Slashdot Mirror
VMWare Eats Microsoft's Lunch
feminazi writes "Jeff Boles attributes VMWare's dominance over Microsoft in the virtualization market to a combination of product depth and focus, but especially to the fact that 'VMWare is actually delivering Microsoft's product in the way that Microsoft should be delivering it.' The ease of GUI but with those enterprise-ready traits that Microsoft is still struggling with: application separation, and decent resource utilization."
Today VMWare is eating Microsofts lunch, it's easier to use and performs better. But discounting MS would be a tragic mistake, this wouldn't be the first field MS entered late only to dominate later on. With their size, cash and market if MS wants to own the VM market on Windows eventually they will.
For TFSummary: "Microsoft is still struggling with: application separation, and decent resource utilization."
And above all: security. Surely.
Microsoft never really had a chance...did they? VMware simply rocks and supports a lot of platforms. Mhy Microsoft felt Virtual Server was even necessary was beyond me. Even when they give away free sessions, VMware STILL beats them hands-down.
I ask again, what were they thinking exactly?
I'm not a troll, but I play one on Slashdot.
I am a VMWare junky, I have been using them for years and they continue to deliver quality software and upgrades are always jammed with new features.
Unfortunately....
Since MS gives Virtual Machine away to big Co.'s I am forced to use this horrible product at work. Once again MS finds a market, makes a far inferior product, then jams it down your throat by giving it away to their big customers.
//TODO: Insert catchy phrase
The article offers no details about how or why VMWare is better than the Microsoft offering, and in fact, doesn't even mention the name of the MS offering. No details about the features offered by either product. No details about the history of either product. No details about the "product depth and focus" attributed to VMWare. "VMWare is doing this by really giving us what we need from the MS Windows OS, that Microsoft has never been able to deliver. VMWare is actually delivering Microsoft's product in the way that Microsoft should be delivering it." um, great, but, um, how is that? What is it that VM Ware delivers and how should Microsoft deliver it and how is VMWare delivering it? This is an op-ed piece, with an a-subtle antimicrosoft slant. Little more.
I'll also add to that and note that since Microsoft bought Connectix out that the product has gone to hell in a hand bag (looking at it from the Mac end, of course :). With each release (Connectix) it just got better and _faster_. That is no longer the case it seems -- today it just seems to be full of Microsoft induced bugs and problems (as usual).
:) is much faster than the current iteration of VirtualPC.
With Apple going to Intel I'm frankly waiting (with baited breath) for VMWare to come out with a version for OS X [Intel]. It is then that I'll probably get serious about buying a new Mac. In the mean time I find that using VNC to tie to a real Windows box (if/when needed, albeit more and more rare
Over the last year or so, I've heard a lot of people in the industry talk about how VMware is fighting a losing battle against Microsoft in the server virtualization market. Really though, I don't see Microsoft beating VMware anytime soon. Here's why:
First, I don't think anyone in their right mind is ever going to truely believe that Microsoft can be entirely agnostic when it comes to what OS you run in a virtualization layer. I just can't see the Linux crowd ever fully buying into the notion that Microsoft will support Linux as a virtual server with the same zealous dedication as they'll support virtualization of Windows servers. We've all seen too many instances in the past where Microsoft has teaked some application to take advantage of their inside knowledge of Windows, at the expense of some other vender's application or operating system. I can't imagine, given this track record, that Microsoft will continue to resist the temptation to shaft everyone else in the virtualization market, ensuring that Windows continues to dominate. This idea alone will seriously retard Microsoft's ability to compete with VMware. I doubt that anyone at VMware really gives a rat's ass what you run in ESX server; Microsoft, on the other hand, will never be able to make the same claim.
Additionally, as I see it, there's also little advantage for Microsoft to expand the number of operating systems they support under their own virtualization layer. Every time they add support for an additional OS running in the virtualization layer, it gives their current customers more choices to run some other operating system that *isn't* Windows. Sooner or later, someone on the Windows server sales team is going to figure that out, potentially putting preasure on the virtualization team to do a half-assed job with anything that doesn't sport a Microsoft logo. Ultimately, I predict that this is going to ensure that Microsoft's virtual server offerings will be the most limited in the market. VMware, of course, won't be bound by the same demands. Every time they expand support for additional operating systems, it makes their products that much more attractive to buyers.
Finally, I suspect that Microsoft will decide at some point in the future that what they really want to do is to build virtualization into the Windows operating system itself. This is the only strategy that makes sense in the long-term. It keeps customers buying Windows while answering the need for server consolidation/management that virtualization brings to the table. In the end, it will put distance between what Microsoft offers and what VMware offers, leaving the independant OS virtualization market squarely in the hands of VMware.
While VMware GSX Server is certainly a better product, it's ESX Server (with its addons) that is in a completely different league to anything Microsoft offers. Further, now that VMware has released VM Infrastructure 3, Microsoft just got smacked down. Hard. Each VM can now use up to 4 CPUs and 16GB RAM. As far as performance goes, on average, you will get 6-8 VMs running per physical CPU core, although you should outfit each server with twice as much RAM as you expect it's workload to require.
Most techs are now familiar with basic single-box virtualisation, but aren't familiar with the virtual infrastructure offerings. It's here that VMware is truly revolutionising the commodity x86 server space. Without the VM Infrastructure concept, but using virtualisation you're effectively trading off reliability for utilisation efficiency. With VM Infrastructure you're gaining both. Let me explain.
Picture this: hundreds of VMs scattered around dozens of physical servers. Under VM Infrastructure, any time a given server's resources start to struggle with the load, the VMware software automatically and seamlessly migrates some of the VMs to another server that has resources to spare. When local server diagnostics identify that a critical piece of hardware is on the way out, all VMs are automatically and seamlessly shifted to other available servers. All this while allowing you to specify per-server policies on minimum and maximum resources (CPUs, CPU time, RAM, storage capacity, disk I/O, network bandwidth, network I/O, etc). This is possible because VMs interact with virtual hardware devices rather than the underlying hardware. This means you can most a running instance from one physical server to another and there are no hardware differences visible to the guest OS.
Basically, you not only ensure that you don't have idle servers sitting around, but you actually increase your availability by mitigating hardware failures and levelling resources throughout your pool of servers when load for given VMs increases.
This works by having all storage on a SAN. This means you don't have wasted disk sitting at individual servers. It also makes your storage subsystem extremely reliable and scalable while simultaneously amortising it's cost across multiple servers. The cost of storage goes down on a per-server basis while the reliability of your storage goes up. It also means your individual servers can have a smaller form-factor as you don't need any disk space on nodes save what's required to boot the virtualisation layer.
Using VMs can make backups much, much cheaper to implement. VMs are just files waiting to be copied to media.
If you don't think that's enough (it was for me!), think of how much easier and cheaper disaster recovery becomes. You just need to replicate SAN-to-SAN and your entire server pool is effectively mirrored offsite. In the event of disaster you can simply disable all non-critical VMs (e.g. DEV, TEST, UAT and low-priority PROD), so you don't need to mirror your (now smaller) server pool at the backup site, just enough to bring up the critical production services. In the event of a disaster you've always got the option of then buying additional servers to host non-critical VMs as time permits. Since it's a SAN that's required for storage, if you're not too concerned about non-critical VMs, you can simply keep them on a separate LUN to the critical ones and not replicate that. In the event of a disaster, you can order more disks, and in the meantime you save on SAN-to-SAN replication traffic costs and bandwidth.
Finally, and this is a key selling point to infrastructure staff and customers alike, a SAN-based virtual infrastructure allows extremely rapid deployment of new servers. Let's say a project manager contacts the service desk with a request to provision two new environments (TEST and UAT) for a new development project, each with a webserver, an application server and a database server. The service desk sources the software licenses (either from a pool of spares or
I should clarify my opinion on this (and it's just my opinion). The virtualization marketspace is large and there are many areas to grow into. Parallels is targetting certain users and use cases. VMware is targetting certain users and use cases. Xen is targetting some. QEMU. Virtual Server. They all have users and are being used for different things. I think Parallels is playing things smart by getting into the Mac market early and targetting the users there with a low priced product, and that may secure a position for them there. But it's not at a point where I feel VMware has to fear it. Keep the company in mind, yes. But "VMware fears Parallels in a big way" is a bold statement to make :)
(My opinion is my own, and not necessarily that of VMware.)
No, it's an erroneous reproduction of the older idiom "I couldn't care less". Given that it relies entirely on vocal intonation (or prior knowledge of the "meaning") to be anything but gibberish, its origin was almost certainly from mispronunciations and/or "mishearings" via regional accents - probably of the phrase originating in written material - followed by repetition.
Incidentally, the word you're after is "sarcastic", not "ironic".
Outside of the US, "I could care less" is seen as the nonsensical error it is and correct usage is "I couldn't care less". Even *within* the US, people with a decent English education realise the former is a perversion after it requires an explanation and sufficient linguistic/mental gymnastics to be meaningful.
The meaning of "I couldn't care less" is quickly and easily understood, whether the phrase is heard or read. The same cannot be said for "I could care less". This is because the latter is (supposedly) a shortening of a longer phrase that requires a suitably sarcastic intonation to communicate meaning, whereas the former is coherent on its own.
DOS isn't an original MS product either, guess where that went. Come to think of it .. the GUI isn't a MS innovation .. neither is networking, email, ftp, telnet, gopher, http .. the list goes on and on. Yet who holds the market?