The VMUG’s “What DBAs need to know about virtualization” session had about 25 attendees — more than a similar session I sat in on at last year’s VMworld, which is obviously a much bigger show. In Maine, VMware’s tier-one database specialist, George Trujillo, talked about how it’s easier than ever to virtualize databases, and he shared some tips for getting management on board with such projects.

“I am absolutely swamped going into large companies, showing them how to virtualize their Oracle databases,” he said.

The first question, of course, is why virtualize databases? Trujillo said DBAs spend 60% of their time on tasks such as creating new databases, moving data and applications between databases and migrating databases to new environments — all actions that virtualization can simplify, using technologies such as live migration, templates and clones.

Most of the discussion, however, focused on the many roadblocks to database virtualization — especially virtualizing Oracle databases. Oracle’s support and certification policies have a bad reputation among virtualization users, but Trujillo said they shouldn’t be deterrents.

Oracle has supported its databases on VMware for years and added support for RAC on VMware in November. Oracle does not certify on VMware, but that’s because the company only certifies to the OS level — not to the hardware level, which is what VMware is viewed as, Trujillo said.

Troubleshooting a virtualized Oracle database is another area that users shouldn’t be overly concerned about, Trujillo said. If you have a problem running Oracle on VMware, Oracle makes you replicate the problem on physical hardware, to show that VMware isn’t the cause. But Oracle does the same thing for problems on physical hardware implementations; if you’re using IBM hardware, Oracle makes you replicate the problem on a Dell or Hewlett-Packard box, to show that IBM isn’t the cause.

“There is not one known error in the Oracle knowledge base that they can attribute to VMware,” Trujillo said.

The final roadblock is an internal one: IT managers and CIOs who fear that virtualizing databases will lead to more trouble than it’s worth. To overcome this challenge, the key is to make a good first impression, because if the performance of the first database you virtualize isn’t up to par, you may not get the chance to virtualize another, Trujillo said.

He noted that the performance hit on virtualized Oracle databases has dropped from 30% to 60% on ESX 2 to 2% to 10% on vSphere 4, but that’s if you do everything right. Avoid memory overcommit and over-consolidation, which can create a fight for resources among your database VMs, and monitor your VMs closely to identify potential problems before they affect performance, he said.

If DBAs can get their higher-ups to be more comfortable with database virtualization, it will be a big step on the road to 100% virtualization. But there’s still another mountain to climb: SAP virtualization, which I asked several VMUG attendees about. They all said they don’t know of any shop that’s doing that.

Photo (cc) by tuppus on Flickr and republished here under a Creative Commons license. Some rights reserved.

The VMware CEO was the latest guest on Fortune’s “Connected with Adam Lashinsky,” where he talked about VMware’s standing in the overall IT market. I recently wrote that Hyper-V vs. VMware isn’t much of a fight these days (and Oracle vs. VMware is even less so), but Maritz made it clear that he takes the threat from both software giants seriously.

“They have unequivocally said that they want to remove the need for any VMware software in your data center,” he said.

Although Maritz vowed to fight off these challenges, he took a pragmatic stance when comparing VMware’s vision of a data center operating system to Microsoft’s success with desktop OSes.

“I don’t think we’ll ever have the degree of control that Microsoft did over the PC market,” he said.

One thing I took issue with was Maritz’s description of Microsoft’s fight against VMware.

“Microsoft decided to start competing with VMware as early as 2002, 2003 and have been working hard ever since, that’s seven or eight years now, and are still not at the same level of capability that VMware is,” he said.

That may be technically true, but Microsoft didn’t really put its weight behind server virtualization until it released Hyper-V in 2008. It’s not completely fair to frame it as a seven- or eight-year battle. And when you look at just the hypervisor, Microsoft has made a lot of strides in just two years.

When Hyper-V first came out, the big outcry was, “It doesn’t have live migration!” Live Migration came in Hyper-V R2, and then the critics said, “But there’s no memory overcommit!” Microsoft will answer that next year with Dynamic Memory in Hyper-V R2 Service Pack 1. And after that, I haven’t really heard any “It still won’t have X, Y, Z!” complaints.

Granted, Microsoft hasn’t taken advantage of this feature parity yet (except maybe in Europe). VMware had a big head start, captured a lot of the market and developed a strong management portfolio around its hypervisor, both in-house and through third-parties. Those are obstacles the folks in Redmond must still overcome if they want to make their threat to VMware a serious one.

Novell, the “strategic buyer” and the private equity firm have all agreed to the deal and will sign it in three or four weeks, according to the Post. In the meantime, let the speculation begin!

Who’s buying Novell? Let’s take a look at some possible candidates and the odds they’re the “strategic buyer”:

VMware (3 to 1): After VMware announced its exclusive virtual appliance deal with Novell in June, the acquisition rumors started circling. That has VMware as the favorite among many observers, including Gartner’s Chris Wolf, who says it’s not a coincidence that VMware is training Novell’s SUSE Linux Enterprise Server organization as part of that deal.

But there are a few major points that don’t make VMware a sure bet. Most importantly, Novell doesn’t fit the traditional VMware  acquisition strategy, which is to pick up smaller growth companies in an “opportunistic” fashion. Plus, some VMware execs have privately intimated that the virtual appliance deal means much more to Novell than it does to VMware.

Oracle (7 to 1): Unlike VMware, Oracle definitely doesn’t shy away from large acquisitions of major companies. Larry Ellison and co. have made it clear they want to control the entire IT stack, and while the Sun Microsystems deal gave them pretty much everything they need to meet that goal, Novell’s SLES would definitely provide a boost when it comes to the server OS. (Novell has about 16% of the market, and Oracle Enterprise Linux another 11%, according to SearchDataCenter.com’s 2010 Data Center Purchasing Intentions Survey. Add that to the 27% market share held by Solaris, and Oracle would control more than half of the server OS market.)

Also, Oracle’s virtualization strategy focuses heavily on VM templates; adding Novell to the fold would strengthen those offerings, while at the same time weakening VMware’s virtual appliances.

Red Hat (9 to 1): You’re the market leader. Your top competitor is vulnerable and for sale. Why wouldn’t you swoop in? No, I’m not talking about Vince McMahon’s WWF and Ted Turner’s WCW in 2001, I’m talking about Red Hat and Novell today. There have been no indications Red Hat is interested — execs may be content to see Novell split up and see what happens under new ownership — but the deal would make a lot of sense.

Citrix (12 to 1): There’s a clear trend toward tighter integration between the OS and the virtualization layer. Hyper-V comes with Windows Server, and in the ultimate example, KVM is part of the Linux kernel. Acquiring Novell would help Citrix out a lot in this area. The interesting subtext to this speculation is the feud between Citrix and Red Hat: As Red Hat tries to woo Xen customers to KVM, Citrix CTO Simon Crosby has suggested that Red Hat Enterprise Linux customers should move to Oracle or SLES.

Microsoft (15 to 1): These two companies have a long history. Microsoft crushed Novell’s NetWare business in the ’90s, and there was significant proprietary-vs.-open source friction for much of the past decade. But after their 2006 patent agreement, the two companies have grown closer, and now acquisition speculation pops up occasionally. The Novell-VMware virtual appliance deal sure wouldn’t last long if this acquisition happened.

Intel (20 to 1): Just a few months ago, Intel wouldn’t belong on this list. But after acquiring McAfee and stating its intentions to offer a “complete stack,” the chip manufacturer can’t be ruled out. A server OS is a pretty important component of a “complete stack,” after all.

Field (6 to 1): There are plenty of other companies that would make sense as the “strategic buyer,” from IBM to Amazon. And as some of the recent head-scratching IT acquisitions have shown, no company can be completely ruled out. Guess we’ll just have to wait three or four weeks to find out for sure.

SAN FRANCISCO — After the splashy, high-level announcement of VMware’s vCloud Director on Tuesday morning, reality set in at a VMworld 2010 session by Kevin Lees, VMware’s global vCloud delivery team lead. Lees detailed the lessons learned from beta deployments at a handful of enterprise and service-provider customers.

He emphasized repeatedly that users looking to deploy vCloud director should “take a stepwise, evolutionary approach.” He advised attendees to start simple by equating one back-end virtual data center with one vSphere cluster. He also recommended that ESX clusters running management utilities and the Oracle vCD database should be separated from resource pools in virtual data centers.

Lees also said careful planning is crucial for organizations looking to deploy vCloud Director at this stage. In particular, he advised users to follow a “60% rule,” particularly if they are setting up an Oracle vCD at a service provider data center.

“What we’ve found is that when you hit 60% utilization of the virtual data center, it’s better if the service provider doesn’t add any more organizations to it so there’s plenty of room for growth,” Lees said.

Organizations should ideally dedicate a “strong project manager” and possibly even a separate operations team to vCloud Director deployments. Well-defined roles and responsibilities, as well as cooperation with all levels of the IT organization, are essential, Lee said.

“This is not something that can be done from organic proof-of-concept processes — that can then just be turned into production,” he added.

Networking and security require special attention in the design phase, Lee continued.

“Design, design and most importantly, design,” he said. “Determine your templates early and get security to review it as early in the process as possible.”

Lees also provided more vCloud Director fine print to consider:

• Only one storage tier can be assigned per virtual data center.
• Currently, there isn’t a single sign-on in version 1 with existing portals.
• Be careful when configuring Lightweight Directory Access Protocol, because it cannot be reset to default values.
• A Red Hat Enterprise Linux (RHEL) utility called compat-libcom_err is required but does not automatically install in RHEL.
• Network File System “transfer storage” is required for users to upload their own virtual machines (VMs) and vApps, but Lees recommended that user self-service follow after initial deployment. At first, he said that IT should control spinning up new VMs, then “evolve into letting end users into the environment directly.”

The Xen vs. KVM debate is not new, but it has definitely been picking up steam as of late. Red Hat Enterprise Linux 6 has given Xen the “let’s just be friends” speech and moved in with KVM. Several major hosting providers are switching their platforms from Xen to KVM. And many Linux diehards say Xen is a huge pain to manage.

The battle lines are drawn, and the fate of the open source virtualization market hangs in the balance. (How’s that for overdramatic? I think I’ve been watching too many “Lost” commercials.)

Anyway, the ramifications of this potential shift may also affect the open source OS market. Leading the charge on this side of the battle is Citrix CTO Simon Crosby, the co-founder of XenSource. Red Hat is trying to move customers off Xen, and now he’s trying to move customers off Red Hat.

“If you approach your virtualized world with a Linux/RHEL based mindset, then I recommend you consider switching to Oracle Enterprise Linux,” he blogged last week. “It is a superior, enterprise class version of RHEL. … Alternatively, if you’re wary of giving Larry more control than he already has over your environment, Novell SUSE Linux offers a superb enterprise Linux platform.” (Note that Crosby linked Oracle CEO Larry Ellison’s name to a picture of his massive yacht.)

I’ve previously said that Red Hat has nothing to lose by switching from Xen to KVM. I meant that in terms of Red Hat’s standing in the virtualization market. VMware, Microsoft and Citrix are way ahead of Red Hat there. (And as Crosby blogged, “Having failed to capitalize on Xen, Red Hat needs a ‘differentiated’ story in virtualization in order to regain credibility.”)

But Red Hat has a ton to lose in the enterprise Linux server OS market, where it’s the leader. Sure, the Linux community may be in love with KVM, but Red Hat butters its bread thanks to the IT admins and systems engineers who work with RHEL. The company can’t afford to turn off these core customers in its pursuit of the virtualization market.

Our guest was Jack Kaiser, vice president of strategic technologies for GreenPages Technology Solutions, a virtualization solutions provider in Kittery, Maine. He’s also a member of our Server Virtualization Advisory Board.

Q: Do you see a lot of Oracle or Sun virtualization products currently in your customers’ environments?

A: Not really. We’ve seen that Oracle has less than a 2% adoption rate on their own virtualization products so far. … Once in a while we will see some Sun Containers in the SPARC space, but predominantly VMware’s still the major player in the market.

Q: Why is it that businesses typically choose VMware, or even Microsoft or Citrix, instead of Oracle or Sun?

A: Obviously VMware is the predominant player and has been for quite some time. Most industry analysts and experts feel they have the best virtualization platform.

Microsoft, being a large marketing company, is coming on strong, and they have some great connections in the server space and with CIOs. And Citrix is really gaining speed as well with the desktop and their XenApp installations. … So I think that’s one of the reasons that VMware, Microsoft and Citrix have a significant leg up on Oracle.

Q: Now that Oracle and Sun are combined, do you see any more possibilities for integrated management or converged hardware and software that they could use to attract more customers?

A: Oracle has great connections with the database part of the business, but we’ve definitely seen some customers migrate off Sun hardware as they try to understand what the acquisition from Oracle will bring.

During yesterday’s five-hour Oracle-Sun press conference, Oracle’s chief corporate architect Edward Screven took aim at VMware: “VMware is integrated with nothing. It’s a point solution.”

Of course, that’s not entirely true. VMware integrates with lots of stuff, both hardware and software, but it’s usually other vendors making that other stuff. (The Cisco Unified Computing System is a prime example.) Judging by VMware’s longstanding dominance in the market, this is a strategy that has worked out pretty well for them.

Now that it has Sun’s virtualization and hardware assets, Oracle is taking a different approach: stressing the value of integrated management (with Oracle Enterprise Manager) and virtualization-specific hardware/software bundles, all from one vendor.

Alone, Oracle and Sun never had much success in virtualization. (When Screven touted Oracle VM as “the best virtualization solution for databases” yesterday, my colleague Bridget Botelho responded on Twitter: “Then why doesn’t anyone use it?”)

And they didn’t announce anything revolutionary yesterday that would immediately cause people to jump on the Oracle-Sun virtualization bandwagon. (Sounds like underwhelming press conferences was the theme of the day, eh, Steve Jobs?)

Clearly, Oracle needs to do more to become a major player. The whole trend of integrated hardware and software for virtualization is something the vendors are pushing; it’s not something a whole lot of customers want now.

Larry Ellison and his Oracle execs know they can’t take on VMware just by talking a good game. So they must have bigger plans in the works. Will they build Oracle VM up using the Virtual Iron technology they acquired last year? Or will they make an even bigger splash by acquiring Citrix?

VMware once enjoyed free reign of the market. Then along came Microsoft, the biggest software company in the world, which became a serious challenger in just a few short years. Now the third-biggest software company, Oracle, is making a similar move. Whatever happens, things are about to get very interesting.

Oracle is sizing up Citrix for a possible acquisition, according to Briefing.com (via The Register). The move would give Oracle its biggest stake yet in the virtualization market — not only in server virtualization, where Oracle VM hasn’t exactly lit the world on fire, but in desktop virtualization, where Citrix is the leader.

Still, an Oracle-Citrix acquisition wouldn’t guarantee virtualization success. A lot depends on what Oracle would do with Citrix’s technology and products. If Oracle were to take the same approach it took with Virtual Iron — discontinuing its products and canceling its reseller agreements — Citrix customers would be up in arms.

Plus, the whole open source issue would come into play with an Oracle-Citrix acquisition. Even though Oracle VM is based on Xen, Oracle is first and foremost a proprietary software company. There would inevitably be friction with the open source community.

As an aside, it would also be fascinating to see the interactions between two of the industry’s most outspoken and controversial executives (Oracle CEO Larry Ellison and Citrix CTO Simon Crosby). Here’s some reaction from around the Web to the Oracle-Citrix acquisition rumors:

Tarry Singh, Sustainable Global Clouds: “VMware and other players will be stunned … Oracle has already declared war on many fronts and by doing a Citrix acquisition it will definitely make a massive statement to the virtualization market.”

Jesse Freund, Cisco analyst relations manager: “Oracle buying Citrix: Um, why? They already have duplicates and triplicates of those assets. Other than to annoy Microsoft, of course.”

Gabriel Pagan, virtualization consultant: “Oracle buying Citrix? Makes no sense, Citrix is too much up Microsoft’s [rear] for Oracle”

We still don’t know what Oracle’s plans are for the Virtual Iron technology it acquired in May, but in light of these developments, a much bigger question is arising about a much bigger acquisition: Will Oracle kill off Sun Microsystems’ virtualization line too?

To answer this question, I did what any responsible journalist would do. I consulted the Magic 8-Ball. Its response?

On one hand, the future doesn’t look good for Sun virtualization. In the wake of the acquisition by Oracle — which is expected to close this summer — Sun has backed down on its plans to offer its xVM Server as a standalone hypervisor. It will only be available as part of the xVM Ops Center management console or the OpenSolaris operating system (but not the commercial version of Solaris).

Judging from this move, you’d think Sun sees the writing on the wall: Oracle wants to present a unified virtualization front, and all these different products from three different vendors won’t help meet that goal.

But on the other hand, Sun just yesterday released its VirtualBox 3.0 virtualization software. Why would Sun execs go through all the trouble of putting out a new release if they thought Oracle was just going to squash it in a few months?

It’s all very confusing for the three companies’ customers and partners. Oracle’s relative silence about the acquisitions isn’t helping matters. (Instead of responding to our questions about Virtual Iron’s future, the company just emailed us a link to its specifics-deficient Virtual Iron FAQ.)

Oracle knew what its plans were for both Virtual Iron and Sun when it bought them. Whether the Magic 8-Ball says “signs point to yes” or “outlook not so good,” any sort of definitive answer would be welcomed.

With vSphere 4, VMware is publicly stating that a single virtual machine can drive an outlandish 300,000 IOPS, up from 100,000 IOPS with ESX 3.5. Then, at the vSphere launch last week, VMware CTO Steve Herrod told the audience that his engineers had just broken the 400,000 IOPS mark that very morning.

Clearly, VMware cares about IOPS.

Just how big of a number is 400,000 IOPS? The average Oracle database only generates about 1,200 IOPS, and the biggest-ever Oracle database was clocked in at 250,000 IOPS, said Scott Drummond, VMware Group Manager, Technical Marketing. VMware’s 400,000 number was achieved using an EMC storage array with a combination of enterprise flash memory and 600 spinning disk drives.

Drummonds conceded that ESX hasn’t always been an IOPS king. Dating back to ESX 2.5, IOPS was sub-par, but the problem was alleviated in ESX 3.0 and rectified in ESX 3.5, he said.

Nevertheless, Drummonds said he still hears the poor IOPS comment from customers all the time, no matter how modest the deployment.

“I was with a small customer in Australia recently, and he said to me, ‘Yeah, but you guys have trouble with IOPS,’” Drummonds said.

Until those comments stop coming, expect VMware engineers to continue trying to prove their case.

“We just want to put it to bed,” he said.