March 2010

Recent changes to Solaris licensing could further encourage Solaris 10 users to consider Linux and fewer new users to consider Solaris at all. If you’re a Solaris customer, don’t overlook this license change.

While the “Linux vs. Windows” competition is often played up in the press, the reality is that Linux workload overwhelmingly comes from Unix migrations. Being the largest Unix platform, Sun Solaris has faced stiff competition for lower-end workload against Linux for the better part of a decade. As Linux usage and features have grown, so to has the applicability of Linux in more mission critical distributed environments, an area historically associated with Unix and Sun Solaris. Sun attempted to slow, or even reverse, this trend in 2005 by offering Sun Solaris 10 free of charge. Sun expected to sell subscriptions to customers seeking support and defect fixes in order to drive revenue.

The Wayback Machine documents the Sun Solaris 10 license as it was offered back in May 2005.

“Obtaining an Entitlement Document is simple. On the Solaris 10 Get It page, select the platform and format you desire from the drop-down menus, and then click the Download Solaris 10 button. When you arrive at the Sun Download Center, either sign in or register, ensuring that a valid e-mail address is part of your Sun Download Center account to receive the Entitlement Document. Fill out the Solaris download survey, specifying the number of systems on which you are installing the software. Once you have completed the survey, you will be redirected to the Solaris 10 download page for downloading, and your Entitlement Document will be sent to your registered e-mail address.”

Simply put, register with a valid email address, download Sun Solaris 10 and receive an Entitlement Document to use Sun Solaris 10 without support and for as long as you wish. The terms and conditions were unchanged until at least June 2008, the final copy of the license found on the Wayback Machine.

It seems Oracle has appended this sentence to the license paragraph above:

“Please remember, your right to use Solaris acquired as a download is limited to a trial of 90 days, unless you acquire a service contract for the downloaded Software.”

The obvious question is why would Oracle make this change?

Oracle must have determined that a large enough number of Solaris 10 users are doing so without a service contract. These users are attractive targets to convert into paying customers. However, what percentage of these unpaid users would Oracle have the ability to upsell into if the user had not been able to use Sun Solaris 10 unsupported for more than 90 days and instead migrated to unpaid Linux?

Free User to Paying Customer?
It’s interesting to note that Red Hat just reported that one of its eighteen deals over $1 million in their recent fourth quarter was from an unpaid user converting to a paying customer. By limiting the use of Sun Solaris 10 to 90 days, it’s hard to imagine that Oracle will build a set of customers that could later be upsold. In effect, the new license severely limits Sun Solaris 10 as a viable alternative to Linux for net new customers. But maybe this is a calculated business decision on Oracle’s behalf. If a net new customer needs the advanced capabilities of Sun Solaris, they’ll pay for that capability and the associated hardware from a Unix vendor. Sun being the largest Unix OS vendor, they’ll win their share of this new business. If the net new customer doesn’t need the advanced capabilities, it’s unlikely the customer would even consider anything beyond Linux on an x86 system.

Existing unpaid users of Sun Solaris 10 with an Entitlement Document could potentially continue their usage without restriction. The license change does not appear to be retroactive. However, I’m not a lawyer, so be sure to validate your unpaid usage of Sun Solaris 10 with your legal department and/or Oracle.

On the surface, this new license has little impact to existing paying customers of Sun Solaris 10. However, in reality, limiting the number of potential users and customers in the Sun Solaris ecosystem can’t be viewed as a positive outcome. With upwards of half a billion in Solaris operating system revenues per year, it’s difficult to argue that Sun Solaris is “dead or dying”. However, IDC’s estimates indicate that Sun Solaris revenue has declined at approximately 10 percent annually from 2006 to 2008. The 2009 data, due out in the summer of 2010 would likely continue this negative trend. Over the same period, Linux, and its poster child vendor, Red Hat, has grown at least 15 percent plus range annually. With this data in hand, one might expect Oracle to encourage Sun Solaris usage; a 90 day trial hardly achieves this.

What do you think about the new license?

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Returning from OSBC last week, Gartner analyst Brian Prentice writes a post questioning the conventional wisdom of open core licensing. Brain argues that open core licensing provides significantly less value to the customer than it does to the vendor or the vendor’s venture capital (VC) backers. I agree with his conclusion, but don’t think that vendors or VCs should shy away from open core licensing as a result.

Brain writes:

“The VC communities’ interest in open source, as I see it, is based on the view that a project’s associated community will lower development and sales costs. That allows them to build an attractive proposition when selling the company.”

Let’s begin by looking at single vendor controlled projects. These are the most common examples of successful, in terms of revenue potential, open source projects. If a VC still believes that a vendor will benefit from lower development costs as a result of selecting an open source development approach, they need to wake up. Look at virtually any single vendor controlled open source project and you’ll note that the vendor severely limits contributions from third parties. Amongst other reasons, ownership of the copyright and being able to track pedigree of the code contributions are key justifications for limiting outsider contributions. These are necessary steps for offering a commercially licensed product with confidence. Single vendor controlled open source projects do benefit from testing, documentation and translations from the user community. But these costs pale in comparison to the actual development costs. As such, I’d argue that lower development costs are not really a reason for VCs to invest in an open source project.

I would posit that lower selling costs are possible through an open core-based open source distribution model. However, Brian calls into question the value of open core distribution:

“The first is that open-core is a largely a re-tread of tired, old SMB packaging strategies which have almost universally failed in the market. Businesses don’t blindly jump into a free open source offering and then upgrade to a full-cost, proprietary product like it was some stimulus-response behaviour. From my experience they assess these products, from day one, based on the full version. That eliminates any sales benefit from the open source component of the overall strategy which, in turn, makes these open-core vendors just like any other small software provider slugging it out in a crowded market space.”

Businesses don’t blindly jump into any offering, free or otherwise. But developers absolutely do. I’m fairly sure that the Gartner customers asking for assistance regarding open source policies or open source product guidance are doing so because developers at the customer site are already using, or would like to use, an open source product. As long as the product is free, easy to acquire, easy to learn, has industry success or buzz and helps the developer be productive, the developer will at least evaluate the product. This is true for SugarCRM, Tomcat, Geronimo, Spring, Alfresco, Drupal or Eclipse, as much as it is for closed source products. For instance, we’ve see strong uptake in WebSphere Application Server for Developers since making it available for no charge last year.

A happy developer quickly becomes a proponent for the vendor’s product. This opens the door to a sale predicated on existing and productive usage of the product within the company. In the case of open-core licensing the sales rep will still have to spend time explaining the differences between the open community version and the commercial product. And the rep will still have to justify the price of the commercial product in terms of business value. But this will be a discussion with a customer much closer to making a purchase decision than by cold calling a customer without any previous usage history of the product.

Going back to Brian’s conclusion; that open core licensing favors VCs more than customers. Agreed, but is that a bad thing? Customers benefit, in the form of new features, faster defect fixes, better documentation, etc., when the open source vendor is paid for its work.

The previous attempt by open source vendors to balance customer and vendor needs by charging for support only proved challenging at best. Most large open source vendors have moved away from selling support directly to selling product licenses which also include support. This approach aligns with the standard industry practice of paying for product licenses in line with business value received. Until, or if, the cloud becomes the leading monetization route for open source vendors, I’d argue that selling product licenses, as is the case with open core licensing, is the best approach for open source vendors and their customers. Oh, and their VCs.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Key members of the Sun Drizzle team have recently joined Rackspace. This is great news for Drizzle users, but also begins positioning Rackspace as an enterprise software provider for the next decade.

Drizzle, a fork of the MySQL database, is described as a database designed for the cloud. The project was founded by MySQL’s, and now Oracle’s, Brian Aker. Development of Drizzle occurred with Sun’s blessing after Sun had acquired MySQL and while Sun was still independent. Jay Pipes, another core Drizzle developer writes about his experience on the MySQL team before joining the Drizzle project:

“For almost 3 years, I had sent numerous emails to the MySQL internal email discussion lists asking the engineering and marketing departments… to recognize the importance and necessity of major refactoring of the MySQL kernel, and the need to modularize the kernel or risk having more modular databases overtake MySQL as the key web infrastructure database… My ideas were met with mostly kind responses, but nothing ever materialized as far as major refactoring efforts were concerned. “

Jay began working on Drizzle, and stayed on with Sun after the Oracle acquisition was approved. Jay and other members of the Drizzle team, including Monty Taylor, Eric Day, Stewart Smith and Lee Bieber eventually joined Rackspace on the Rackspace Cloud team.

Funding open source for profit:
Like any hosting company, Rackspace has a rich history of using open source within its hosting offerings and to run and manage their infrastructure. However, with Rackspace’s prior investments in the open source Cassandra project and new investments in Drizzle, Rackspace is shifting from a consumer of open source to a producer of open source. In doing so, Rackspace is also evolving its focus from a hosting provider to an enterprise software provider. The fact that the enterprise software will be made available through usage-based Cloud services and APIs is secondary. Rackspace is increasingly likely to compete with enterprise software vendors such as Oracle, IBM, VMware or Microsoft over the next decade. This competition is likely to open the door to Rackspace being acquired by an existing enterprise software vendor. I suspect this will happen as Rackspace’s revenue base shifts from hosting to enterprises seeking software platforms through Cloud offerings. The latter would offer higher profit potential than the former. This is important for a potential acquirer that needs to maintain high profit ratios for Wall Street. At 68 percent and 5 percent, Rackspace’s current gross profit and net income percentages would drag down the 80 percent and 20 percent respective figures that, for instance, Oracle and Microsoft achieve today.

Rackspace’s open source investments or enterprise software capabilities won’t surely stop with Drizzle. Databases are critical to Platform as a Service (PaaS), but so to are runtimes. This would suggest Rackspace’s open source involvement has room to grow in the area of Java or dynamic scripting language runtimes.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

By all accounts Red Hat is the undisputed leader in enterprise Linux, but Ubuntu is proving its up to the challenge.  Is it time to evaluate Ubuntu in your enterprise?

Like many of you, I haven’t given the Linux market too much thought beyond Red Hat Enterprise Linux (RHEL) being the leaders and Novell SUSE being a distant second.  Last May, while reading the Eclipse Survey 2009 results I came across two very interesting pieces of data about Linux adoption that made me reconsider this point of view.

I was very surprised to find that nearly 15 percent of Eclipse developers responding to the survey were using Ubuntu on their development machines.  I rationalized the lack of Fedora/RHEL or OpenSUSE/SUSE usage versus Ubuntu as a proof point of Ubuntu’s user experience investments.  But then I realized that Ubuntu performed equally well on deployment server market share amongst respondents.  Granted, Fedora/RHEL led Linux deployments, but only by a percentage point versus Ubuntu. And yet, from a revenue and unit shipment standpoint, IDC estimates Red Hat’s market share of Linux at over 60 percent.

Since May 2009, I’ve been keeping my eye out for data that supports Ubuntu’s growth in the enterprise.  Earlier this week I learned that Weta Digital, the digital effects studio behind movies such as AVATAR, District9, Jumper and Lord of the Rings, is using Ubuntu on a large scale.  Dustin Kirkland, an Ubuntu Server core developer for Canonical wrote about Paul Gunn’s 2010 talk:

“It was a great talk, about the type of data center needed to render special effects in today’s blockbuster movies. They have a 2 Petabyte disk array, 10gbps networking, and 35,000 cores (4,000+ HP blades) in their data center, and still it takes 48 hours to render some of their graphic sequences.

According to Paul, Ubuntu is at the core of all of this, running on all of the rendering nodes, and 90% of the desktops at Weta Digital. He notes that his farm (he calls it a “render wall”) is in fact an Ubuntu Server farm, and not RHEL as he has seen reported in the media.”

Weta’s data center is pretty amazing, especially in terms of its green data center practices.  The work running on Ubuntu at Weta Digital could easily be considered enterprise grade:

“…more than 10,000 jobs and an estimated 1.3 to 1.4 million tasks per day. Each frame of the 24 frame-per-second movie saw multiple iterations of back and forth between directors and artists and took multiple hours to render.”

Canonical’s strategy is to grow Ubuntu on client desktops, an area where Red Hat has left untapped.  However, as the Eclipse survey and Weta Digital’s usage underscores, Ubuntu shouldn’t be ignored as a server operating system.

While considering Ubuntu in your enterprise, particularly as a server operating system, enterprise application support on Ubuntu will be a key concern.  Red Hat definitely enjoys a much larger ISV ecosystem than Ubuntu does today, meaning that enterprise applications are more likely to be tested and certified on RHEL than Ubuntu.  However, with Ubuntu’s growth and the uncertainty around Novell SUSE, enterprise ISVs won’t be able to delay Ubuntu certification for long.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Palm’s future in the smartphone market remains uncertain, but its technology could prove valuable to Research in Motion, makers of the popular BlackBerry smartphones – especially as Research in Motion (RIM) continues its consumer push.

Reading David Coursey’s InfoWorld post “Palm is doomed; let the good-byes begin”, I couldn’t help but wonder “what next for Palm?” As Coursey and the Wall Street Journal mention, Palm has released strong products since re-launching on the WebOS platform and enjoys excellent carrier support. Yet, this hasn’t helped Palm’s share grow:

“In the nicest way possible, it (the Wall Street Journal) says Palm, with a mere 0.7 percent of the smartphone market, compared to 14.4 percent for Apple and 20 percent for RIM, simply can’t catch up.”

Being acquired by RIM is definitely one answer to the “what next for Palm” question. There is however the slight issue of Palm’s $1 billion market cap putting a serious dent in RIM’s $1.3 billion cash and near cash position. However, RIM doesn’t have any debt, so there’s room to finance the acquisition. RIM’s stock, while not priced where it’s used to being, remains on most investor’s tech stock short list, and could help fund the acquisition. For our purposes let’s assume RIM could close the deal.

The larger question is “Why would RIM want to acquire Palm?”

The 0.7 percent market share of Palm isn’t reason enough to acquire Palm. RIM could get its share of that 0.7 percent as Palm users look for future devices from Apple, RIM or Android phone manufacturers.

One reason to acquire Palm would be to leverage Palm’s open source experience. I’ve argued that RIM could benefit from using open source more effectively in its business. Palm would jumpstart this effort.

The more compelling reason to acquire Palm would be Palm’s WebOS platform. The BlackBerry platform, built on the aging BlackBerry OS, is in serious need of a refresh. This is less the case for enterprise BlackBerry users, many of whom couldn’t function without the email and messaging capabilities that the BlackBerry excels at. The user interface and rich interactivity of BlackBerry applications are secondary to the mail and messaging requirements. This will change over time as more businesses expose enterprise applications to mobile devices. For instance, the fact that a company’s CRM application is more usable on an iPhone or Android versus a BlackBerry may well entice an enterprise user to migrate off their BlackBerry. A revamped OS would resonate a lot more with younger consumers deciding between an iPhone or a BlackBerry Curve. The fact that one’s friends are on BlackBerry Messenger (BBM) is reason enough to leave behind 100,000 plus iPhone apps for the ability to communicate with tens or hundreds of ones (closest?) friends on BBM. BBM is an instant messaging platform available to BlackBerry device users. Giving these consumers users a richer, more fun, user experience would go a long way towards keeping these BlackBerry users happy in the face of iPhone toting friends.

The user experience would be vastly different between a WebOS-based BlackBerry and a BlackBerry OS-based BlackBerry. But this would be a point in time statement and one that retains and attracts both enterprise and consumer users. For a consumer the WebOS-based BlackBerry lineup, especially if new RIM designed devices are released in addition to the existing Palm devices, would be a much more compelling user experience than what’s available through the current BlackBerry OS-delivered user interface. For the enterprise user, the addition of a WebOS-based BlackBerry line, along with RIM’s commitment to bring the user experience to all BlackBerry’s would be a reason to remain a BlackBerry user until the new interface arrives on BlackBerry’s enterprise-targeted product line. Waiting for coveted features on a product or platform you’ve already invested time and money in is not uncommon in the IT market. For instance, as terrible as the BlackBerry browser is, many BlackBerry users are waiting at the edge of their seats for a new WebKit-based browser rather than jumping ship to an iPhone or Android device.  Early iPhone users lacking copy and paste capabilities are another example.

There’s also an issue of existing BlackBerry applications running on WebOS-based devices. Maybe a stripped down BlackBerry OS could run in a virtual machine on the mobile device? I’m sure RIM’s engineers could come up with some creative solutions.

As a BlackBerry user, I’d love to get the WebOS user experience in addition to the email and messaging capabilities of a BlackBerry.

Would you?

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”