Slashdot Mirror

It's not cheap so might not be a viable option for a smaller shop, but VMware has been making some very interesting strides in this area.

Check out VMware View, also known as PCoIP (Yes, that is personal computer over internet protocol)

http://www.vmware.com/products/view/
http://www.vmware.com/resources/techresources/10083

Put really simply, each real workstation is loaded with a minimal system and the vmware view clients.
When a user goes to login to a computer on your network, after authentication their virtual workstation pops up (Be it windows or ubuntu) and lets them work.

All of the actual 'workstations' being used are virtual machines, thus are the same unified image you are looking for with one set of drivers.

While I have not tested it with a multi-monitor setup, they claim it is supported.

The one main thing you do lose is full accelerated 3D support, and direct support for old eccentric hardware. (Think ISA card support and non-standard PCI interfaces)
I can say USB support is simply amazing in how well it works.

Clients can even play full interactive flash media and video, and it runs well (As well as one would expect it to work in native OS anyway)

VMWare View is what you want.

For now, at least. VMWare doesn't support combining pages >= 2MB because the overhead (hit rate on finding duplicates versus the cost of searching for duplicates) and I suspect other hypervisors will do the same. Additionally, Intel and AMD are both moving to support 1GB page tables. What are the odds that you'll start up two VMs and their first 1GB of memory will remain identical for very long?

The only way I see page sharing working in the future is if the hypervisor inspects the nested pages down to the VM level, which will typically be the 4KB pages we know and love. Either that, or paravirtualization support needs to exist for loading common code and objects into a shared pool.

Even so, there's a lot of overhead from inspecting (hashing and then comparing) pages which will only grow as memory sizes grow. If we increase page sizes, the hit rate decreases and the overhead of copy-on-write increases. It's not a good situation.

Sources: Performance Best Practices for vSphere 4 which references Large Page Performance which states:

In ESX Server 3.5 and ESX Server 3i v3.5, large pages cannot be shared as copyonwrite pages. This means, the ESX Server page sharing technique might share less memory when large pages are used instead of small pages. In order to recover from nonsharable large pages, ESX Server uses a “sharebeforeswap” technique. When free machine memory is low and before swapping happens, the ESX Server kernel attempts to share identical small pages even if they are parts of large pages. As a result, the candidate large pages on the host machine are broken into small pages. In rare cases, you might experience performance issues with large pages. If this happens, you can disable large page support for the entire ESX Server host or for the individual virtual machine.

That is, page sharing involves breaking up large pages, negating their performance benefit and is only used as a last ditch when you've overcommited memory and you're nearly to the point of having to hit the disk. And VMWare overcommit is great until you hit the disk, then it's a nightmare.

For now, at least. VMWare doesn't support combining pages >= 2MB because the overhead (hit rate on finding duplicates versus the cost of searching for duplicates) and I suspect other hypervisors will do the same. Additionally, Intel and AMD are both moving to support 1GB page tables. What are the odds that you'll start up two VMs and their first 1GB of memory will remain identical for very long?

The only way I see page sharing working in the future is if the hypervisor inspects the nested pages down to the VM level, which will typically be the 4KB pages we know and love. Either that, or paravirtualization support needs to exist for loading common code and objects into a shared pool.

Even so, there's a lot of overhead from inspecting (hashing and then comparing) pages which will only grow as memory sizes grow. If we increase page sizes, the hit rate decreases and the overhead of copy-on-write increases. It's not a good situation.

Sources: Performance Best Practices for vSphere 4 which references Large Page Performance which states:

In ESX Server 3.5 and ESX Server 3i v3.5, large pages cannot be shared as copyonwrite pages. This means, the ESX Server page sharing technique might share less memory when large pages are used instead of small pages. In order to recover from nonsharable large pages, ESX Server uses a “sharebeforeswap” technique. When free machine memory is low and before swapping happens, the ESX Server kernel attempts to share identical small pages even if they are parts of large pages. As a result, the candidate large pages on the host machine are broken into small pages. In rare cases, you might experience performance issues with large pages. If this happens, you can disable large page support for the entire ESX Server host or for the individual virtual machine.

That is, page sharing involves breaking up large pages, negating their performance benefit and is only used as a last ditch when you've overcommited memory and you're nearly to the point of having to hit the disk. And VMWare overcommit is great until you hit the disk, then it's a nightmare.

68k emulation in PPC was decent. Classic mode worked for most applications and Rosetta was as seamless as it gets. I understand that Microsoft has a ton of backwards compatibility they need to maintain, but if a company the fraction of your size can do it, why can't you?

er... so these periodic compatibility updates I see for Windows aren't related to backwards compatibility? Including the Windows 7 one back in February that fixed Warcraft 3's video playback (that's the only fix I noticed for stuff I had)?

Yes, there are still problems, particularly for games (Ex: Syberia crashes on my Win7 x64 PC when you enter the factory courtyard, even in Windows 98 compatibility mode; I run the entire game in a VMWare Player Windows XP VM now... won't work in VirtualBox or Wine because WineD3D only supports D3D 8 and 9), but MS does have a lot of compatibility shims to make everything from Windows 95 to present run. Having said that, the 64-bit version of Windows will not run 16-bit Windows applications (for Windows 3.1 or earlier) at all.

Of course, since Snow Leopard doesn't even include Rosetta to run PPC programs from 5 years ago, expecting Windows from 2009 to run 1993 applications is a stretch.

Maybe I'm wrong about the Open Source bit, VMWare Player has an Open Source tab on their downloads page.

I would second the idea of looking into alfresco. I have not used it.
However, what it will do for you is that it will make sure that you can be using a common file system with revision control. So what would happen is that you would allow your users to network mount the alfresco filesystem across the firm. Users would read and save files to this filesystem. Anytime, it is saved, versions are created.

Alfresco Documents

Also, it does handle signatures with the plugin from http://www.viafirma.org/ (note, that is in spanish but works fine with google translate) http://viafirma.googlecode.com/svn/

Those saying stop working on this and hire people are thinking that you have a large firm. That is not really a great option.
What I would recommend is that you do setup single signon if you can.
The first start is to have an LDAP server.
ActiveDirectory does provide that. If you want to provide kerberos/active directory and ldap there are open source solutions.

• The way to technical solution is: http://freeipa.org/page/Samba_4_Installation
• Note: some of this can be done with: http://www.vmware.com/appliances/directory/393283
• The no to technical solution might be: http://www.clearfoundation.com/Software/overview.html

Well it's not about need. If you really need constant availability with absolutely no interruptions, you'll have engineered something to provide that. But most people would like to have constant availability, but the reality is you can't afford to provide that, so do what they can given their resources. That often means a high-availability failover cluster, with a short disruption to services whenever the failover occurs.

An inexpensive service to allow you to avoid failovers for some classes of scheduled maintenance just gives you another tool you can use to get closer to constant availability at a price that's affordable.

Combine this with something like VMware's fault tolerance and you could get a pretty robust solution. Now you just need to be able to patch the programs providing your service in-memory and you're gold!

Not really. With the multicore, gigabytes of ram type, systems becoming norm, think ThinApp + VMWare you can start having applications running in a completely disposable virtual machine and it would work just like a regular application, only it can't ACTUALLY access your system.

Not really. With the multicore, gigabytes of ram type, systems becoming norm, think ThinApp + VMWare you can start having applications running in a completely disposable virtual machine and it would work just like a regular application, only it can't ACTUALLY access your system.

Cisco / VMware has done some work in this space, abeit it is a Cisco / VMware solution.... The Nexus 1000V basically provides an overlay to the virtual networking stack from VMware and places it into an appliance with a Cisco CLI. It can then be hooked into the usual Cisco management suspects. The solution makes sense because it also gives back control of the network aspects back to netops, instead of the server ops/virtual ops... http://www.vmware.com/products/cisco-nexus-1000V/

VMware's solution for test and dev allows you to access the VMs over your browser: http://www.vmware.com/products/labmanager/

VMware even makes a product just for that use case: http://www.vmware.com/products/labmanager/

Oof!:

http://store.vmware.com/store/vmware/en_US/DisplayProductDetailsPage/productID.105855000

Not a big deal for lots of groups, but a show stopper for lots of others.

Semi-OT, but a handy way to use different versions of Office on the same PC, and portably on a USB key, is to modify their installation via VMWare ThinApp:

http://www.vmware.com/products/thinapp/

I found out about Thinstals/Thinapps/"portable" versions when I accidentally browsed a torrent site where they are popular for various reasons, but the concept works well and it's easier to copy/paste a folder than do a conventional install.

I think you're confusing VMWare's HA and everyone else's. VMWare has a solution that allows a server to be unplugged and the VM remains up. Red Hat, Hyper-V, Xen, etc, currently can't do that. It's very tricky work.

http://www.vmware.com/products/fault-tolerance/

I mean, that's even a step above HA. You set your cluster to replicate the memory and actions of a VM and you can pull the plug and everything keeps going.

As far as prices go, it sounds pretty good, but again, please don't quote "starting at." Quote one with comparable support. Microsoft's support is 24x7 phone support and after the 2nd or 3rd year of payments (depending on licensing program), the annual cost drops to a third or a half of the initial annual costs. So even Windows Server Datacenter drops down to around $1300 a year (still per socket, eugh) and Windows Server Enterprise drops to around $600 a year (8 socket max, 4 virtual instances included, so still not as good as RH Advanced Platform.)

We could continue to argue prices all day, but the point I was trying to make was that there's a ton of FUD about Windows pricing. And calling it Windows 2008 Not Crippled Edition is almost as bad as calling them M$. Some features cost more to implement, more to support, and more to maintain. This is true in the open source and proprietary world, and I just get annoyed when I read nonsense like he spouted.

Whoops, I lied. The latest version can create images now. I'm still not sure that it has 3D support, though.

It does have 3D support.

By the way, the release notes were very easy to find on Google - it was the second result when searching for VMware Player 3.

It is expensive... when compared to VMWare Player, which is free (as in beer) and also supports up to 4 virtual CPUs.

Of course, you can't create VM images with VMWare Player...

By what measure of success? Effectiveness, sure. But what is the market share of all the Linux distros put together? What is the ratio of Windows to Linux boxes globally or in the US?

How do you define success? Apparently for you, at least in this context, it's market share. That's a bit spurious considering the fact that Linux is Free. What is the market share of Linux anyway? Probably a lot higher than you realize. Oh, you thought only desktop computers ran operating systems? Ever heard of Tivo? Android? Routers? Embedded systems? Servers? I'll bet if you put every device that runs Linux vs. every device that runs Windows you might be surprised about the "ratio of Windows to Linux boxes globally". Besides, Ferrari has low market share. They're a success right?

Says you. You're omitting how many devices don't work on Linux due to a lack of drivers or simple inoperability with Linux. It's improving, but there's a long way to go.

Linux supports more peripherals than OSX; I don't see you bringing that up? You wouldn't happen to have an agenda would you? Besides, I've installed quite a few Linux boxes in my day. It's the very rare exception that I find a device that doesn't just work out of the box. Contrast this with literally every other operating system ever made. And everytime that has happened, I waited a few months for the next kernel update and it did work. A lot of hardware actually works better in Linux. For example, my Verizon USB aircard. In Windows, you have to wait over 30 seconds for it to do its thing and connect and it disconnects requiring pulling it out and reinserting it about once an hour. On Linux, it connects in about 5 seconds and works perfectly for as long as you want. Funny story, I was at my brother's house a couple of weeks back and his Windows 7 box bluescreened so many times, I lost count. Finally, I was like, dude, what does the error say when it crashes? Come to find out the problem was the USB network adapter he had was crashing his box. Plugged it into my netbook running Ubuntu 9.10 and it worked perfectly.

The main flaw i find in Linux is the opposite. It's small because it's small. Developers don't want to double their efforts to sell to a handful of neck beards.

I don't see Linux's smallness as a flaw. Actually, that tends to increase the signal to noise ratio quite a bit. There are quite a few quality software projects that only develop for Linux and/or OSX and refuse to port to Windows because of the inevitable flood of clueless users that would pull in thus swamping the project in handholding. This is a good example. Very high quality software.

As for the "handful of neck beards" comment, didn't you say something about the supposed childish and condescending tone of the GP? Besides, there are quite a few commercial projects that develop for Linux. But, if you stop and think about it, why would there be large amounts of commercial Linux development in any case? One of the possible reasons developing commercial software is such a niche for Linux is that practically anything you need is in the repositories anyway. And quite a bit of Free sofware spanks the commercial alternatives. K3B smokes Nero. Pidgin smokes YIM, AIM, and MSN Messenger. Firefox and Chrome smoke IE, Opera, Safari, what-have-you. And for the stuff where the Free stuff isn't as good as the proprietary bits, it's still pretty good. OpenOffice is pretty good, GIMP is pretty good, Eclipse is pretty good. Why pay for proprietary software when my needs are already met for free?

Wine isn't there either. i use as much FOSS as i can.

That

VMWare converter is free, and it works with ESXi.

Check it out here.

Mm. Not really. Mono pushes their own libraries - which means you often can't take a mono executable and run it on a windows box without first installing mono libraries. Kind of an ironic twist, really...

Meh. Encapsulate them in a VMWare virtual machine instance like the Deki Wiki appliance. That's a mono app, I believe, running on some distro of Linux or other. I drop it on whatever machine is handy to run it. Dead easy.

The corollary to that rule is that many applications won't run because they're poorly architected and require administrative rights to run

ThinApp and forget it!

... but since there is no Apple-approved way to virtualize OS X...

Wrong.

http://vmware.com/

http://parallels.com/

http://virtualbox.org/

Take your pick, any one of them will run Windows 98 on Mac OS X.

The 13" MacBook Pro fits within your budget ($1199), has hardware virtualisation so can run any Intel-based operating system under VMware Fusion, Parallels Desktop or Virtual Box
You don't get a hot-swap Ultrabay, but you probably don't really need the added complexity. It has a built-in 7-hour battery, has a built-in SD Card reader, has a built-in SuperDrive (Dual-layer DVD±RW, CD-RW) It doesn't have an option for a Floppy Drive from Apple, but any USB floppy will work with it (seriously, does anyone use them anymore? Even Windows doesn't need floppies to load drivers from during the initial install).
It doesn't have an option for a docking station from Apple, but it has all the ports on one side of the machine, rather than at the rear, so it's very easy to plug and unplug - I do this daily and don't miss not having a docking station. If you NEED a docking station, there's a 3rd party one from BookEndz

It has outstanding hardware build quality, comes with a fantastic development environment for free, and can run any of the open-source ones as well, can run Windows XP SP2+ natively on the bare metal, but who wants to reboot these days, so it'll run everything back to DOS in virtualisation. It will also open you up to a new user experience and a new operating environment that you may just end up liking. If you don't you format the drive and install Linux or Windows instead...

There are plenty of reasons why you might choose to host two services on two different machines, even if one machine would have enought power. Things like being able to take one down without affecting the other.

VMs let you keep some of that model, while consolidating down to less hardware.

Plus it makes deployment easy: get your system how you want it, then save it as an image. Now you can clone it as much as you like. Now that there are OSS VM hosts, the commercial virtualisation companies are concentrating their efforts on providing more and more powerful tools for creating, managing and deploying images.

But why should I be saying all this. Read it from a marketeer:
http://www.vmware.com/technology/why.html

Especially cool, is VMs with high availability. Two physical machines, in separate datacentres, each running the same VM in lockstep. Pull the power cord on one, and the user will notice nothing but a momentary pause as the secondary VM takes over.

Well the obvious solution to that is that everyone respect and honor as best as possible the established standards. It makes backwards compatibility far easier if a browser must "flip modes" to respect say, current or past specs, rather than worrying about version "x" of browser "y" etc.

IE is moving towards that now, thankfully, but it also creates the awkward backwards looking issues you describe. The solution? For you possibly this: http://www.vmware.com/products/thinapp/using.html

You can package older or newer versions of IE inside a standalone EXE, and run that as needed. Very handy. Licensing costs I can't cite, but I've seen the demos working for FireFox and MS Office and it's super useful.

JB

"No, it was because while the companies were entirely happy to contribute back changes to libraries they were using, they were not happy to open up their entire finished product, just because they used a GPL'd library along the way."

I believe that to be a bit of FUD. http://www.vmware.com/download/eula/workstation65.html

Apparently, VMWare is a little bit smarter than any of the decision makers you have worked for. VMWare manages to sell a proprietary product which is based on GPL'd code. How do they manage? Could it be, because they are literate?

Actual license terms verbiage:

You may use the Software to conduct internal performance testing and benchmarking studies, the results of which you (and not unauthorized third parties) may publish or publicly disseminate; provided that VMware has reviewed and approved of the methodology, assumptions and other parameters of the study. Please contact VMware at benchmark@vmware.com to request such review.

To the parent poster, IANAL however, if nobody outside your University or dissertation committee ever even reviews your paper, that would certainly be acceptable use under these terms. Otherwise you can shoot them an e-mail, explain your situation and ask permission, you might even be surprised by the response.

"You may use the Software to conduct internal performance testing and benchmarking studies, the results of which you (and not unauthorized third parties) may publish or publicly disseminate; provided that VMware has reviewed and approved of the methodology, assumptions and other parameters of the study. Please contact VMware at benchmark@vmware.com to request such review." From http://www.vmware.com/download/eula/server.html

I think Microsoft gets upset if any other company talks to the BIOS besides them. Here's a page from VMWare that compares their own product to Microsoft's Hyper-V. Hyper-V only debuted as a beta a year ago and they're already compromising company policy to release Linux kernel level code.

Wikipedia page for Hyper-V

I remember VMware implementing this several months ago. It was experimental, I don't know about it's status right now.

Ummm... actually, it's been a feature in VMWare for several years... It was experimental in VMWare 5.0 but it has been standard in the past three major releases: 5.5, 6.0 and 6.5. FWIW, VMWare tends to do major updates in 0.5 increments and you can go from 5.0->5.5 and 6.0->6.5 for free... It's a nice way for only paying for half your major upgrades. Minor upgrades are a smaller decimal value added on (i.e. 5.51, 5.52, etc) and those are always free.

You might consider VMware view. Each person has a thin client on their desk which controls a VM on a centralized cluster. You can centrally manage everyones' VMs and control things like updates and anti-virus.

http://www.vmware.com/products/view/

Here we have a disconnect. If the OS you are using DOESN'T support a feature you need, use one that DOES.

If you want to run POSIX/Unix mostly server applications in a way that allows resources to be controlled at this level, don't use Linux. Oracle/SUN has this nice open-source OS called Solaris that supports containers (zones) that will do what you want.

VMware has this nice product called ESX that can do the same thing https://www.vmware.com/support/esx2/doc/esx20admin_res.html except that you still need another OS.

Which leaves you with ESX + Linux, or ESX + Solaris, or Solaris. Of course the ESX solution allows ESX + Windows as a possibility, although Orcle/SUN has another product called VirtualBox.

Now, it can be difficult choosing -- because there are even OTHER solutions that would work. Remember, Linux (though good, and my preferred desktop and small server platform) is not necessarily the end-all and be-all in OS's.

Also, remember that, even if a platform is "free" (as in freedom, and even, sometimes, as in beer), there are people out there who can help you architect a solution (usually, not so "free" as in beer).

Your personal use would not be met by these solutions, though. Sorry about that (the accurate DOS machine including timing). But you did mention LPARs; and that idea can be supported.

They're building a vCenter for Linux, which is currently in Beta. It is only a matter of time...

Download and install the VM Player
Download and load up one of the many virtual appliances that have web browsing capability (some are as small as 50MB).

http://www.vmware.com/products/player/
http://www.vmware.com/appliances/

Download and install the VM Player
Download and load up one of the many virtual appliances that have web browsing capability (some are as small as 50MB).

http://www.vmware.com/products/player/
http://www.vmware.com/appliances/

Wouldn't it be nice if there was a a tool that let you run outdated software, including all of its OS dependencies, in a little walled garden, allowing you to run modern software for all other uses?

It'd be even cooler if we could get major OS vendors to provide this functionality as part of the OS

http://communities.vmware.com/community/beta/vcserver_linux

Beta, but available and you can run it in a VM on top of ESX so you don't have any additional costs other than the hit of running one extra (fairly low impact) guest OS in your ESX environment.

I'm on the FT team at VMware and just wanted to provide some additional information on FT requirements. You can also find out more about FT at: http://www.vmware.com/products/fault-tolerance/

VMware collaborated with AMD and Intel in providing an efficient VMware Fault Tolerance (FT) capability on modern x86 processors. The collaboration required changes in both the performance counter architecture and virtualization hardware assists from processor vendors. These changes could only be included in recent processors from both vendors: 3rd-Generation AMD Opteron(tm) based on the AMD Barcelona, Budapest and Shanghai processor families; and Intel® Xeon® processors based on the Penryn and Nehalem microarchitectures and their successors.

The current set of VMware FT supported processors are:
Intel® Xeon® 3100 Series, Wolfdale (UP)
Intel® Xeon® 3300 Series, Yorkfield
Intel® Xeon® 5200 Series, Wolfdale (DP)
Intel® Xeon® 5400 Series, Harpertown
Intel® Xeon® 7400 Series, Dunnington
Intel® Xeon® 5500 Series, Nehalem
AMD Opteron(tm) 1300 Series, Budapest
AMD Opteron(tm) 2300 Series, Barcelona (65nm, DP) and Shanghai (45nm, DP)
AMD Opteron(tm) 8300 Series, Barcelona (65nm, MP) and Shanghai (45nm, MP)

VMware maintains a KnowledgeBase (KB) article that provides a current list of supported processors, see http://kb.vmware.com/kb/1008027.

You can download a utility, VMware SiteSurvey at http://www.vmware.com/download/shared_utilities.html to check if your configuration can run VMware FT.

While we understand the end user's desire to be able to use VMware FT on as many processors as possible, VMware's goal is to guarantee that Fault Tolerance works reliably.

I'm on the FT team at VMware and just wanted to provide some additional information on FT requirements. You can also find out more about FT at: http://www.vmware.com/products/fault-tolerance/

VMware collaborated with AMD and Intel in providing an efficient VMware Fault Tolerance (FT) capability on modern x86 processors. The collaboration required changes in both the performance counter architecture and virtualization hardware assists from processor vendors. These changes could only be included in recent processors from both vendors: 3rd-Generation AMD Opteron(tm) based on the AMD Barcelona, Budapest and Shanghai processor families; and Intel® Xeon® processors based on the Penryn and Nehalem microarchitectures and their successors.

The current set of VMware FT supported processors are:
Intel® Xeon® 3100 Series, Wolfdale (UP)
Intel® Xeon® 3300 Series, Yorkfield
Intel® Xeon® 5200 Series, Wolfdale (DP)
Intel® Xeon® 5400 Series, Harpertown
Intel® Xeon® 7400 Series, Dunnington
Intel® Xeon® 5500 Series, Nehalem
AMD Opteron(tm) 1300 Series, Budapest
AMD Opteron(tm) 2300 Series, Barcelona (65nm, DP) and Shanghai (45nm, DP)
AMD Opteron(tm) 8300 Series, Barcelona (65nm, MP) and Shanghai (45nm, MP)

VMware maintains a KnowledgeBase (KB) article that provides a current list of supported processors, see http://kb.vmware.com/kb/1008027.

You can download a utility, VMware SiteSurvey at http://www.vmware.com/download/shared_utilities.html to check if your configuration can run VMware FT.

While we understand the end user's desire to be able to use VMware FT on as many processors as possible, VMware's goal is to guarantee that Fault Tolerance works reliably.

I'm on the FT team at VMware and just wanted to provide some additional information on FT requirements. You can also find out more about FT at: http://www.vmware.com/products/fault-tolerance/

VMware collaborated with AMD and Intel in providing an efficient VMware Fault Tolerance (FT) capability on modern x86 processors. The collaboration required changes in both the performance counter architecture and virtualization hardware assists from processor vendors. These changes could only be included in recent processors from both vendors: 3rd-Generation AMD Opteron(tm) based on the AMD Barcelona, Budapest and Shanghai processor families; and Intel® Xeon® processors based on the Penryn and Nehalem microarchitectures and their successors.

The current set of VMware FT supported processors are:
Intel® Xeon® 3100 Series, Wolfdale (UP)
Intel® Xeon® 3300 Series, Yorkfield
Intel® Xeon® 5200 Series, Wolfdale (DP)
Intel® Xeon® 5400 Series, Harpertown
Intel® Xeon® 7400 Series, Dunnington
Intel® Xeon® 5500 Series, Nehalem
AMD Opteron(tm) 1300 Series, Budapest
AMD Opteron(tm) 2300 Series, Barcelona (65nm, DP) and Shanghai (45nm, DP)
AMD Opteron(tm) 8300 Series, Barcelona (65nm, MP) and Shanghai (45nm, MP)

VMware maintains a KnowledgeBase (KB) article that provides a current list of supported processors, see http://kb.vmware.com/kb/1008027.

You can download a utility, VMware SiteSurvey at http://www.vmware.com/download/shared_utilities.html to check if your configuration can run VMware FT.

While we understand the end user's desire to be able to use VMware FT on as many processors as possible, VMware's goal is to guarantee that Fault Tolerance works reliably.

As a side note, you don't need to buy even a $288 computer to test it; you can get a free one here and here.

One of the things I have done to describe things in the past where a consensus had not been reached is tell clients something was a "common practice". I think any practice has to spend time in the common practice area before it can become a best practice, and I would be explicit with my clients if something did not match best practices. I have also many times told clients that there is more than one "school of thought" when it came to something with contradictory common practices.

Sources of best practices that I have used beyond my personal experience:

• Companies such as Cisco, Altiris and Microsoft typically produce whitepapers and publish other work that describe best practices.
• Any number of forum sites also produce best practices.
• I read books that cover best practices, and study for new skills (presently getting ready to take my CISSP which is all about best practices)
• I read trade journals, attend user groups and hear what the vendor has to say
• I spend time on forum sites
• I continue my education taking classes at night.
• The best resource of all without question have been the people that were senior to me that were willing to let me ask 101 questions on "why" they did something. Learning to listen when someone describes why something was or wasn't done a certain way and to look past the immediate technical solution I thought was best was the most important thing I ever developed.

Even when you have a million apps, it's not that hard.

You know times are tough when one gets modded "troll" for speaking the truth.

VMware FT is not based on continuous memory snapshotting, it uses deterministic record/replay to do simultaneous record and replay. You could find a overview of this technology at http://www.vmware.com/products/fault-tolerance

Also VMware demonstrated a working prototype as early as in 2007
http://www.vmworld.com/community/conferences/2007/agenda/

w.r.t to Xen, doing a proof of concept is one thing but implementing it and supporting it in production quality with sufficient performance is another thing.

Virtual machines don't support hardware accelerated gaming.

That's not true. In fact, DirectX 9.0 Shader Model 2 hardware support it's one of the advertised features of VMware Fusion. Indeed, Fusion can even provide exactly what the submitter is looking for (see the section "Two" Computers in One)

However, to actually do this I'd suggest a pretty beefy computer, like a Quad Core Mac Pro with at least 4GB of RAM, and that's going to set you back quite a bit. It would be very significantly cheaper to, as some others have pointed out, just buy two refurbished computers (I picked up a factory refurb Compaq late last year for $200 CAD that would be powerful enough for the submitters needs). Heck, you could buy 10 - 20 of them for what it would cost for the Mac Pro, and you could make a Beowulf cluster out of them for what the Mac Pro solution would cost (and I say that as a big fan of the Mac Pro itself -- it is simply major overkill for the submitters needs. Technical ability doesn't necessarily mean it's the best possible solution).

Yaz.

Virtual machines don't support hardware accelerated gaming.

That's not true. In fact, DirectX 9.0 Shader Model 2 hardware support it's one of the advertised features of VMware Fusion. Indeed, Fusion can even provide exactly what the submitter is looking for (see the section "Two" Computers in One)

However, to actually do this I'd suggest a pretty beefy computer, like a Quad Core Mac Pro with at least 4GB of RAM, and that's going to set you back quite a bit. It would be very significantly cheaper to, as some others have pointed out, just buy two refurbished computers (I picked up a factory refurb Compaq late last year for $200 CAD that would be powerful enough for the submitters needs). Heck, you could buy 10 - 20 of them for what it would cost for the Mac Pro, and you could make a Beowulf cluster out of them for what the Mac Pro solution would cost (and I say that as a big fan of the Mac Pro itself -- it is simply major overkill for the submitters needs. Technical ability doesn't necessarily mean it's the best possible solution).

Yaz.

Here's a sort of corollary idea... HyperSpace or ESXi with DSL or similar runs on the machine (from flash, of course) and if you want to run something more complicated you load it in a virtual machine. One possible virtual machine would be a LAMP appliance that would make the browser in your machine more useful by hosting web applications; another one would be a storage appliance...

We use VMotion and our only noticeable downtime was due to self-induced implementation issues. I think there might be a slight slowdown during the actual vmotion, but for what we're doing, it wasn't noticeable. A game server might be a different story.