Mike Neil, via his blog, advised that in order to ship on time without sacrificing quality, some of the high-end features of Windows Virtualization (a.k.a. Viridian) have been postponed to the subsequent releases:

·         No dynamic addition or removal of CPU and memory

·         No live migration

·         No more than 16 cores overall

Nevertheless, Windows is already a great server virtualization platform, and customers have other options available to them today.

As a Gold certified partner, SWsoft works closely with Microsoft to support customers on the Windows platform. SWsoft is backed by 24-hour, 7-day support from Microsoft for customers operating Windows-based applications within Virtuozzo virtual environments.

We are committed to make all our virtualization products work well with Microsoft virtualization products. Specifically,

·         Parallels products will support Microsoft VHD disk format

·         Virtuozzo management Tools will manage Viridian and Virtual Server

When Viridian is out, you can expect full interoperability, and Viridian will be a seamless addition to your virtualized infrastructure. Considering that VMware has a very cold relationship with Microsoft, you can hardly expect any interoperability between those products.

With regard to the features still-to-come from Microsoft, here are few more reasons to give Virtuozzo a try.

1)    Virtuozzo has always had dynamic resource allocation – you can change how much memory, CPU power, disk space and quite a few of other parameters of a VE (virtual environment) on the fly, in real time. No other server virtualization technology – not VMware ESX, not XEN – support dynamic resource management.

2)    On Linux, Virtuozzo does live VE migration without SAN or any other shared storage, and we’re making a very good progress in implementing the same capability on Windows. Look for that to be delivered soon.

3)    Virtuozzo supports as many CPU cores (or any other resource) as the host OS supports. And for those who need real scalability, Virtuozzo supports Itanium. Again, no other server virtualization technology scales nearly as well. Virtuozzo supports any number of CPUs, any amount of memory or disk space, any number of devices – on the host and in the guests. And there is no performance penalty. When an application in a guest is using 8 CPUs, it does it with the native efficiency of the underlying OS. As for VMware ESX, if you ran any heavy workload in a VM with 4 virtual CPUs, you know the difference very well, especially if you had few more VMs on the same machine.

Users can realize the benefits of virtualization on the Windows platform today and look forward to more coming from Microsoft.

Hi everyone,
Virtuozzo Blog is moving to http://virtuozzoblog.swsoft.com – please, update your references.

Thanks a lot,

Ilya Baimetov

The MSFT licensing “whitepaper” by VMware generated a lot of buzz. Some even went as far as comparing it to Netscape and suggesting that we all should expect a lawsuit by VMware – oh, please. Few comments in this regard:

·         Before blaming MSFT, VMware needs to take an effort to be more forthcoming with their APIs and formats. Unlike MSFT’s VHD spec, VMDK SDK and specs cannot be freely downloaded and VMware VMI specification is not implemented in any VMware product.

·         Microsoft has been, slowly but surely, changing their licensing for server products towards making it less restrictive – Windows Data Center allows unlimited number of virtual environments, as well as SQL Server Enterprise allows unlimited number of SQL Server instances on one box regardless of what server virtualization product is used.

·         I am, myself, not very excited about a perspective of paying $300 for Vista Business if I want to run Vista on my Mac (inside Parallels), but Vista is not being very widely deployed yet, so it’s not really an imminent problem. And as a Mac user I will stick with XP simply because all Windows software I need runs fine on XP.

Bottom line, for the moment all these problems are all but hypothetical. We should of course let MSFT know if we are unhappy with some of their policies, but it’s way too early to ring the bell. So, let’s tell them, not scream and yell at them – that way we have a much better chance of being heard.

For quite a while, virtual appliances have been one of the hottest server virtualization themes. The idea is to package an application as a pre-built, pre-configured and ready-to-run virtual machine. Virtual appliances provide important benefits:

·         An ISV can pick and custom-configure the OS, which becomes merely a library, much like QT, STL or MFC. There is only one platform to develop and test for, and the configuration is always known.

·         Zero installation and zero configuration – just copy down the image and run it. Plus you get all the manageability benefits of virtual machines – migration, backup/restore, HA, etc.

However, there is a price to pay for having an entire OS embedded into each appliance.

·         OS sprawl – every application now comes with its own OS instance that needs to be maintained and updated. I’ve already written about it here and here.

·         Size – each appliance image contains an entire OS and a swap partition/file for it – at least 1GB, even of the OS instance is highly customized and stripped down. It takes a non-trivial time to send such image over network. It also takes a non-trivial amount of storage to keep a library of such images.

Images are large because the OS inside the appliance, however customized and stripped down, is still responsible for the bulk of the image size. Compared to the OS, application is much smaller and much less complex. Remember the old joke?

A man is walking down the street with two large suitcases. Someone asks him what the time is. The man puts down his luggage, looks at his watch and tells the exact time, temperature, air pressure, weather forecast for next week and closures of the local roads for next month. “Wow, nice watch” – “indeed”, says the man, looking at the suitcases, “but these batteries are killing me.”

Only in this case a proper analogy for the OS would be a pick-up truck, not two suitcases.

·         Desktop – Windows cannot be distributed with an application, which means that desktop appliances, which in my opinion may have even greater potential than server ones, are out – at least for now.

·         Updates – since the OS inside the appliance is probably customized, standard update and patch management software won’t work. This means that each ISV now needs to become a service provider just to distribute the updates for the applications. Administrators, in turn, will have to deal multiple service providers just to get the applications updated.

·         Security – typically, enterprise IT tightly controls which OSes are allowed and how locked-down they are. With virtual appliances, they lose this control and have to completely trust the appliance vendor to take all the necessary security measures.

To be fair, virtual appliances are a great idea and these problems are not at all insurmountable. Virtuozzo already solves most of them, and we know how to solve them in the Parallels products line, too. It’s only a matter of time and resourcesJ

• VMware’s IPO is definitely a sign that virtualization market is quickly going from red-hot to white-hot, which is very good news for any virtualization vendor. Especially for the two-in-one company like SWsoft with Virtuozzo and Parallels.
• A 10% sale makes VMware about as much independent from EMC as Parallels is from SWsoft. It’s unlikely that EMC competitors - IBM, HP, Fujitsu and others - will now consider VMware a much better partner candidate.
• Going public means having a solid long-term strategy. Yet, so far, the EMC/VMware merger has been a financial move. VMware has not done much to integrate its product line with EMC, and has mostly acted on its own.

As far as I know, VMware still does not allow 3^rd parties to benchmark their product. Fortunately, VMware has benchmarked their own product and made the results publicly available. Here (http://blogs.vmware.com/performance/2007/02/a_performance_c.html) and here (http://www.vmware.com/pdf/hypervisor_performance.pdf) you can find a comparison of VMware ESX and Xen.

I applaud VMware for publishing industry-first comparative virtualization benchmark. Here are my comments that I hope will help VMware to make the comparison more accurate:

·         Take the latest version of Xen – Xen 3.0.4 has been around for a while.

·         Include Linux guests. Xen has never claimed any decent support for Windows. As all of virtualization-savvy people know, under Xen approach (para-virtualization), achieving maximum performance would require modifying source code of Windows, which only Microsoft can do. The only version of Xen that is supposed to provide decent Windows performance is Xen Enterprise 3.1 running on VT-enabled CPUs.

·         Use Xen Enterprise on VT-enabled hardware. It is Xen Enterprise (not the open source version) and ESX (not VMware Server or Workstation or Player) that are positioned for enterprise workloads.

·         Use multiple guests and Virtual SMP. The paper says that Xen could not boot SMP Windows and it was not possible to run multiple guests under Xen. Well, I’m sure if Xen Enterprise was used, some of the problems would go away. I’m also sure that they could call Xen tech support and ask for help J.

·         Limit the “performance comparison of the hypervisors” to performance (not features or manageability) and hypervisors (the lowest-level components in V-stack). It may be true that Xen suffers from “the lack of such RAS, scalability, management, and distributed virtualization capabilities”, but it’s not really relevant to the subject of the paper.

·         Add Virtuozzo to the pack? There is nothing in the workloads that would prevent running ESX against Virtuozzo. I’m quite sure that both multiple guests and SMP guests would workJ

The very last question I have – if I run the very same tests myself on ESX and then Virtuozzo and Parallels – will VMware allow me to post the results?

In our recent discussions with customers and analysts, the question of isolation has come up again. The reason is that whenever an article in the press needs a one-sentence explanation of OS virtualization it is something like ’OS virtualization provides better density and performance but cannot run different operating systems simultaneously and does not provide as much isolation between partitions as hardware virtualization technologies such as Xen or VMware.’ The two “not” statements require some clarification.

The statement about different operating systems is generally correct, but one needs to understand that Virtuozzo can run different Linux distributions – such as Red Hat, SuSE and Debian - as long as they use the same kernel.

The issue of isolation requires a more detailed explanation. There are several aspects of isolation:

·         Namespace isolation - Each partition provides a complete virtual copy of the entire system namespace – file system, registry, processes, users, IP addresses, port numbers, routing table, etc. Virtuozzo fully virtualizes all system namespaces and provides the same level of isolation as hardware virtualization.

·         Functional isolation - Each partition and the applications it hosts can be configured independently from other partitions and applications. Each Virtuozzo partition has a complete OS environment in it and provides the same level of isolation as hardware virtualization technology.

·         Fault isolation - A fault in one partition does not affect others. Here, hardware virtualization has a theoretical advantage – a fault in the OS would crash all virtual environments on a given machine, although an OS crash in one virtual machine would leave other VMs intact. In practice, though, more than 90% of OS crashes are related to hardware drivers, which always run in the host partition – the one that manages the physical hardware. So, when the driver crashes, the entire machine goes down regardless of which virtualization technology is used.

·         Performance isolation - A partition cannot monopolize resources of the entire machine and hamper performance of other partitions, yet will receive resources required for its execution. Here, Virtuozzo has an advantage over existing hardware technologies because it provides much more granular control of, and intelligent policies for, allocation of system resources.

·         Security isolation - A partition cannot breach security of other partitions, even if its own security was compromised. Each partition has an independent set of local users, including the administrative account. Because of the reasons mentioned above, Virtuozzo is at least as good as hardware virtualization.

However good these logical conclusions are, the best argument is experience. As of now, there are over 500,000 Virtuozzo virtual environments out there running on the public networks, without firewall protection, typically with about a hundred virtual environments on a single machine. In my opinion, these numbers speak for themselves.

What do you think?

First, IDC (http://www.swsoft.com/en/news/id,11213) recognized Virtuozzo as the fastest-growing among all vendors — 98 percent — in the super hot server virtualization market.

Now, Gartner Group, in the report “Predicts 2007: Brace Yourself for the Next Wave of Server Technology,” says:

“By 2010, mainstream virtualization technology will embrace I/O virtualization, breaking the traditional bonds between physical servers, network switches and storage area network (SAN) switches; by 2010, shared operating system (OS) virtualization will become mainstream.”

You can read the news release (http://www.swsoft.com/en/news/id,11527) and go here (http://www.swsoft.com/en/virtuozzo/gartnerreport06/) to view the Gartner Group report.

Personally, I think that this hill is too steep for VMware. Here are a few thoughts on this:

1.    VMware is a very serious threat to the currently dominant Windows and Linux platforms. By introducing a virtualization dimension, VMware has been able to clearly articulate the value proposition for its platform without having to compete with Windows or Linux feature-by-feature.

2.    VMware will try to take over hardware resource management and infrastructure services provided by the OS – CPU scheduling, memory management, perhaps even the file system. Then, VMware will try to introduce some proprietary APIs that provide access to advanced features and ensure that customers are locked-in to the VMware platform.

3.    However, the value of the OS as a platform is much more than a hardware resource manager and basic system services – it includes middleware, support from ISVs and OEMs. In short, there is a huge ecosystem around the current platform and it’ll be very hard to build another one.

4.    Microsoft and Linux vendors must win - they have no other choice. To do so, they need to change the game back on VMware and embrace the entire spectrum of virtualization technologies – hardware, OS-level, application-level and integrate them into their platforms so that customers can use whatever technology is best for them, instead of trying to stretch VMware to solve all problems.

Share your thoughts with me!

SWsoft CEO Serguei Beloussov

Virtualization continues to make waves in the IT industry. Last week, giants IBM and Microsoft took steps to shore up their position in virtualization.

First, let's look at the IBM announcement on virtual datacenter management.

Is IBM seeking relevance in the fast moving virtualization space? Last week's announcement seems to point to this, but I would argue that to some degree, IBM already has some relevance. After all, virtualization on the mainframe has been around for something like 40 years.

Maybe IBM wants to leverage that history to build some credibility in the x86 server market. Fair enough. Is last week's announcement an attempt by IBM to protect its Tivoli flank against VMWare Virtual Center?

If the answer is yes, this jibes with the notion that core virtualization will become a de facto technology in the datacenter and what really matters to customers is how virtual datacenters are managed.

This is something that SWsoft has been keenly aware of for some time and was the focus of our August 2006 announcement that Version 4.0 of our Virtuozzo software, due early in 2007, will include tools to manage different virtualization technologies in addition to our own. In fact, our multi-vendor management of virtualization goes beyond what IBM is offering since we not only cover hardware-based virtualization such as VMWare and Xen but also OS-based virtualization such as Virtuozzo. In addition, we will extend beyond management to datacenter automation to deliver the broadest range of performance, flexibility and value to our customers, across the entire virtual infrastructure.

Now, to the Microsoft/Novell announcement.

Call me biased, but since I am in the business of virtualization technology, I tend to look at news from the IT world with that perspective. Last week's announcement about Microsoft and Novell partnering is no exception.

Here is my take:

First, I’m happy to report that I am not alone looking at this announcement from this perspective. IDC analyst Vernon Turner says that "There are several key pillars to this announcement, including interoperability initiatives supporting virtualization, the interoperability of system management surrounding virtual servers…"

Microsoft is looking for partners to help them deal with the fact that in an increasingly heterogeneous world, going it alone is not sustainable. We agree. In fact, our own history in the service provider market with Microsoft bears this out. A few years ago, Microsoft had very small market share in the service provider/hosting market. They were completely against heterogeneous tools.

However, as always, Microsoft was able to relatively quickly turn around and partner with SWsoft to provide products and solutions for Windows for hosting customers, which has been a big success for both of us. 

Further, from where I sit, there are two x86 operating systems - Windows and Linux. I might argue that virtualization is becoming a third operating environment. Or, you could say that virtualization and the OS are becoming indistinguishable. That software layer - however it is defined - will be critical to managing tomorrow's virtual infrastructure. And if VMWare is the dominant player, then Microsoft needs to forge relationships that would help it compete in the virtual datacenter.

SWsoft and Microsoft have been working very closely on virtualization, which has proved to be a very viable competitive option to incumbent VMWare, and which we like to think as the next generation of server virtualization. In fact, operating system-level virtualization technology is a perfect complement to Microsoft in its quest to compete in the virtualization space. We welcome Microsoft's announcement. Watch this space for some exciting news in the near future that further proves this point.

This question is posed here (http://www.virtualization.info/2006/10/virtuozzo-more-scalable-than-esx.html) on virtualization.info with an open invitation to hear from readers about their experience with large implementations of SWsoft Virtuozzo and VMware ESX Server. We encourage our users to contribute and write a comment.

If you follow my blog, you know I’ve said many times that Virtuozzo is a production virtualization technology. When it comes to large-scale production environments -- where density, scalability and performance matter a lot – it makes so much more sense to use OS virtualization like Virtuozzo.

This discussion was stirred up in a separate article (http://www.techworld.com/opsys/news/index.cfm?newsID=7199&pagtype=all) quoting David Turner, manager at IT consulting firm IQ-SYS, saying "we get greater scalability with Virtuozzo" and "VMware is good for test and development, and for consolidation of different servers. But on a greenfield site and Citrix server farms, which are all pretty much the same configuration, you don't have a requirement to support different OSes." Finally, he states, there is “room for both vendor solutions to sit side-by-side in the same organization.”

4-Oct-2006

(from our CEO Serguei Beloussov)

We know that the implications of virtualization technology on software licensing are huge. I wrote about this in an opinion column, http://news.com.com/Rethinking+software+licensing/2010-1012_3-6110449.html.

Just recently, Microsoft took a step in the right direction for the users of virtualization technology. In a change effective October 1, Microsoft licensing for the Windows Server 2003 R2 Datacenter edition running on servers with two or more processors are allowed an unlimited number of virtualized servers. As Microsoft notes (http://www.microsoft.com/windowsserver2003/evaluation/news/bulletins/datacenterhighavail.mspx), “Unlimited virtualization rights significantly extend the savings customers can realize through server consolidation on the Windows Server platform.”

To quote from a famous phrase that has surfaced in the news again this week, “This is one small step for (a) man and one giant leap for mankind.” We, at SWsoft, applaud what Microsoft has done, which is the best interest of customers. At the same time, it must be noted that there is still much to be done by Microsoft and every software company to clarify licensing for the users of virtualization technology – especially operating system-level virtualization, such as SWsoft Virtuozzo.

With high-performance, highly-efficient Virtuozzo software that offers much greater density per server than other virtualization technologies, this revised Microsoft licensing is particularly good news and of greater value for customers. While the Microsoft move is a good step in the right direction, there is still a need to enable virtualization friendly licensing for Enterprise and Standard Editions of Windows Server.  Further, our customers expected Microsoft to treat virtual environments (VEs) very differently from virtual machines (VMs) since they leverage one Windows instance and not multiple ones.

This issue isn’t going away and given the industry sales for virtualization technology and, in particular, our own at SWsoft this is becoming a more pressing issue every day. We’re encouraging software suppliers to re-think traditional licensing strategies to provide more clarity for users of virtualization technology. We maintain virtualization actually benefits software suppliers because it allows for new usage scenarios and could increase their license revenue – while at the same time decreasing cost per user for their customers.

We encourage customers to check license agreements carefully and negotiate hard with suppliers. We’re at an inflection point as the uptake in virtualization technology increases rapidly with high stakes for software suppliers and users alike.

Serguei Beloussov, CEO
SWsoft

(commentary from me)

To make the long story short, Microsoft vastly simplified virtualization licensing for Windows Server 2003 R2 Datacenter Edition. You can run unlimited number of virtual environments – regardless of the server virtualization technology - yet pay only for a single OS license.

This is good news for VMware users – they don’t have to pay for or track usage of per-VM OS licenses.

And here is why this is also VERY good news for Virtuozzo. People will inevitably start looking at Virtuozzo trying to maximize usage of the “free” Windows licenses.

With “free” licensing, it makes sense to buy a very powerful server – 8+ CPUs, 16+ GB of RAM and try to run as many Windows OS instances on it as possible. And it is the OS virtualization technology (Virtuozzo) – with much greater density, much more efficient application management and much more sophisticated resource management - that really does the best job here. The funny thing also is that unlimited licensing encourages people to create homogeneous environments – why waste any computing power on running Linux on a machine that can run an additional instance of Windows for free? And, if you start moving towards all-Windows machines, you don’t really need Xen or VMware – Virtuozzo will do everything, only cheaper, faster and more efficiently.

Just recently, I received an e-mail from one of our partners. They have been testing several server virtualization technologies – Virtual Server, VMware ESX and Virtuozzo – and found that Virtuozzo, even with just a few virtual environments running on the machine, outperformed the others by a factor of two. This does not even take into account that Virtuozzo can provide much greater density in terms of the number of virtual environments that can be run on a server. Overall, Jess realized substantial manageability and performance advantages by choosing OS virtualization over hardware virtualization. That’s the good news.

The bad news is that Jess had to pull his VMware results because the VMware EULA prohibits publishing any benchmark results unless the methodology is approved by VMware. Can you imagine an individual blogger specifying the methodology in detail and requesting VMware to approve it – all just to do a blog post :)  I’m not sure what VMware is afraid of, but I, personally, encourage everyone to test the hell out of Virtuozzo. We’d appreciate if you let us know when you publish it, so that we can look at results. I believe that customer comparison tests provide great insight and are a very valuable form of feedback, which ultimately helps us make a better product.

Virtual Iron expressed its skepticism about paravirtualization. I already said that I don’t see much benefit in this approach, and let me explain why.

Hardware virtualization. - Hardware virtualization will probably catch up with software virtualization in 2-3 years. Yes, I know, VMware published a report that says that software virtualization is still ahead. But for how long? Intel is investing so much in its virtualization technology, I bet the results will come out sooner than later.

No standard hypervisor. -  This problem is more serious. As of now, we have Xen, Microsoft Hypervisor and VMware – all incompatible with each other. To me, it means that each and every OS has to be modified (para-virtualized) for each hypervisor separately.

Who’s going to do the work? OS vendors? Hypervisor vendors? Are you sure MSFT will make a genuine effort to para-virtualize Windows for VMware unless forced by antitrust authorities?

So far, VMware has been para-virtualizing different OSes on the fly using binary rewriting. When you make such changes static to the OS, you instantly have a multitude OS versions to service. Patches may actually be different for different hypervisors. Do we want “patch sprawl” in addition to “server sprawl” and “OS sprawl”?

What do you think?

The publishers of SearchServerVirtualization are sponsoring a seminar on server virtualization. This seminar is free for attendees and you’ll hear directly from experts and users, including Lukas Loesche from Arvato mobile, who will talk about his experience with Virtuozzo and why they chose OS virtualization.

It’s not going to be a huge event, so you’ll get an opportunity to talk face-to-face with some very experienced people. The seminar is intended for IT managers who want to use server virtualization to increase server utilization and manageability while reducing server glut - highly recommended.

In this article, Information Week magazine tells us that Gartner wants the next (after Vista) version of Windows to be componentized using virtualization. They’re talking about coarse-grained componentization targeted to simplify deployment of OS components and applications.

In my opinion, hypervisors will (as Virtual Machines do now) provide a solution for compatibility problems, but software management will be solved by combination of OS virtualization (specifically, Virtuozzo templates) and perhaps  application virtualization (e.g. Softricity recently purchased by Microsoft). Once virtualization technologies mature, there will emerge a “hybrid” approach that will be able to solve a wider spectrum of problems more efficiently than any individual technology alone.

Bottom line, one way or another, all virtualization technologies will find their place and all of them will play a significant role in changing the way operating systems and applications are deployed and managed. However, it seems that no one should think there is even a slightest chance of getting virtually componentized Windows before 2010, and, realistically, more like 2012.