Who am I to second guess it. Virtualization is hot. In the past week VMWare went public closing at the end of the first day with a $20B market cap and Citrix agreed to buy XenSource for $500M.

WOW! This kind of activity make the bubble days pale in comparison.

I mean okay VMWare, hot company, fast top line revenue growth but also accelerating expenses. In 2006 the company reported revenues of $704M and net income of $87M or 12% of revenue. In 2005 VMWare reported $387M in revenues and net income of $67M or 17%. So revenue was up 82% but net income is declining on higher spending all around. If the company continues to grow revenues again this year at 82% the current $20B market cap means a 28x trailing twelve months revenue and a 15x current run rate revenues.

Compare VMWare to the Bladelogic IPO or the Opsware acquisition by HP and it looks pricey by comparison. But, given the market is so starved for growth stories it kinda makes sense.

But, XenSource is another story. The company just started shipping product in January of this year and according to Business Week’s Aaron Pressman and 451’s Rachel Chalmers, XenSource had less than $1M in revenue over the past year. That means Citrix paid 500x trailing 12 month revenues. WOW! Okay yeah Citrix needs a new game, they’ve been looking for their 2.0 story forever. But I mean anyone could pick up Xen and integrate it. It’s open source for crying out loud! Is there really $500M worth of value in the XenSource management tool for the Xen Hypervisor? Citrix seems to think so.

I think this points more to the continuing trend of acquiring hot technologists not so much technologies. Despite what Matt Assay writes Tim O’Reilly may just be right on with his assessment of Open Source companies eventually being bought by proprietary companies. These acquirers are not buying the software or the licensing model, but the people. The licensing model doesn’t matter. Customers pay for innovative ideas that solve problems easier, cheaper and faster and they’ll buy it in almost any form if it works better than what they’ve got.

Sure Cirtrix paid a whole lot. But they might be right about owning the brightest minds in virtualization software. If virtualization is the future and they captivate the biggest thinkings perhaps 500x is not too much?

What’s so great about agile software development? Sure engineers think its great. I think it helps them feel empowered. Product managers think its great too, but secretly I think they’re still trying to figure it out. Apparently Oracle thinks it’s great. The company just paid $495 million for Agile Software Corporation (Nasdaq: AGIL), representing a 14% premium.

As a user I just want to know, what’s in it for me?

We just shipped Splunk 3.0. During the past seven months I’ve heard agile, scrum, sprints — all the cool concepts that are part of this revolutionary framework to spur innovation, more efficient product development cycles and a tighter loop with customers. So why then did it take us half a year to release something? I mean we’re still a start-up after all.

Don’t get me wrong, Splunk 3.0 is fantastic. All kinds of amazing, ground breaking stuff. I’m running it on my OS X laptops and desktops to monitor ps, top, iostat, vmstat and it comes in very handy for figuring out why things crash by indexing everything in /var/log. It also now generates cool interactive reports and Flash graphs of who’s getting the most SPAM on my FC5 mail server running Sendmail and Dovecot IMAP front ended by my Ironport box. Look out mom I’m gonna charge you extra. Check out all the new Splunk 3.0 features and download your own free copy.

What I’m getting at is we had many of the latest greatest features available months ago. But they were unavailable to me and other end users. Turns out even though we’re developing software by living and breathing much of the agile manifesto, we’re still struggling with how to conceive of and adhere to an agile release cycle which incorporates more than just engineering. Sure if we wanted we could just deliver everyone half broken releases every few days without an installer and lousy or no documentation. However, our business thrives by satisfying users with a complete and easy to use product that is high performance and high quality. See IT people are much less curious and tinkering when it comes to solutions and tools than developers. They just want the stuff to work. Turns out the massively complex infrastructure they’re managing gives them enough to toy with.

Sometimes we get customers saying they can’t install releases more than two or three times a year, but I have a hunch with a technology that’s moving as fast as IT Search — we need to release new features faster. So we’re working on moving this agile, defactored architecture we have underneath the covers all the way to the surface. And in the process challenging ourselves to move the whole stack to faster release cycles. We get input through our live product roadmap every day and we’ve got innumerable parallel engineering scrums going on at any one time. But it’s the last mile of figuring out what pieces to release when and how to package them which always seems to slow things down.

So what can you expect to see from us in the future?

• Smaller units of features.
• More frequent releases.
• An active beta program.
• A well supported Splunk developer network.

We’re VERY interested in your comments on how we’re doing and feedback on what you’d like to see. Feel free to contact me or our support team with your thoughts.

Happy Splunking!