I have not heard that Rackspace resells Amazon. If they do, I find it odd that their datacenters are in different locations, and that their services offered are so different (Rackspace is way behind Amazon).
In any event, you seem to have misunderstood my point. My point was that Azure was going to drive Hyper-V development through the roof, not that azure was a replacement for Hyper-V or whatever you seemed to think I was talking about.
We are talking about the VMWare technology vs the Microsoft technology when it comes to virtualization here, I fail to see why discussing the factors that will drive technology changes is forking an argument.
There's a reason they call it the "Magic" quadrant, because they do a lot of hand waving and ordering you to ignore the man behind the curtain.
It's all just marketing bs. Yes, we all know that VMWare is ahead of all the competition in many ways (but not all of them). I personally use vSphere myself because the memory over-commit and page de-duplication is critical. However, once Microsoft adds those features to Hyper-V, it's going to be a significantly more level playing field.
Microsoft is basing Azure on Hyper-V, which is generating a lot of impetus to improve the technology. They've already largely caught up the virtual networking aspect (VMWare used to be ahead there) and in the management area (VMM is getting quite advanced).
There's a reason Amazon, Rackspace, and most other commerical "cloud" services are based on Xen rather than VMWare.
Why are you attributing SAN and NAS technology to VMWare? They didn't invent it, they *USE* it. The same technology the Windows uses.
VMFS has a few nice features, but also a ton of limitations. Hell, it didn't support directories until VMFS3, and even now still limits file sizes to either 256MB or 2TB.
It's a special purpose filesystem designed for a single use, store Virtual Machine disk files. Because of that, it doesn't have to have the same characteristics of a general purpose filesystem.
It's only real advantage is the ability to share the filesystem with multiple servers (clustere file system). But it lacks in a lot of other areas. If all you can trot out is VMFS, then I think you've proven my point for me.
Whether or not Microsoft used backslashes or slashes, there would still be two of them on the keyboard, causing confusion with end users. Microsoft didn't invent the backslash... and it's used pretty heavily for other things (like for escaping in programming).
As long as there is two slashes on the keyboard, presenters will have to diferentiate between them.
Drive letters have not been required by the OS for nearly 20 years. Apps still use them, Admins still use them.
Admins like drive letters, because end users use them as cue that they're storing their data on the network. Using some kind of path like/user/network/data/joe is less obvious to end users. They still say stuff like "I put that file out on the S: drive".
Actually, that kind of blocking operation is by design. THe purpose is to prevent well-known security issues, particularly in audit-trails.
If you can rename the log files while they're open, or even better, delete them.. while open... that means an attacker can do something like this:
* Make a copy of the audit log file * Delete the current audit log file * Do nasty things that get written to the delted log file (because the filehandle is still open, it keeps writing to the deleted file) * Rename the copied logfile back to the original name * Nobody knows you were there, because the log file seems to be fine.
You can't do that in windows, because logfiles are kept open by the OS, and never closed while the OS is running, so the log is tamper proof. Yes, you can delete the log, but then someone knows someone deleted the log.
Greyhole comes a lot closer, but it's still pretty experimental, and it's relatively recent... which goes against the claim that DE is doing what Linux or Unix has done for decades.
LVM is only part of what DE did. LVM doesn't do data de-duplication, nor does it provide redundancy like DE did (yes, you can stripe volumens and mirror them with LVN, but that's just doing software RAID). DE used drive balancing, which copied data between drives to keep them balanced and redundant. This way, if a drive fails, and you have enough storage, the system will just rebalance (assuming you have more than one drive) without any degradation of performance (like you would see in RAID-5).
Let's say you have 5 1TB drives, and you have 50% of the capacity used. That will mean you have 50% free on all drives (or about 500GB). If you lose one drive, then the system will rebalance between the 4 drives (so you will now have 250GB free on 4 drives), if you lose another drive, it will rebalance over 3 drives.. etc.. until you no longer have enough free storage to compensate. Add more drives, and it will rebalance across those drives.
This system is not as robust as RAID, in that if you lose a drive before a rebalance can occur, you might lose data... but it's far more flexible and easy to use. It was designed for home users who may not understand RAID, but are technical enough to be able to pop in a new hard drive if they need more stogage.
This is such a typical comment from people that have no idea what Drive Extender is (or was). It's not RAID. RAID has a ton of issues to deal with. You can't just add more drives to raid and have the volumes magically get bigger, not without some serious rebuilding (often taking 10+ hours on larger volume sizes). Even then you have to extend your partitions manually.
DE just works. You plug in a new drive, of any size, and you get additional redundant storage, instantly, without rebuilds or partition extending, or any of the other headaches the come with RAID. Also, if your motherboard fails, you don't have to reconfigure RAID on a new computer.. you can just access the files from any computer, since the files were stored on normal NTFS filesystems, without the need to create drive sets.
When WHS first came out, i tried to explian how useful Drive Extender was, and just got blank looks from the unix guys.. saying "it's just raid". It was like trying to explain DVR's to people that had only used VCR's. They just didn't get it until they used it.
I'm sorry, since when does VMWare have any major file storage technology? And what does Hyper-V have to do with data storage?
Microsoft happens to have a pretty good SAN system called Windows Storage Server. I don't know of any VMWare product that is storage related. True, VMWare products like vSphere *USE* storage, typically via SAN LU's but their products are not themselves Storage related technology.
VMWare is still ahead of Hyper-V in a lot of ways, but that gap has narrowed a great deal in recent years. If Hyper-V gets memory over-commit and page de-duplication then it will be largely on par with vSphere.
I'm sure that these test results are based on the beta version, not the RC, and the beta version is about 10 months old, compared to all the recent releases you mention in comparison.
It's also worth mentioning that HTML5 and CSS3 aren't ratified standards yet.
Microsoft executives wish to do whatever makes them the most money, and they don't make money by giving customers what they don't want.
You may bitch about standards conformance as if it's something everyone wants, and perhaps they do... but they also don't want their old apps to break. That's a very sharp razor to walk, and I think Microsoft has been successful in gradually moving their legacy customers forward into modern browser standards.
You can say "they did it to themselves", but that's just ignoring history. Think about it like this... When IE6 was released, Mozilla didn't even exist, at least not as a 1.0 product. Mozilla 1.0 was released in November 2004.. that's 3 years after IE6.
In effect, there was no "standards compliant" browsers when IE6 was released. Opera existed, and was pretty conformant, but few people even knew it existed. Standards conformance is an ongoing effort (partially because standards progress and change as well), but more imporantly, the interpretations of the standards change over time. When IE6 was released, it was considered the most standards conformant browser. That quickly changed, as other browsers became more conformant, and IE6 sat idle, but IE6 was not any specific attempt by Microsoft to inflict proprietary standards on people.
I'm sorry, since when did all other browsers implement the things you're talking about? I'm not sure any of them have implemented all of it.
What's more, all of those things are draft standards, and subject to change. In my opinion, it's reckless to implement things and call them standards when they're not actually standards yet. People can use them, and then the standards change...
I'm not sure where you are getting your information, but I suspect you are grossly misinterpreting something.
IE9 and the next office, and VS 2010 will all be based on WPF's rending engine, which will use 3D rendering capabilities if available. IE's rendering engine changes so rapidly (not counting IE6) that this would be ridiculous. We've had 3 versions of IE in 4 years. Basing an app on that, which typically has a 2-3 year rev cycle would be stupid.
The client download is to run desktop applications, not web applications. ASP.NET does not require any special client, or any framework installed on the client.
That's simply not true for the vast majority of web apps. The only web apps that require.net on the client are WPF web apps, and i've never actually seen one in the wild... Just not something anyone will build.
asp.net apps just genearte HTML, javascript, css, etc.. While various versions have had varying levels of standard compliance, the new.NET 4 stuff has been largely redesigned to be completely standards compliant out of the box.
Silverlight requires a client side runtime that's based on.NET, but IE is not required for it...
Technically, it was the iPhone 3G, and 3GS. The iPhone 3G was actually the 2nd generation, and the 3GS was the 3rd. Thus, the iPhone 4 which is the 4th generation (but still 3G)
I'm aware of that. That was part of my point. Looking at what the sportsbooks post as the winner may, or may not be anywhere near what their actual opinion is.
When there is an overwhelming majority of the public that believes one side will will, the odds will reflect a number where half the public will take the other side, even if they think the "winning" side will win. No book wants to be lopsided with a handle of 10's or 100's of millions, unless they're very opinion in a way that is opposite of the public (maybe inside information).
Slashdot Mirror
I have not heard that Rackspace resells Amazon. If they do, I find it odd that their datacenters are in different locations, and that their services offered are so different (Rackspace is way behind Amazon).
In any event, you seem to have misunderstood my point. My point was that Azure was going to drive Hyper-V development through the roof, not that azure was a replacement for Hyper-V or whatever you seemed to think I was talking about.
We are talking about the VMWare technology vs the Microsoft technology when it comes to virtualization here, I fail to see why discussing the factors that will drive technology changes is forking an argument.
There's a reason they call it the "Magic" quadrant, because they do a lot of hand waving and ordering you to ignore the man behind the curtain.
It's all just marketing bs. Yes, we all know that VMWare is ahead of all the competition in many ways (but not all of them). I personally use vSphere myself because the memory over-commit and page de-duplication is critical. However, once Microsoft adds those features to Hyper-V, it's going to be a significantly more level playing field.
Microsoft is basing Azure on Hyper-V, which is generating a lot of impetus to improve the technology. They've already largely caught up the virtual networking aspect (VMWare used to be ahead there) and in the management area (VMM is getting quite advanced).
There's a reason Amazon, Rackspace, and most other commerical "cloud" services are based on Xen rather than VMWare.
Why are you attributing SAN and NAS technology to VMWare? They didn't invent it, they *USE* it. The same technology the Windows uses.
VMFS has a few nice features, but also a ton of limitations. Hell, it didn't support directories until VMFS3, and even now still limits file sizes to either 256MB or 2TB.
It's a special purpose filesystem designed for a single use, store Virtual Machine disk files. Because of that, it doesn't have to have the same characteristics of a general purpose filesystem.
It's only real advantage is the ability to share the filesystem with multiple servers (clustere file system). But it lacks in a lot of other areas. If all you can trot out is VMFS, then I think you've proven my point for me.
Hmm.. nobody noticed the article was published 2.5 months ago?
Umm.. Your link alleges that Microsoft locked Be out of the market, not that they stole any source code. What's your point exactly?
Whether or not Microsoft used backslashes or slashes, there would still be two of them on the keyboard, causing confusion with end users. Microsoft didn't invent the backslash... and it's used pretty heavily for other things (like for escaping in programming).
As long as there is two slashes on the keyboard, presenters will have to diferentiate between them.
Drive letters have not been required by the OS for nearly 20 years. Apps still use them, Admins still use them.
Admins like drive letters, because end users use them as cue that they're storing their data on the network. Using some kind of path like /user/network/data/joe is less obvious to end users. They still say stuff like "I put that file out on the S: drive".
Actually, that kind of blocking operation is by design. THe purpose is to prevent well-known security issues, particularly in audit-trails.
If you can rename the log files while they're open, or even better, delete them.. while open... that means an attacker can do something like this:
* Make a copy of the audit log file
* Delete the current audit log file
* Do nasty things that get written to the delted log file (because the filehandle is still open, it keeps writing to the deleted file)
* Rename the copied logfile back to the original name
* Nobody knows you were there, because the log file seems to be fine.
You can't do that in windows, because logfiles are kept open by the OS, and never closed while the OS is running, so the log is tamper proof. Yes, you can delete the log, but then someone knows someone deleted the log.
mhddfs only provides spanning of volumes, it doesn't do the rest of it.
Greyhole comes a lot closer, but it's still pretty experimental, and it's relatively recent... which goes against the claim that DE is doing what Linux or Unix has done for decades.
LVM is only part of what DE did. LVM doesn't do data de-duplication, nor does it provide redundancy like DE did (yes, you can stripe volumens and mirror them with LVN, but that's just doing software RAID). DE used drive balancing, which copied data between drives to keep them balanced and redundant. This way, if a drive fails, and you have enough storage, the system will just rebalance (assuming you have more than one drive) without any degradation of performance (like you would see in RAID-5).
Let's say you have 5 1TB drives, and you have 50% of the capacity used. That will mean you have 50% free on all drives (or about 500GB). If you lose one drive, then the system will rebalance between the 4 drives (so you will now have 250GB free on 4 drives), if you lose another drive, it will rebalance over 3 drives.. etc.. until you no longer have enough free storage to compensate. Add more drives, and it will rebalance across those drives.
This system is not as robust as RAID, in that if you lose a drive before a rebalance can occur, you might lose data... but it's far more flexible and easy to use. It was designed for home users who may not understand RAID, but are technical enough to be able to pop in a new hard drive if they need more stogage.
This is such a typical comment from people that have no idea what Drive Extender is (or was). It's not RAID. RAID has a ton of issues to deal with. You can't just add more drives to raid and have the volumes magically get bigger, not without some serious rebuilding (often taking 10+ hours on larger volume sizes). Even then you have to extend your partitions manually.
DE just works. You plug in a new drive, of any size, and you get additional redundant storage, instantly, without rebuilds or partition extending, or any of the other headaches the come with RAID. Also, if your motherboard fails, you don't have to reconfigure RAID on a new computer.. you can just access the files from any computer, since the files were stored on normal NTFS filesystems, without the need to create drive sets.
When WHS first came out, i tried to explian how useful Drive Extender was, and just got blank looks from the unix guys.. saying "it's just raid". It was like trying to explain DVR's to people that had only used VCR's. They just didn't get it until they used it.
I'm sorry, since when does VMWare have any major file storage technology? And what does Hyper-V have to do with data storage?
Microsoft happens to have a pretty good SAN system called Windows Storage Server. I don't know of any VMWare product that is storage related. True, VMWare products like vSphere *USE* storage, typically via SAN LU's but their products are not themselves Storage related technology.
VMWare is still ahead of Hyper-V in a lot of ways, but that gap has narrowed a great deal in recent years. If Hyper-V gets memory over-commit and page de-duplication then it will be largely on par with vSphere.
That's funny, since from your comment above, you say "an important part of the HTML5-spec offline cache and HTML5-local-storage and -session-storage".
So how exactly does it have nothing to do with specs and drafts when you actually called it a spec? And, it's a Draft Specification, not a final one.
I'm sure that these test results are based on the beta version, not the RC, and the beta version is about 10 months old, compared to all the recent releases you mention in comparison.
It's also worth mentioning that HTML5 and CSS3 aren't ratified standards yet.
Umm.. that's exactly what IE8 does, except in certain situations where it can only assume it's quirky html.
Mozilla plans to release 4 versions of FF this year, according to another slashdot article i read, FF4, FF5, FF6, and FF7.
So, will your dad have a newer OS in 3 months?
Microsoft executives wish to do whatever makes them the most money, and they don't make money by giving customers what they don't want.
You may bitch about standards conformance as if it's something everyone wants, and perhaps they do... but they also don't want their old apps to break. That's a very sharp razor to walk, and I think Microsoft has been successful in gradually moving their legacy customers forward into modern browser standards.
You can say "they did it to themselves", but that's just ignoring history. Think about it like this... When IE6 was released, Mozilla didn't even exist, at least not as a 1.0 product. Mozilla 1.0 was released in November 2004.. that's 3 years after IE6.
In effect, there was no "standards compliant" browsers when IE6 was released. Opera existed, and was pretty conformant, but few people even knew it existed. Standards conformance is an ongoing effort (partially because standards progress and change as well), but more imporantly, the interpretations of the standards change over time. When IE6 was released, it was considered the most standards conformant browser. That quickly changed, as other browsers became more conformant, and IE6 sat idle, but IE6 was not any specific attempt by Microsoft to inflict proprietary standards on people.
I'm sorry, since when did all other browsers implement the things you're talking about? I'm not sure any of them have implemented all of it.
What's more, all of those things are draft standards, and subject to change. In my opinion, it's reckless to implement things and call them standards when they're not actually standards yet. People can use them, and then the standards change...
I'm not sure where you are getting your information, but I suspect you are grossly misinterpreting something.
IE9 and the next office, and VS 2010 will all be based on WPF's rending engine, which will use 3D rendering capabilities if available. IE's rendering engine changes so rapidly (not counting IE6) that this would be ridiculous. We've had 3 versions of IE in 4 years. Basing an app on that, which typically has a 2-3 year rev cycle would be stupid.
The client download is to run desktop applications, not web applications. ASP.NET does not require any special client, or any framework installed on the client.
That's simply not true for the vast majority of web apps. The only web apps that require .net on the client are WPF web apps, and i've never actually seen one in the wild... Just not something anyone will build.
asp.net apps just genearte HTML, javascript, css, etc.. While various versions have had varying levels of standard compliance, the new .NET 4 stuff has been largely redesigned to be completely standards compliant out of the box.
Silverlight requires a client side runtime that's based on .NET, but IE is not required for it...
Technically, it was the iPhone 3G, and 3GS. The iPhone 3G was actually the 2nd generation, and the 3GS was the 3rd. Thus, the iPhone 4 which is the 4th generation (but still 3G)
It's easy. They don't want you running servers on your residential cable connections. They want to charge premiums for that.
I'm aware of that. That was part of my point. Looking at what the sportsbooks post as the winner may, or may not be anywhere near what their actual opinion is.
When there is an overwhelming majority of the public that believes one side will will, the odds will reflect a number where half the public will take the other side, even if they think the "winning" side will win. No book wants to be lopsided with a handle of 10's or 100's of millions, unless they're very opinion in a way that is opposite of the public (maybe inside information).