 w/2x2.66Ghz, 24GB RAM (6x4), 4x1TB, RAID, redundant PSUs: ~$7,500
Dell R610 w/2x2.66Ghz, 24GB RAM (6x4), 6x500GB, RAID, redundant PSUs: ~$8,500
Note also that it's highly likely the Dell prices would be discounted further after speaking with a sales rep, particularly if multiple units are being purchased. The chances of getting any discount from Apple - let alone a significant one - is basically zero.
Tell that to O.J.'s lawyers and anybody else who wants to claim, or actually believes, that the police might want to tamper with evidence to frame them. Or tell that to all the commentators on fark who are certain that the police would delete evidence to help protect one of their own. You should understand that these kind of enterprise storage systems that protect data from any kind of tampering (and "losing" data is a kind of tampering) even by system administrators are intended to be used in lieu of mountains of WORM media. When you've properly designed for their use they offer a worthwhile benefit.
I'm well aware of how the systems work. My first point was that "modification" and "deletion" are different operations, and allowing one by disallowing the other should be a minor technical issue. My second point was that with proper backups and auditing, it should be possible to implement deletion functionality without any need to compromise the trust of the system.
Native support for more than 640k ram.
Windows 3.1 could run in protected mode and most definitely make use of more than 640k RAM.
And don't say it was the application developers' fault; it's very hard to know the requirements of an operating system that will come out years in the future when you're writing your application.
It *is* the developers' fault, completely and utterly. They ran out of excuses not to write "UAC-friendly" software back around 1998 or so.
They are now the running Server 2008 kernel inside Win 7, and that is not a light upgrade, or even a service pack upgrade. That is a complete redesign of the core architecture.
It's nothing of the sort.
Vista and Server 2008 had the same major kernel revision.
Windows XP x64 and Windows 2003 are the same major kernel revision.
Windows 2000 (and earlier) Workstation and Server are the same major kernel revision.
Finally, they're *all* just different revisions of the same core architecture - Windows NT.
Compare Vista with an NT kernel OS and you'll see that XP or 2000 would run circles around Vista in many aspects, including stability.
Vista *is* an "NT kernel OS".
And yet they could have used the Internet for feedback well more than a decade ago.
*They* could have. Their customers, on the other hand, probably couldn't. In 1995, less than 10% of people would have had internet connections. Even ca. 1997-98, it was probably only around 30%.
They sorted that problem out about 10 years ago when Windows 2000 shipped.
Ten ? More like fifteen. I can count the number of BSODs I had with NT 4.0 on one hand, and most of those were hardware related.
Absolutely. Intel looked like they were going to start producing "decent" integrated graphics when, in 2004, they announced the GMA 900. It looked to Microsoft like the world's largest GPU maker would finally have something capable of desktop compositing, so they figured they could finally add this capability to Windows without a huge performance hit.
"Desktop compositing" was in the plan long before that. It first appeared in Longhorn in 2002.
Once the data is written, it can not be modified or deleted. Now, the reasoning behind this is so the police can not digitally manipulate the timestamps or data in any way. This is to protect the integrity of the data so it can withstand legal challenges.
Deleting and modifying are two distinctly separate things. It shouldn't be a problem for a system to allow deletions but not modifications.
Combined with proper backups and auditing, that should be sufficient to retain the legal value of any digital data.
I'll go with the XServe, that is cheaper still.
A comparably configured Xserve costs ca. 40% more.
Win7 is basically just a refurbished Vista under the hood.
Indeed. Just like any other software $VERSION+1 is "just a refurbished $VERSION under the hood".
Who ever suggested otherwise ?
That's a reason to make each respective piece of software _workable_. But there is no incentive to make it actually work as well on the foreign platform as it does on the native one...
There aren't many ways one could define Office for Mac as "foreign".
Surely you're not comparing 8 individual servers with a single server with 8 cores?
Indeed. Virtualisation makes it a perfectly reasonable comparison.
And it's a Dell? Wow. Just wow.
Even the lowest end Dell server is a vastly superior server to a Mac Mini.
The discussion was comparing blades to minis on the basis of cost, not density, but if you want to talk density...
Well, the cost argument works just as well for rackmount machines. An appropriately specced 1U rackmount is about $6.5k. 8 Mac Minis is about $8k.
I've just measured the rack in my garage - it's 30" deep by 19" across, and a mac mini fits handily into a 7" square by 4 holes in the rack with space to spare. 6U is 18 holes, so I can get 4 slide-out racks of minis into 6U, and 8 minis (including PSUs and wiring) per rack in a 2-across by 4 deep configuration. That's a total of 8x4x2 CPUs (because they're core-2 duos in the mini), or 64 CPUs.
Yes. I estimated you'd get about 8 Minis per 2U. You've measured it out at about 8 Minis per 1.5U - though I suspect you'd have trouble keeping the density you're talking about properly cooled, given the lack of high-volume forced cooling in the Mini. The ones at the upper rear of the rack, in particular, would suffer.
The blades pack 8 blades into 6U. The ones we started off discussing therefore put 8 CPUs into 6U. Whoosh. If we jump to the later offering, that's 8x4=32 cores in 6U, or 32 less than the minis. Only if we go to the 8-core blades (for a total of 4299+8*1610 for the E5405 CPUs, or $17179 before you add OS and disk storage) do you get the same density of 64 CPUs in 6U.
I didn't see the blades you were referring to earlier. A dell M10000e, or HP's c-Class, will put 16 blades into 10U, slightly higher density than the ones you are talking about.
The largest number of cores I'm aware of in standard 1U is dual quad-core (soon to be dual 6-core I suppose), so in 6U you get 6x8 = 48 CPUs. That doesn't work out too well compared to 6U of minis (64 CPUs).
Actually, it works out extremely well. Those rackmount servers have Nehalem based CPUs, which are roughly twice as fast as Core 2 CPUs at the same clock speed. Even being conservative, the 48 cores in 6U of rackmount servers will be quite a bit faster than the 64 cores in the Minis. The same is true of using Xserves, though the price comparison probably won't be quite as favourable.
I'm not advocating that you build your next computing cluster out of mac minis. I'm saying they scale (in price) reasonably well, that's all. I assume that there's a design difference between a blade server and a more consumer-orientated server (though I've been using a mini as an entertainment hub for the last 2 years without a single problem).
The thing is, a Mac Mini is not designed to be a 24/7 server. It also lacks numerous "standard" server features like power supply redundancy, remote management capabilities, and hot-swap hard disks. A Mini comes with a 1 year warranty standard, while even a low-end Dell server will come with a 3 year warranty. A Mini is difficult and time-consuming to service. A typical rackmount is trivial to break down and nearly any arbitrary component can be replaced in a matter of minutes.
If you need a low-end Mac server (or two), then a Mac Mini is certainly a reasonable option. If you just need "a server", though, and particularly if you need a lot of bang for your buck, then the Mini just doesn't match up. 1U rackmounts are a better deal up until about 8-10U, and after that the ~13 cores/RU (and Nehalem cores at that) blades provide is untouchable - and that's even before getting into the reliability and manageability aspects. There's no way those "Mac Mini farms" can be competitive in the face of Nehalem-powered blades and virtualisation, except for the handful of people who genuinely need a Mac.
Yes, and an 8-core XServe is 1U as well, but that's also out of scope of the discussion...
Why ? The context of the discussion is density. An example given was 8 Mac Minis. A 1U server will do as much as 8 Minis, more reliably, in less space, for less money.
Heck, your last post was comparing *blade servers*. How can blade servers be relevant and 1U rack servers not ?
I was using 8's because that's the unit of the enclosure. If you want to start using more blades, the 16-way enclosure is even more expensive, relatively, with prices for just the enclosure starting at $17k. Just multiply up to suit the numbers. I think you'll easily fit 4 mini's within a 6U rack-space though - you could probably fit them on a 2U tray...
However, a single 1RU server will fit in less space than 8 Minis and provide similar performance with all the benefits of the blades.
Overall, the mini comes out looking pretty good :)
Only if your requirements are low. If you need lots of capacity, 1U servers will comfortably provide ca. 1.5x the density, and blades will comfortably provide 2x the density, and either will provide a _vastly_ superior physical package.
You can cramp LOADS of those things in a small space and have massive storage and crunching power without needing an entire room.
Standard 1U rackmount servers have density at least as good (and generally better). Blades have (depending on vendor) up to 60% better density than that.
Not to mention real server hardware having vastly better cooling dynamics, substantially longer warranty, remote access and management tools and actually being designed to run 24/7.
8 blades @ 379 + enclosure @ 4299 => $7331
8 x Mac mini server => $7992
Dell R610 w/8 cores, 24GB RAM, 4x1TB drives, dual PSUs: $6,500.
20% cheaper, half the physical footprint (the Minis + a shelf to hold them will take 2 RU), more redundancy, vastly better cooling dynamics, actually designed to run 24/7 and 3 times the warranty. I know where I'd be spending my money.
I was quoting your remark about how "unusual" registry editing was in windows and you start calling me names. I guess that says it all about you and your 'arguments'.
No, you were being facetious. The discussion is (or was) about editing the Registry directly.
I will repeat it (you just try to read it slowly so that your synaptic pathways do not get damaged in the meantime): if you use windows, you edit the registry all the time (even if you do not know that you do so).
Great, but that's not what the discussion was about. Ergo, trying to equate it to what the discussion *was* about, is disingenuous.
Windows 7 was brought forward by 2 years at least, MS usually expects their OSs to last 5-7 years between updates.
Where do you get this number from ? The average time between major Windows NT releases is ~2.5 years, as you can see here. Windows 7 was pretty much bang on schedule. Even Vista wasn't really "late", when you consider it was mostly late in a marketing, rather than development, sense (much of the early work into "Vista" - then Longhorn - appeared in the market as Windows 2003, XP SP2 and 2003 R2).
NT was a cleanr OS, but they did graft the Win95 UI on top of it, and move the graphics drivers into the kernel to support it with enough speed.
They reimplemented the Windows 95 UI on top of it, yes. Why is this a bad thing ? Why is it any more "grafted" than, say, OS X's GUI ?
Also, most OSes run video drivers of some description in kernel space for performance.
I can't say that they broke more of it in Vista, but I wouldn't be surprised given the amount of changes they made to everything for Vista.
You haven't even managed to describe anywhere that they started "breaking" it, let alone did "more".
.NET is Win32's replacement. It's the "new API". The API "akin to WINE" is Win32 itself.
There was never, ever, even the slightest possibility that Microsoft would do a dump-and-replace of Win32 in Windows 7, now is there in Windows 8, or any other forseeable release. It's simply not how they operate - way too disruptive to customers. Win32 will be phased out over time - you *might* see the first version of Windows without any win32 support ca. 2020, though it will probably be more like 2025-2030.
You mention OS 9 -> OS X move, but it's worth remembering that OS 9 wasn't itself a proper OS by modern standards, and was largely in the same boat as Win9x in terms of OS design.
Win9x ? Classic MacOS was much more like Windows 3.1 than it was 95.
But their biggest change was grafting their current broken OS on top of NT rather than reinvent properly.
Er, what ? NT's lineage from its first relase today is pretty clear and doesn't involve any "grafting".
It needs to be secure, flexible, low-latency, scalable, modular and customizable. It needs to be their Unix. Only, Steve Jobs beat them to the punch with OS X.
The only one of those that OS X is, is "secure" - and that's mostly because no-one is really targeting it yet.
Even as they implemented terminal services, they still worked around a broken multi-user model.
Broken how ?
And even when they saw their kernel was behind the curve on performance, they instead decided to bloat it even more.
For example ?
There were some claims that after Vista, they'd throw out their current API and start Windows anew, using an emulation layer (akin to Wine) to intercept old API calls.
Yes. It's called .NET.
7 really isn't the savior press make it out to be. Most of the Vista UI regressions remain.
What UI regressions ?
Vista's failure made 7 a necessity.
Eh ? Windows 7 proceeded pretty much as expected. What "necessity" are you referring to ?
Less than what? If I really wanted to, I could fit a minimal-but-working Linux system on a floppy disk. Can you do that with Win 7?
I could do it with DOS, which is about as valid a comparison as the one you're making.
Microsoft continually shoots itself in the foot by completely changing the user interface with each new release of software, resulting in massive productivity losses as everyone has to relearn how to do their job.
Rubbish. Office 2007 was the first "complete change" in the interface since, oh, at least Office 4.0 back in 1994. Windows hasn't had a "complete change" of its UI since Windows 95.
Not having to throw out your OS and apps, replace them all, and retrain everyone every few years on Microsoft's schedule is one of the real, tangible economic benefits of using open source.
The changes between different major releases of OSS software are *at least* as significant as those between Microsoft software.