virtualization.info

Last week Sun released the third generation of its hosted virtual machine monitor (VMM) called VirtualBox, acquired from the German company innotek in February 2008.

Sun put a serious effort on this product, launching two major updates in less than 18 months, plus several minor releases.
Sometimes the company strategy is concerning as it seems to suggest that its hosted VMM can compete with a bare-metal VMM like ESX, XenServer or Hyper-V, or that its product can run a virtual desktop infrastructure. Nonetheless Sun has executed very well on the engineering side of this project.

The new version 3.0 introduces the following new features:

• support for up to 32 vCPUs (as long as you have an Intel VT or AMD-V powered CPU)
• support for OpenGL 2.0 on all guest OSes (Windows, Linux and Solaris)
• experimental support for Direct3D 8 and 9 for Windows guest OSes

For now VirtualBox remains open source and free of charge for Windows, Linux, Solaris and Mac OS X. The customers are still wondering if this will continue after Oracle will have completed the Sun acquisition.

Labels: Releases, Sun

Nor Oracle neither Sun announced yet what will happen to the xVM virtualization offering after the acquisition completes.

While waiting for the big news, Sun gives us the first beta of VirtualBox 3.0 which includes major improvements in the graphic area and the support for an insane number of vCPUs:

• Guest SMP with up to 32 virtual CPUs (Intel VT-x and AMD-V only)
• Support for OpenGL 2.0 in Windows, Linux and Solaris guest OSes
• Ability to use Direct3D 8/9 applications and games in Windows guest OSes (experimental)

The beta program can be enrolled here.

Labels: Oracle, Sun

Just two weeks ago, after one year and a half of silence, EMC (or better a couple of its top executives) decided to publicly criticize the Oracle support policy against its subsidiary VMware.

The trigger for such change of directions probably was the acquisition of Sun, which may transform Oracle in a dangerous competitor in the long term.
Rather than replicate on the corporate blog, Oracle answered with the acquisition of Virtual Iron, which is pretty much equal to a declaration of war.

While Oracle VM Server is being sold as a general purpose hypervisor that customers can use for any workload, a few are really using it to run any application but Oracle ones.
The acquisition of Virtual Iron, even more than the acquisition of Sun and its xVM virtualization portfolio, may change this perception and attract a different kind of customers that not necessarily use Oracle products.

So EMC is back on the topic, this time attacking the entire Oracle virtualization strategy.
Once again is Chuck Hollis, Vice President, Global Marketing CTO, to push the button on his personal blog:

…Put in the context of other recent activities, the picture is crystal clear: it appears that Oracle intends to use their market power with databases to force customers to consider their soon-to-be-announced virtualization stack.

…

Almost all of my IT customers want to standardize on a single virtualization layer.  They'd like to use one consistent set of technology to virtualize server applications, virtualize desktop applications and virtualize all the supporting cast of management, security, backup, etc. as well.
And, not surprisingly, they've all chosen VMware as the direction they'd like to go.
It appears that Oracle is going to try and bust up this happy customer-centric vision.  It looks like they're going to use customers' dependence on the Oracle database to force a separately architected, separately managed and separately supported virtualization layer on their customer base.

…

There are a host of useful features in VMware that we'll probably never see in the Oracle hypervisor.

…

Sorry, Mr. Customer.  You'll have to live with a separate, clunky, inefficient and expensive Oracle Officially Supported Alternative.  Oracle wins, customers lose.

…

Sorry, Mr. Customer.  I guess you can't consider Oracle for vSphere fault tolerant environments.  Maybe SQLserver, maybe UDB, maybe something else -- but not the Oracle database since that feature isn't in the Oracle Officially Supported Alternative.  Oracle wins, customers lose.

…

Now, you know I talk to large customers frequently, and -- frankly -- they're pissed at all of this.  With Oracle's latest moves regarding Sun and now Virtual Iron, they can clearly see what's going on here.  And they're starting to figure out how they want to play this.

One smart fellow told Oracle that they were starting a large-scale proof-of-concept around Microsoft's SQLserver as the strategic alternative to the Oracle database.  The Microsoft team was more than happy to help, as were we at EMC.  I don't know how it's going to end, but I bet that Oracle does a special deal with this guy regarding VMware support as a result.

Another guy told me he was starting to contract with one of the focused Oracle boutique consulting organizations for most support issues -- they had no problem with VMware -- and denying Oracle services revenue in the process…

Labels: EMC, Oracle, Sun, Virtual Iron

Less than one month ago Oracle acquired Sun Microsystems, closing one of the most strategic deal of the last ten years.
With Java, Solaris, MySQL, Oracle also inherited the entire Sun xVM virtualization portfolio.

Oracle has its own Xen-based hypervisor, Oracle VM Server, and its own management console, Oracle VM Manager, but it’s reasonable to believe that these two products will merge with Sun xVM Server and Ops Center in the coming months.

Any customer at this point would assume that Oracle has enough resources, engineers and developed code to release a strong virtualization product against VMware, Citrix and Microsoft.
It seems that this is not the case.

Today the company announced a second, major acquisition in the virtualization space: Virtual Iron, for an undisclosed sum.  
This confirms the rumors that virtualization.info reported in March.

So far Virtual Iron raised $65 million in five rounds of investment, one of the highest sum ever granted to a virtualization vendor.
In the last couple of quarters the company reported a healthy growth: 130% revenue growth in Q4 2008 and 65% growth in Q1 2009.
During the last year anyway, many of the original executives left the company, including the founder and CTO Alex Vasilevsky who is now heading the startup called Virtual Computer.

After the release of Xen by XenSource (now acquired by Citrix), Virtual Iron was one the first competitors to adopt it, so the company can probably count on seasoned virtualization engineers and the favor of a certain number of the Xen adopters. 
Despite that, most of the features of the Virtual Iron hypervisor overlap the Oracle VM Server and Sun xVM Server ones.
So it’s not completely clear why the database giant needed to close this additional deal.

Is Oracle trying to become the fourth major player in the virtualization space by consolidating the market?
Or is this an indirect admission that both its product and the Sun one are not competitive enough against ESX, XenServer and Hyper-V?

For sure Oracle has a big challenge now: merging the code and the portfolio of the three products will require a flawless and timely execution.
Without it the company will not be credible to the eyes of its investors, customers and partners, and it will risk a mass exodus of the Sun and Virtual Iron virtualization experts to other companies.

The first virtualization vendor to be disturbed by the new Oracle/Sun/Virtual Iron giant is Red Hat, that plans to lead the open source market with a new virtualization solution based on KVM.

The virtualization.info Virtualization Industry Radar has been updated accordingly

Labels: Oracle, Sun, Virtual Iron

An unexpectedly quick answer came from Sun after the virtualization.info article titled Sun xVM Server may be indefinitely postponed now published last week.

On his corporate blog, Steve Wilson, the Sun Vice President of xVM, wrote a long post detailing how the company has decided to reconsider many aspects of the hypervisor design after the feedbacks received during the Early Availability program.

• Participants requested a "hands off" installation process that could be used to deploy the hypervisor to many servers quickly. The single-system install was "klunky" and not suited to an enterprise data-center.
• Participants requested migration capability for guests between hypervisor instances. Multi-host management was not an add-on option -- it was a requirement for serious use.
• Participants requested more access to the underlying OpenSolaris instance to allow for more customization. While people appreciated having a wrapper "appliance" around the core Solaris instance, it was a problem having a totally custom OpenSolaris distro for xVM Server.
• Customers are now for more interested in larger "cloud" type deployments than smaller consolidation projects

Sun doesn’t seem concerned at all about the delays accumulated so far and doesn’t seem to value the time-to-market aspect as a critical aspect of its strategy in the highly competitive virtualization industry that we have today. As result the company is now:

Thus, we refocused our efforts around use-cases where Ops Center becomes the central way to manage the hypervisor and the underlying hardware. In addition, we've started on a trajectory where we will converge the xVM Server and OpenSolaris lines so that exactly the same codebase is used for both.

Unfortunately Wilson forgot to mention when we should expect xVM Server now, but most of all he completely skipped the discussion about what Sun will do with it as soon as the Oracle acquisition is complete.
Probably it’s too early to say or maybe he can’t speak on Oracle behalf, but it’s clear that the customers concerns now are more about what will happen a year from now, rather than the current state of the product. 
Without a firm and prompt answer on this, nobody will bet on xVM Server and will move its long-term investment money elsewhere.

On his blog Wilson also informs that the new version of xVM Ops Center now includes a xVM Server beta and sports extended support for it.
Last week virtualization.info wrongly reported that Ops Center 2.1 doesn’t have any virtualization support, sorry about that.
Anyway it’s not clear why Sun is doing this: if they already know that xVM Server must be redesigned in part to address the customers feedbacks, why include the soon-to-be-discarded bits in a shipping product?

Labels: Oracle, Sun

One week ago Oracle announced the acquisition of Sun surprising customers, partners, analysts and even the companies employees.
Many projects inside Sun are being discussed right now and some may be just cancelled.
Reading the messages flowing around social network sites like Facebook or Twitter, it’s easy to see how fear, uncertainty and doubts are spreading across the Sun ranks.

One of the most important projects that Sun has in place and that may be negatively impacted by this deal is the release of xVM Server, the Xen-based hypervisor that Sun originally announced in September 2007.

The launch of the hypervisor has been delayed several times. The last update from Sun suggested a tentative first release in Q2 2009.
Potential customers already know that such release would miss some key features like Fibre Channel and iSCSI SAN support. For that Sun suggested to wait an update in H2 2009.

The company original plan was to release xVM Server along with xVM Ops Center, an ambitious enterprise management solution able to handle physical and virtual boxes, providing enhanced features like virtual machines live migration and resource pools.
Ops Center was ready but xVM Server wasn’t, so Sun decided to release the latter as is, without support for the virtual data center management.
This is why we don’t have yet a xVM Server 1.0 while Ops Center is already at 2.0 version.

During the last weekend Sun silently updated Ops Center once again, reaching now version 2.1.
There’s still no sign of support for xVM Server, meaning that the hypervisor GA date is still far away.

The chances that Sun actually releases this product now seems very, very limited.
Oracle already has its own Xen-based hypervisor, Oracle VM Server, and for the vendor releasing an overlapping product just doesn’t make sense.
It’s much more likely that VM Server and xVM Server will merge. If this is not possible for any technical constrain, Oracle may want to stick with its own solution and drop the Sun code: every day Sun postpones its entrance in the virtualization market its competitors gain market share, and because changing the hypervisor of choice is not exactly an easy process, Oracle may decide that a continued development of xVM Server is not worth the effort.

Labels: Oracle, Sun

In mid-March the Wall Street Journal broke the news about an ongoing acquisition discussion between IBM and Sun. The deal never happened and IBM walked away retiring a $7 billion offering.
At this point Oracle jumped in and acquired Sun for $7.4 billion.

For the virtualization industry this is a very interesting move.

Sun is finalizing an entire virtualization portfolio, the xVM family, that includes a much delayed hypervisor based on Xen (Server), an enterprise management solution (Ops Center), a connection broker (VDI) and even a desktop virtualization solution (VirtualBox).
On its end, Oracle announced its own virtual infrastructure in November 2007, which includes a Xen-based hypervisor (VM Server) available free of charge and an enterprise management solution (VM Manager).

So far Oracle kept a low profile and didn’t seriously push its presence in the virtualization market, at the point that most people believe that Oracle VM is just for Oracle workloads. But the company strategy is very different: the database vendor wants to become a fully accredited virtualization vendor and compete with its former partner VMware.

So far Sun has played very nice with VMware (for example, by giving priority to ESX over xVM Server inside Sun VDI) but Oracle is all but friendly with the virtualization player and the attacks are intensifying.

With this acquisition Oracle gets the entire Sun virtualization portfolio, and its entire computing stack (servers, storage, hypervisor, operating system, management layer, connection broker, etc.).
If Oracle plays well its cards here in a couple of years it may become a dangerous competitor for VMware.

Another implication of this deal is that Oracle doesn’t need Virtual Iron anymore.
In the past weeks virtualization.info reported about rumors of a possible acquisition.

Of course the deal has several implications that are not related to virtualization.
One is that Oracle may finally cut its dependence on Red Hat and its Enterprise Linux in favor of Solaris.
Another, which will have a much bigger impact, is that MySQL (acquired by Sun in January 2008) may change into something completely different.

Labels: Acquisitions, Oracle, Sun

Less than one month ago The Wall Street Journal broke the news of an ongoing acquisition talk between IBM and Sun.

virtualization.info reported about the early involvement of Cisco in the bid for Sun, a rumor never confirmed by other sources.

Earlier this week the New York Times reported that the discussion between IBM and Sun has ended and that IBM withdrew its $7 billion offering.

If Cisco was really interested in Sun, now it may be a good moment to reopen the negotiations.

As many pointed out, if Cisco really wants to emerge as a leading player in the server market, it needs all the experience, the credibility and the customers that it can have.
Building all the three things from scratch may take several years, even for a giant like the networking vendor.

Sun can provide all and a virtualization portfolio that may become useful if, for any reason, the intimate partnership with VMware gets compromised.

And by the way, after this failed bid, acquiring Sun is probably much cheaper than one month ago.

Update: It seems that the discussion is still open between Sun and IBM.

Labels: Acquisitions, Alliances, Cisco, IBM, Sun

Yesterday Sun released a new minor update of its desktop virtualization product: VirtualBox.

The new build introduces support for the just ratified OVF 1.0 standard.
Additionally, VirtualBox 2.2 introduces new, welcome features like:

• support for 3D graphics acceleration for Linux and Solaris applications using OpenGL
• support for Apple Mac OS X codename Snow Leopard
• support for up to 16GB vRAM per virtual machine
• support for host-interface networking mode

Anyway there are a couple of other things that make the press announcement quite interesting:

1. Sun dropped the word xVM from the product name.
   Sun included VirtualBox in its xVM product family immediately after the innotek acquisition in February 2008.
   So far Sun has been very careful about its naming convention so this is unlikely a mistake.
   Maybe the company doesn’t want to promote a virtualization portfolio that seems unable to release, or maybe it’s thinking about some name changes.
2. Sun started to use the word hypervisor to describe VirtualBox
   It’s not clear if Sun is doing so to change the users perception about its product (now that it’s sold as a valid platform for VDI environments) but for sure VirtualBox is not a type-1 virtual machine monitor (VMM), aka hypervisor.
   It’s a type-2 VMM pretty much like VMware Player/Workstation/Server, Parallels Workstation/Desktop/Server for Mac or Microsoft VirtualPC/Virtual Server. 
   Messing with the technical terminology just confuses the customers which may lose the trust in the vendor, a mistake that Parallels did as well in the past.

Update: VirtualGuru published an interesting insight about the new virtual hardware that this release introduces.
It seems that the compatibility between VirtualBox 2.2 and VMware 5.x (and ESX) is now very high.

Labels: Releases, Standards, Sun

While everybody waits to know if IBM will swallow Sun, Sun continues to execute its (controversial) virtualization strategy.
The third version of its VDI connection broker, simply called VDI, is finally ready.

Announced in January, as expected the product introduces support for xVM VirtualBox, the hosted VMM that Sun acquired from innotek in February 2008.

As previously highlighted, it’s unclear why Sun believes that its customers may want to run a resource hog like a virtual desktop infrastructure on top of a platform that is much slower than a hypervisor.
The reason can’t be the price: even if VirtualBox is free there are several free hypevisors available at this point.

Anyway there are other features that are more interesting:

• the support for Microsoft RDP remoting protocol (finally!)
• the support for Microsoft Active Directory
• the integration with Solaris ZFS

Sun is pricing this version at $40 per concurrent user / year.

Labels: Releases, Sun, VDI

Just yesterday the Wall Street Journal reported that IBM is bidding to acquire Sun.

Many believes that IBM may want Sun to consolidate its position in the virtualization and cloud computing space before Cisco Systems gets any market share with its upcoming Unified Computing System. But the Cisco involvement in this bid be be much deeper than that.

Several rumors (none of them coming from trusted virtualization.info sources anyway) suggest that Cisco was already in discussion to acquire Sun before IBM came in.

Many have spotted a similarity between the Sun server chassis and the new Cisco UCS chassis (here an example) and started wondering if between the two have an OEM agreement in place to manufacture the UCS hardware.

Design analogy or not (it doesn't seem so evident honestly), at least one of the rumors that confirms the discussion between Cisco and Sun comes from inside Cisco itself.
And this would have forced IBM to enter the bid to block Cisco.

A number of people (like the influential Om Malik at GigaOM) believes that Sun belongs to Cisco much more than to IBM.

Labels: Acquisitions, Cisco, IBM, Sun

The Wall Street Journal is reporting that IBM is in acquisition talk with Sun and considering the source this is very unlikely just a rumor:

If the deal does go through, which could happen as early as this week, IBM is likely to pay at least $6.5 billion in cash to acquire Sun, the people said. That would translate into a premium of more than 100% over Sun's closing price Tuesday.

The impact of such merge would be huge. Of course the big question is what will happen to the many overlapping business units and offerings (servers, storage, management software).

One of the things that IBM may want to save of the current Sun identity is the upcoming and so much delayed server virtualization portfolio dubbed xVM, which includes a bare-metal hypervisor based on Xen (xVM Server), an enterprise management console that can perform VMs live migrations and resource pooling (xVM Ops Center), a VDI connection broker (xVM VDI), a hosted virtualization product (xVM VirtualBox) and a cloud computing facility that can rival with Amazon EC2 (depending on the recently acquired Q-Layer technology).

So far IBM has been happy in its role of virtualization distributor, despite the company invented the technology in the ‘60s. But Cisco is invading the server space and has a relevant interest in VMware. Not enough to buy the virtualization vendor but enough to keep a leadership position in the fastest growing IT market today.

Of course IBM doesn’t look at Sun just to own a x86 hypervisor. Virtual Iron is available and infinitely cheaper than Sun. But earning a complete portfolio for the virtualization market must be a nice bonus to consider.

If the two big will close this deal, HP may be obliged to do something similar to consolidate its position. And Citrix seems so interesting these days…

Labels: Alliances, IBM, Sun

It really seems like Sun has serious issues to deliver its Xen-based hypervisor, xVM Server, in time.

The product, along with the enterprise management console Ops Center 2.0, was initially expected for September 10, 2008 but that day Sun, unable to release in time, just re-announced the product line (the first announcement was published in November 2007).

During the second announcement Sun suggested that its virtualization offering may be released within two months (meaning November 2008) but once again the milestone was missed.

Last week, without any press release, Sun decided to release Ops Center 2.0 anyway, despite xVM Server is clearly unready.

At the same time the French magazine LeMagIT (which usually makes great newsbreaks) confirmed that the hypervisor is in late, and that its release is now postponed to Q2 2009.

LeMagIT also unveils that the first edition of xVM Server will not support SAN and iSCSI storage facilities.
For that, customers will have to wait a subsequent release scheduled for H2 2009.

Labels: Sun

Last week Sun released a maintainance update for its desktop virtualization product VirtualBox.

The new 2.1.2 version introduces a number of bugfixes but also adds the full support for Microsoft Windows 7 beta 1 as guest OS.

As usual the product is open source and free of charge. You can download it here.

Labels: Releases, Sun

Along with its Xen-based hypervisor, xVM Server, Sun is preparing to launch the third version of its connection broker: xVM VDI 3.0.
The Early Access program will be open within the end of this month.

The strange thing is that instead of supporting the upcoming bare-metal virtualization platform, Sun decides to introduce support for xVM Virtual Box, the hosted virtualization product for the consumer market that the company acquired from innotek in February 2008.

It’s unclear why a customer may want to use a hosted platform (which is definitively slower compared to a hypervisor) to run a resource-consuming VDI environment.
Hopefully Sun will share some performance comparisons to understand how many virtual desktops can fit inside a VirtualBox host.

Even stranger is the fact that Sun seems to be absolutely sure that xVM Server must be not supported.
The company is developing its hypervisor since a very long time: the virtualization engine was already in the work when Sun VDI reached version 2.0, ten months ago.
Considering the Sun had more than one year to integrate xVM Server 1.0 with xVM VDI 3.0 and didn’t.

The support for ESX instead is still there, as it’s now clear how Sun bets on VMware solutions much more than on its own.

Labels: Sun, VDI

Yesterday Sun announced a new offering for the SMB segment: a bundle of some of its mid-range servers and SANs with VMware ESX or Microsoft Hyper-V.
Exactly: Sun, which is investing million of dollars on its own hypervisor, is actively pushing two leading competitors.

What’s the strategy behind this initiative?

This is not just a typical offering to pre-install the hypervisor of choice inside a brand new server like every major OEM does since a while now: Sun issued a press announcement, published a dedicated website, highlighted the differences between the two virtualization products suggesting which one is better in which scenario.

An agreement to resell competing hypervisors would make sense if Sun was three years away from releasing xVM Server. But while in late, xVM Server is almost here (as the available documentation demonstrates).

Supposing that Sun can successfully sell ESX and Hyper-V to its customers, what its sales reps will tell them when xVM Server will be out? “Do you mind throwing away the investment that we suggested and that you just made and switch to our hypervisor?”

At that point it will not matter if xVM Server+Ops Center will be a free, valuable platform: customers will invest in learning, deploying and troubleshooting ESX or Hyper-V. How they could consider switching to xVM Server before, let’s say, three years?

This is not the first time that Sun pushes for a competing hypervisor: its VDI solution at the moment only supports VMware ESX and the company couldn’t say when they plan to support xVM Server.

As the Sun VDI is around since a while, it (barely) makes sense that the company tried to push it using the support for ESX as a major selling point. But what’s the sense of this new initiative announced just (hopefully) weeks before the xVM Server launch?

The Sun virtualization proposition seems solid and promising over the long term. Why the company has to sell other solutions instead of waiting for its own?

Labels: Alliances, Microsoft, Sun, VMware

Sun continues to be terribly in late with the release of its Xen-based hypervisor, xVM Server 1.0, and the related enterprise management platform, xVM Ops Center 2.0.

Nonetheless the company continues to tease with bits of information. Today part of the documentation for the two products appeared online.

The corporate wiki reveals some useful information, like the amount of virtual CPUs supported per virtual machine (two at maximum), the availability of Live Migration and resource pools, or the list of supported guest operating systems:

• Solaris 10 (and OpenSolaris and Solaris Express)
• Red Hat Enterprise Linux (RHEL) 5.2
• Microsoft Windows XP
• Microsoft Windows 2003

As declared everywhere in the site, the documentation is still a work in progress and the content must be verified for technical accuracy.
Hopefully somebody forgot to add Windows Server 2008 to the list above. If not Sun will start in a very bad way.

Update: It seems that Sun reconsidered the idea of publishing the documentation before the release time frame. With a bittersweet note Owen Allen informs:

So, earlier this week, Laura and I and the rest of the team were really excited about opening up the docs. It seems that not everyone shares our excitement. Our orders are that we restrict access until the actual GA release, which should be soon but isn't here yet. We're slightly grumpy about this…

It’s a pity because the early exposure of the xVM Server and Ops Center documentation demonstrated how concrete the products are and how many features will be available at their first release.
Knowing this information as soon as possible is critical these days to make a strategic decision on which virtualization platform to embrace.
Going back to stealth mode will just push potential customers somewhere else.

Labels: Sun

It’s more than clear that Sun doesn’t want to be a secondary player in the emerging markets, most of all in the virtualization and cloud computing ones.

In the virtualization space the company is developing its own hypervisor, based on Xen, and an enterprise management system that could rival with VMware vCenter.
Unfortunately both xVM Server 1.0 and Ops Center 2.0 are terribly in late and the customers’ hopes to have an open source hypervisor with enterprise-grade support within a reasonable amount of time are fading away.

Sun is also moving to build a concrete presence in the cloud computing space: its general purpose grid computing facility Network.com is being upgraded, probably to mimic the wildly popular Amazon EC2.

And now the company announces the acquisition of Q-layer, a European company offering a management platform, called NephOS, that simplifies the on-demand provisioning of discrete servers, as well as networks and storage resources, and offers per-user chargeback capabilities.

It seems very likely that the next version of Network.com may feature NephOS to further compete with Amazon.

Labels: Acquisitions, Q-layer, Sun

Sun releases today a new version of its desktop virtualization product: VirtualBox.

Despite the numbering this minor update introduces a couple of interesting improvements:

• Support for OpenGL
• Built-in iSCSI initiator

The product continues to stay free of charge and open source, a couple of things that grant Sun no less than 25,000 downloads / day. Get it here.

The virtualization.info Virtualization Industry Roadmap has been updated accordingly.

Labels: Releases, Sun

Sun has just temporary shut down its cloud computing facility at Network.com.

The online platform was launched in March 2006 and offered the opportunity to buy certain amount of processing cycles at $1 / hour through simple PayPal account.
The user just had to upload its application and specify in how much time he wanted the execution of a certain workload.

This approach was not very flexible, requiring the customers to develop (or re-engineer) their applications for parallel computing on UltraSPARC processors.
Over time the company tried to mitigate this remarkable cost of entry by offering pre-installed applications, like 3D rendering programs.
Despite that, Network.com never became as popular as Amazon EC2, who offers empty Xen virtual machines where customers can install Linux and Windows guest OSes.

Sun stays mum on which virtualization technology will be used for the next iteration, but it’s very likely that it will be similar to EC2, leveraging the upcoming xVM Server (which still is an implementation of Xen).

Labels: Sun

While customers continue to wait for the Sun virtualization offering that seems never ready, the company continues to tease with short glimpses of its interface and feature-set.

This new (commercial) presentation of xVM Server 1.0 and xVM Ops Center 2.0 is the best demo so far of the two products and includes some precious details about the features that will be included in the new virtualization.info Buyer’s Guide:

• xVM Ops Center 2.0 will offer virtual machines Live Migration
• xVM Server 1.0 will be able to import VMware virtual machines

Labels: Sun

Last week a Gartner chart comparing virtualization vendors market shares and their hypervisors’ features generated a lot of buzz as, for example, Oracle VM was reported as more used than Microsoft Hyper-V.

The chart was included in a recent article from Datamation, but Gartner said that it was part of November 2007 report.
The analysis firm has requested the news magazine to update its article with the newest version of that chart, based on projections made on March 2008. Let’s compare the two diagrams:

	November 2007
	March 2008

As you can see the data is remarkably different and even more interesting for several reasons:

• While VMware continues to be unreachable in terms of market share, Microsoft now jumps to the second position, ahead of Citrix, despite Hyper-V is really new on the market while Citrix counted 400 new XenServer customers in Q4 2007, to add on top of the XenSource ones won before the acquisition.
• Oracle, which was as good as Virtual Iron and better than Citrix in the first chart, goes back to the lowest rating for the Management / Automation category and for the Maturity / Stability category.
  It’s an interesting degradation considering that Oracle VM updated its hypervisor to 2.x version in July (after the second projection went out): something terrible must be happened somewhere between 1.x and 1.x.x. to negatively influence Gartner.
• Virtual Iron, XenServer and Hyper-V became more expensive

But the chart is interesting also for another reason: now that we have a guarantee about the freshness of Gartner data, we can compare its market share projections with the IDC ones, published two weeks ago generating another strong flow of comments. The difference is more than remarkable:

• Microsoft market share: IDC (23%) – Gartner (7%)
• VMware market share: IDC (44%) – Gartner (89%)

Please consider that the IDC percent refers to an aggregate data that includes both ESX and Server for VMware and both Hyper-V and Virtual Server for Microsoft. We don’t know if Gartner did the same.

In any case it’s clear that there is a major discrepancy between the projections, putting in serious doubts the reliability of every report on the virtualization market shares.

It would be interesting to have additional numbers from other analysis firms. If Forrester, Burton Group or any other wants to play this game we’ll update this post accordingly.

We wonder, no irony intended, if virtualization.info could be used to provide some reliable metrics about the virtualization adoption to compare with the ones above.
To verify this we published a very simple (9 questions) survey covering just the hardware virtualization (hypervisors) adoption.

This should be considered as a first attempt to measure the market through our audience and it’s expected that the survey design may be unsatisfying for somebody.
If the experiment will be successful we’ll work on more sophisticated questionnaires, evaluating multiple aspects of the market.

Every reader but virtualization vendors employees is welcome to answer the 9 questions. It shouldn’t take more than 3 minutes.
The results of course will be published free of charge online as soon as we reach a fair amount of responses.

http://www.virtualization.info/surveys

Labels: Citrix, Market Trends, Microsoft, Oracle, Sun, Virtual Iron

While Sun puts the final touches to its first hypervisor xVM Server 1.0 and to xVM Ops Center 2.0 (which could be released in November), some more details about the products emerge from a corporate blogs.

In a list of FAQs published there an interesting (yet confusing) indication about the free vs paid strategy about the xVM family:

Q: Within the Sun xVM Portfolio, what's going to be open-source, and what will cost money?
A: Sun xVM VirtualBox and Sun xVM Server will be open-sourced. Some features, though, such as guest migration, are part of Sun xVM Ops Center. Also, service contracts can be purchased for any part of the Sun xVM Portfolio.

The list also includes some additional details about the supported virtual machines formats:

Q: Will Sun xVM Server support native Xen formats?
A: We don't, as there weren't a lot of requests for it.

Q: Can you move guests back and forth between Sun xVM Server and similar VMware solutions?
A: It's only one-way. Sun xVM Server can read vmdk files, but it doesn't save to the vmdk format, so you couldn't modify the guest and then move it back. As far as I know, VMware doesn't support our format either.

Labels: Sun, VMware, Xen

At this point it’s well known that Sun is about to enter the virtualization market with a massive offering: a hypervisor based on Xen (xVM Server), a management platform for physical and virtual machines (xVM Ops Center), a connection broker (Sun VDI) and even a hosted virtualization platform for desktops (VirtualBox).
On top of that it’s easy to guess that the company will release virtualization friendly servers and storage arrays.

As said many times before, Sun has a unique opportunity at the moment, being the only big company that can offer a complete computing stack for virtualization, from the hardware to the software, without bothering its customers with multi-vendor license and support agreements (and issues).

In such position one would think that Sun is focusing all its effort in integrating the components above in a well concerted offering, leaving few things to desire outside the company’s portfolio. But it’s not the case.

No matter if its hypervisor is due next month or so, no matter if there’s already a connection broker that could interconnect with it, Sun continues to enforce its relationship with VMware.

Last month the two announced that VMware Virtual Desktop Manager (VDM) will be sold by Sun with its Sun Ray thin clients.

It’s not hard to figure out where the Sun customers will look at when in need of new hypervisor, and it’s hard to believe that the company will be able to redirect their attention to xVM Server as soon as it’s out.

What surprises the most is that Sun continues to do business with VMware despite the hypervisor interoperability alliance with Microsoft.
Of course the latter doesn’t imply any kind of exclusive loyalty to the Redmond giant, but in practice behind such agreements there are remarkable financial interests, and Microsoft is probably not too happy to know that Sun is encouraging the adoption of ESX rather than Hyper-V.

We’ll see after the xVM family will be out if and how this strategy will be modified and what impact it caused so far.

Labels: Alliances, Sun, VDI, VMware

Multiple bidders are currently running for the FastScale acquisition, as virtualization.info has learned from mutliple sources.

FastScale is a US startup launched almost one year ago, funded by ATA Ventures, Leapfrog Ventures and Hunt Ventures, which offers an innovative approach to reduce the overall virtual machine size.

In the last two years, much before the public launch, the company tightly worked with VMware, probably to shape its solution in a way that would fit the virtualization giant's strategy for virtual appliances.
This is the reason why virtualization.info (wrongly) speculated a FastScale acquisition by VMware in January 2008.

It's unknown if VMware is part of the ongoing acquisition bid now. The only name reported so far is Sun.
The other FastScale partners (HP, IBM, Microsoft and Red Hat) may be part of the bid as well.

Labels: Acquisitions, FastScale, Sun