Windows 8 on USB to be Enterprise-only feature

[Dominus Atheos]

As some suspected yesterday when the feature comparison was released, neither Windows 8 (home edition) or Windows 8 Pro will be able to boot off a USB drive; only Windows 8 Enterprise, which you can only get through Software Assurance, will be to do so.

So for all of us who were considering buying Windows 8 so you could take your operating system with you wherever you go, sorry, 5 license minimum.

[Alyeska]

That feature is awesome for companies that need to stage computers quickly and easily. And it sucks that it is being restricted soley to Enterprise level and not being made a choice for consumers.

[Dave]

Alyeska, are you talking about first-time OS installs? Since most companies have fast internal networks, what was wrong with pushing the OS over network? Request OEM default to network installs, hook them up, turn them on and let the boot server do its thing.

[Alyeska]

Alyeska, are you talking about first-time OS installs? Since most companies have fast internal networks, what was wrong with pushing the OS over network? Request OEM default to network installs, hook them up, turn them on and let the boot server do its thing.

That option does not always exist. Especially with a network spread across 6 states and only one data center. No local servers at any branch except the DC and Disaster Recovery locations.

On conversations we periodically ran file share servers temporarily while on site. And even those got swamped.

Anyway, re-reading DAs post. I was thinking stage off USB. He is talking about boot of USB. Which is a concept I still have trouble wrapping my head around.

[Dominus Atheos]

Why? It's not like there should be any functional difference between booting off of flash memory instead of physical platters. Just look at SSDs.

[atg]

So for all of us who were considering buying Windows 8 so you could take your operating system with you wherever you go, sorry, 5 license minimum.

Wouldn't hardware drivers cause problems for most users? I remember WinXP just would not boot if you swapped the boot drive to another system (IIRC it hit problems if the chipset was different) without doing prep work before hand. Vista would at least boot but if the hardware was different you still had to go through installing correct drivers (i.e. gpu, etc). I havn't tried it with Win7 or the Win8 previews so I don't know if they handle it better.

Anyway isn't taking your OS with you wherever you go called buying a tablet/laptop

[Xon]

Why? It's not like there should be any functional difference between booting off of flash memory instead of physical platters. Just look at SSDs.

There is a significant technical difference.

Why does Windows not recognize my USB device as the same device if I plug it into a different port?

You may have noticed that if you take a USB device and plug it into your computer, Windows recognizes it and configures it. Then if you unplug it and replug it into a different USB port, Windows gets a bout of amnesia and thinks that it's a completely different device instead of using the settings that applied when you plugged it in last time. Why is that?

The USB device people explained that this happens when the device lacks a USB serial number.

Serial numbers are optional on USB devices. If the device has one, then Windows recognizes the device no matter which USB port you plug it into. But if it doesn't have a serial number, then Windows treats each appearance on a different USB port as if it were a new device.

(I remember that one major manufacturer of USB devices didn't quite understand how serial numbers worked. They gave all of their devices serial numbers, that's great, but they all got the same serial number. Exciting things happened if you plugged two of their devices into a computer at the same time.)

But why does Windows treat it as a different device if it lacks a serial number and shows up on a different port? Why can't it just say, "Oh, there you are, over there on another port."

Because that creates random behavior once you plug in two such devices. Depending on the order in which the devices get enumerated by Plug and Play, the two sets of settings would get assigned seemingly randomly at each boot. Today the settings match up one way, but tomorrow when the devices are enumerated in the other order, the settings are swapped. (You get similarly baffling behavior if you plug in the devices in different order.)

In other words: Things suck because (1) things were already in bad shape—this would not have been a problem if the device had a proper serial number—and (2) once you're in this bad state, the alternative sucks more. The USB stack is just trying to make the best of a bad situation without making it any worse.

[Dominus Atheos]

Why? It's not like there should be any functional difference between booting off of flash memory instead of physical platters. Just look at SSDs.

There is a significant technical difference.

Why does Windows not recognize my USB device as the same device if I plug it into a different port?

You may have noticed that if you take a USB device and plug it into your computer, Windows recognizes it and configures it. Then if you unplug it and replug it into a different USB port, Windows gets a bout of amnesia and thinks that it's a completely different device instead of using the settings that applied when you plugged it in last time. Why is that?

The USB device people explained that this happens when the device lacks a USB serial number.

Serial numbers are optional on USB devices. If the device has one, then Windows recognizes the device no matter which USB port you plug it into. But if it doesn't have a serial number, then Windows treats each appearance on a different USB port as if it were a new device.

(I remember that one major manufacturer of USB devices didn't quite understand how serial numbers worked. They gave all of their devices serial numbers, that's great, but they all got the same serial number. Exciting things happened if you plugged two of their devices into a computer at the same time.)

But why does Windows treat it as a different device if it lacks a serial number and shows up on a different port? Why can't it just say, "Oh, there you are, over there on another port."

Because that creates random behavior once you plug in two such devices. Depending on the order in which the devices get enumerated by Plug and Play, the two sets of settings would get assigned seemingly randomly at each boot. Today the settings match up one way, but tomorrow when the devices are enumerated in the other order, the settings are swapped. (You get similarly baffling behavior if you plug in the devices in different order.)

In other words: Things suck because (1) things were already in bad shape—this would not have been a problem if the device had a proper serial number—and (2) once you're in this bad state, the alternative sucks more. The USB stack is just trying to make the best of a bad situation without making it any worse.

Your post is a complete non sequiter. That describes the behavior of the Windows USB stack on a machine that has already booted. For reference, Linux has been able to boot from a flash drive for as long as motherboards have had the option to set a USB device as a boot device.

Seriously were you drunk when you made this post? Because the only thing your post and mine have in common is the term "USB".

[Skgoa]

I guess they don't allow taking your OS with you because they fear people would abuse that. Business customers buy volume licences anyways, but individuals are supposed to buy a licence for each computer.

I remember WinXP just would not boot if you swapped the boot drive to another system (IIRC it hit problems if the chipset was different) without doing prep work before hand.

Huh? I never had trouble booting from old installs.

[Xon]

Your post is a complete non sequiter. That describes the behavior of the Windows USB stack on a machine that has already booted. For reference, Linux has been able to boot from a flash drive for as long as motherboards have had the option to set a USB device as a boot device.

That is the behaviour of the underlying USB hardware stack, and that the serial number as defined in the USB spec isn't even compulsory nor guaranteed to be unique.

That Windows has the design choice of requiring tracking by serial number or track by connection without the ability to convert between the two is an obvious gap which they probably didn't bother to fix. Largely because untill relatively recently USB-boot has been a buggy clusterfuck in terms of working, both on the motherboard side and the usb-device side.

Aiming at purely Enterprise is likely driving in part that Enterprise customers are likely to actually spring for recent enough equipment that the feature will actually work.

Seriously were you drunk when you made this post? Because the only thing your post and mine have in common is the term "USB".

I gave you the technical response for why USB boot isn't something Windows does. There are also a lot of serial numbers which Windows uses to identify devices for persistant settings which would require reworking to let 'Windows to Go' work.

I guess they don't allow taking your OS with you because they fear people would abuse that. Business customers buy volume licences anyways, but individuals are supposed to buy a licence for each computer.

Depending on the hardware, USB-boot can still be a little dodgy. Enterprise equipment (including the USB drive!) is generally much better at being 'sane'. There is also the fact that Windows personalizes an install greatly based on installed hardware and the activation binds to a pile of serial numbers.

For consumers, this would be like giving them a copy of Windows which doesn't require activation. This isn't something I see Microsoft doing.

Huh? I never had trouble booting from old installs.

It radically depends on the hardware and drivers you have installed. Moving between identical machine should be very doable, but moving between an intel-based chipset to an amd-chipset would almost guaranteed to fail.

[someone_else]

last time I checked, the only thing preventing Win 7 from booting from a USB drive was that during boot sequence the OS starts fucking with how it recognizes the USB (resets USB driver or somesuch) and then fails to find the drive it was booting from (duh! you just reset it dumbfuck!).

To have it boot and run from a USB (not just making an installation USB stick) you have to hack it or use automated hacks others made.

I hope the feature is made unloackable in the future for others as well, or at least makes easier to hack it.

I mean, it's not like Windows 8 Enterprise isn't going to show up as fully working illegal download within weeks of release, so it makes little sense as an antipiracy move. Not that it matters anyway.

Huh? I never had trouble booting from old installs.

He means placing the booting HDD in a different computer. For example you install WinXP, after a few years of happy use your motherboard dies for some reason, you buy another (a different one since they change them pretty damn fast) and connect the HDD, WinXP refuses to boot.
It happened so much times to me...

[Skgoa]

Huh? I never had trouble booting from old installs.

He means placing the booting HDD in a different computer. For example you install WinXP, after a few years of happy use your motherboard dies for some reason, you buy another (a different one since they change them pretty damn fast) and connect the HDD, WinXP refuses to boot.
It happened so much times to me...

Yes... we are all talking about the same thing. When my mainboard died, I continued using my installed XP. My NAS is running the same six years old install. I sometimes boot from old partitions to copy old data like pgp keys on yet another new computer. (the oldest CPU/MoBo combo now rests in a box.) I never had a problem. I only used Intel processors, though. That might be the problem. (Although it shouldn't be.)

[Ariphaos]

*idiocy snipped*

Next you'll be telling us we can't possibly boot from CD or floppy because there's no way for such a thing to be uniquely identified.

[Xon]

Next you'll be telling us we can't possibly boot from CD or floppy because there's no way for such a thing to be uniquely identified.

This isn't remotely the same, and if you knew a damn thing about computer's you'ld know it. CD/floppies are tracked by controller and port, and explicitly do not support hotswap of devices once the the controller is powered up. Yes, the media can change but that isn't remotely the same as the entire damn device being removed and readded.

On the hardware layer, USB has an arbitary re-usable address assigned per controller. It's up to the OS to build it's own device-tracking on top. Windows uses the serial number (queried from the USB device) but is forced to fail back to controller+arbitary address is a duplicate serial number is there. This fallback process isn't particularly reliable in ensuring that Windows knows about that device after it is re-inserted somewhere else in the USB stack (or power cycles, or has the device restarted, or goes to sleep).

It's perfectly fine to bitch about the choice to restrict this to an enterprise-only feature, but it is utterly false to claim there is no difference between the USB stack and ATA/ACHI stack.

[Xon]

Also, linux's solution to these buggy hardware (usb devices with non-unique serials)? Don't use them.

And in case you missed it, one of the large reasons Windows doesn't offer this feature is licencing. Windows Activation which requires binding to hardware serials, which makes USB-based boot rather worthless.

[Ariphaos]

This isn't remotely the same, and if you knew a damn thing about computer's you'ld know it. CD/floppies are tracked by controller and port, and explicitly do not support hotswap of devices once the the controller is powered up. Yes, the media can change but that isn't remotely the same as the entire damn device being removed and readded.

On the hardware layer, USB has an arbitary re-usable address assigned per controller. It's up to the OS to build it's own device-tracking on top. Windows uses the serial number (queried from the USB device) but is forced to fail back to controller+arbitary address is a duplicate serial number is there. This fallback process isn't particularly reliable in ensuring that Windows knows about that device after it is re-inserted somewhere else in the USB stack (or power cycles, or has the device restarted, or goes to sleep).

Just like Windows won't know it's the original if you pull out the boot cd and stick it in a different drive.

It's perfectly fine to bitch about the choice to restrict this to an enterprise-only feature, but it is utterly false to claim there is no difference between the USB stack and ATA/ACHI stack.

Which is fine because I didn't claim it.

When you boot from USB, it emulates a hard drive.

Note the part about where you need to find the device's serial number? Oh wait. No mention at all, because it isn't relevant. The BIOS exposes the USB stick/drive/whatever as a hard drive, and an operating system can handle that appropriately.

These days, we use GPT, of course.

Also, linux's solution to these buggy hardware (usb devices with non-unique serials)? Don't use them.

Linux's solution is given in an Ubuntu bugtracker? Where people are specifically asked to file a bug against the Linux kernel?

]And in case you missed it, one of the large reasons Windows doesn't offer this feature is licencing. Windows Activation which requires binding to hardware serials, which makes USB-based boot rather worthless.

You made a ridiculously uninformed claim about the USB boot process. Whatever Microsoft's reasons, that claim is still completely and utterly bunk.

[Uraniun235]

Why would I want to bring my own OS to another computer?

[Ariphaos]

Typically, either troubleshooting, or a guaranteed clean start. I had one client who was so virus prone that that was the only reliable solution for them.

[Mr Bean]

Why would I want to bring my own OS to another computer?

Viruses, viruses and troubleshooting. Being able to take a dead computer alive again with a a usb stick would be highly useful in a corporate setting or home troubleshooting situation. Computer no boot? Plug in a usb-drive fire it up and look at your boot files. Even on hot-swapping drive bays taking a possibly dead HD out, re-connecting it and starting up takes at least ten minutes, knock that down to thirty second it a hell of a time saver. Being able to carry around a known good with you at all times is terribly useful. Sure you could do that with an external SATA but those don't always play right as boot on a drive externals. One of the promises of Win 8 is an OS smart enough to know the difference and not care.

[Uraniun235]

Typically, either troubleshooting, or a guaranteed clean start. I had one client who was so virus prone that that was the only reliable solution for them.

Troubleshooting can be accomplished with any number of bootable CDs, including livecd Linux distros, and is a hell of a lot cheaper than blowing a couple hundo on a Windows license.

Doesn't Windows write a bunch of data on boot? How would that be compatible with a read-only USB drive?

Viruses, viruses and troubleshooting. Being able to take a dead computer alive again with a a usb stick would be highly useful in a corporate setting or home troubleshooting situation. Computer no boot? Plug in a usb-drive fire it up and look at your boot files. Even on hot-swapping drive bays taking a possibly dead HD out, re-connecting it and starting up takes at least ten minutes, knock that down to thirty second it a hell of a time saver. Being able to carry around a known good with you at all times is terribly useful. Sure you could do that with an external SATA but those don't always play right as boot on a drive externals. One of the promises of Win 8 is an OS smart enough to know the difference and not care.

You really think a corporate environment is going to be cool with letting any schmuck user waltz in and boot their USB drive from god-knows-where?

Also, a corporate environment stands a good chance of running (or at least having access to) Enterprise, making this whole run-around even more pointless.

[Skgoa]

I want to bring my own OS to a computer so I know that everything is configured for me. My browsing history, my settings in whatever IDE I want to use... I am regularly using computers in my university in this way, though those boot over lan. It would be great to just have everything with me, anywhere anytime.

[Ariphaos]

Troubleshooting can be accomplished with any number of bootable CDs, including livecd Linux distros, and is a hell of a lot cheaper than blowing a couple hundo on a Windows license.

Doesn't Windows write a bunch of data on boot? How would that be compatible with a read-only USB drive?

There's no physical law that states Windows needs to write anything, much less write specifically to the drive it boots off of.

Regardless, I suppose I was answering a different question - 'why would someone'. A ~$200/seat license is not a high barrier of entry. People pay thousands for similar capabilities currently.

[Uraniun235]

It doesn't really matter anyway. Absent evidence that they are implementing read-only booting for the Enterprise version, it's a moot argument since this thread is whining about mean ol' Microsoft making USB booting an Enterprise-exclusive feature, not about mean ol' Microsoft refusing to rewrite their OS to cater to a tiny minority of users.

I want to bring my own OS to a computer so I know that everything is configured for me. My browsing history, my settings in whatever IDE I want to use... I am regularly using computers in my university in this way, though those boot over lan. It would be great to just have everything with me, anywhere anytime.

That makes a bit more sense. I'm not sure what proportion of organizations actually utilizes wholly network-booted workstations (or whole labs thereof). Any location that doesn't use boot-over-LAN might be more hesitant to allow you to boot your own OS, since doing so gives you full control over the local hardware and could be used to render the local hard drive unbootable. This doesn't even begin to address environments that exert control over what applications they want to permit to have access to the network, or that want to implement other network policies on all clients. The irony is that a proliferation of people bringing their own OS could lead to a crackdown that makes it harder for people to do so.

That said, although I know you don't have control over the development of applications (any more than people have control over the development of Windows... hmm...) it seems like it would be substantially more efficient to bring applications which don't need local installs in order to retain preferences and history.

[someone_else]

Why would I want to bring my own OS to another computer?

Apart from what Skgoa said, which is entriely doable with any of the best linux distros out there, I'll tell a wild speculation trying to catch the idea behind Microsoft's move.

More and more companies here are buying fucktons of cheap not-so-crappy nettops like this (scroll down a little for intelleggible specs) for at most 300 euros to replace aging tower desktops without blowing 500-600 euros in hardware that is not much better anyway.

To make these things even cheaper you can buy them without an internal HDD and buy instead a fuckton of 8GB keydrives for nearly nothing.

This way you can install the Win8 OS on the USB stick and use it in even cheaper workstations.

Also, USB 3.0 is coming. Which means this arrangement may become more common.

Of course you'll have to take precautions to not allow any fool to plug in his own keydrive and have fun with his own linux OS, but that's not horribly difficult if you have one internal USB port (where the USB main drive will go) and the ones on the outside (where the user can put his own USB things) hidden behind a Hub (so the bios won't look at their contents during boot).

This also makes easy to restore functionality to a dead workstation due to malware or other crap, just swap the USB drive. (if you have enough spare USB drives ready for this)

Troubleshooting can be accomplished with any number of bootable CDs, including livecd Linux distros, and is a hell of a lot cheaper than blowing a couple hundo on a Windows license.

LiveCDs tend to be MUCH slower to load (especially the windows-based ones), unless you are using a very lightweight linux distro.

Albeit most serious linux distros have an option to be installed and live their whole life on a USB stick/HDD (like say Puppy). Which is the preferred way to trubleshoot stuff of my friend.

Doesn't Windows write a bunch of data on boot? How would that be compatible with a read-only USB drive?

Why on Earth must the USB drive be read-only?

[Uraniun235]

More and more companies here are buying fucktons of cheap not-so-crappy nettops like this (scroll down a little for intelleggible specs) for at most 300 euros to replace aging tower desktops without blowing 500-600 euros in hardware that is not much better anyway.

To make these things even cheaper you can buy them without an internal HDD and buy instead a fuckton of 8GB keydrives for nearly nothing.

So... exactly the kind of company that would have access to the Enterprise version of Windows?

Also there's no way an 8GB drive would be adequate for Windows 8, and a bottom-cost keydrive is going to be slow as shit. You could get a larger and slightly faster one, but then you're really not saving much over a bottom-end hard drive for like ~$20.

Why on Earth must the USB drive be read-only?

Xeriar noted that one use was for a "guaranteed clean start" because he had a client who was "so virus prone" that it was "the only reliable solution". If the USB drive is writeable, then that USB-bootable OS is just as prone to viruses as any other OS.