Analysts familiar with Sun Microsystems Inc.'s Project Indiana say that as early as this week the company could...
reveal plans to revamp the OpenSolaris operating system by incorporating key pieces of Linux software.
In March 2007, Sun officially announced Project Indiana, whose goal is to create an OpenSolaris binary distribution. The long-term objective of the project is to increase the technology's user base and cultivate mindshare. Similar to Red Hat's Fedora Core and Novell's openSUSE projects, OpenSolaris is Sun's open source operating system and includes experimental features that might eventually make their way into its commercial Solaris operating system.
Industry watchers said Project Indiana is indicative of Sun's desire to increase awareness of Solaris among Linux-centric IT managers and developers. But these same pundits also said Sun must walk a delicate line between luring Linux users to Solaris and supporting its existing Solaris user base.
Giving OpenSolaris a Linux 'look and feel'
"We've been getting hints that Sun is getting ready to put out Project Indiana and that it is going to make Solaris appear more like Linux from the perspective of developers, systems managers and operators," said Tony Iams, a senior analyst with Rye Brook, N.Y.-based Ideas International Inc. "This is because Sun has been concerned about lowering the barriers for developers and managers to using Solaris as an alternative."
One such hint came in March, when Sun hired Ian Murdock, the founder of Debian Linux and former head of the Linux Foundation, a nonprofit consortium that supports the growth of Linux. At the time, Murdock wrote on his blog: "You can probably guess from my background and earlier writings that I'll be advocating that Solaris needs to close the usability gap with Linux to be competitive; that while as I believe Solaris needs to change in some ways, I also believe deeply in the importance of backward compatibility. … Even with Solaris front and center, I'm pretty strongly of the opinion that Linux needs to play a clearer role in the platform strategy."
Solaris and Linux, kernels and interfaces
In a July 6 interview with Reuters, Al Gillen, an analyst at Framingham, Mass.-based IDC, said that Project Indiana will keep the Solaris kernel intact.
If Sun wants to bridge the gap between Linux and Solaris functionality, its choices are to do so at the kernel layer or at the user layer, which includes interfaces, application programming interfaces (APIs) and administration tools, said Ideas' Iams.
"From a functionality standpoint, the [kernel] barrier here is not insurmountable," Iams said, but it's significant nonetheless. "At the kernel level, the issue is device drivers and being able to run Linux device drivers in Solaris," he said.
Making kernel-level changes would also create licensing issues. Linux is released under the GNU General Public License (GPL), but Solaris is based on the CDDL, or the Common Development and Distribution License. Currently, the Free Software Foundation, which governs Linux's GPL license, considers the CDDL incompatible with the GPL. "Unless there are changes made to the license, some other workaround will be needed," Iams said.
Therefore, analysts believe Project Indiana will focus on the user level by introducing APIs and interfaces. This approach would be in line with interface technologies like GNOME -- an open source graphical user interface that sits on top of a computer operating system -- that Sun has incorporated into Solaris in the past, said Illuminata Inc. senior analyst Gordon Haff.
An "above the hood" kernel-free approach should work for Sun, Iams said, because the GNU tools in Linux also run on Solaris. The move to a Linux-like interface and tool set will also allow Sun to focus on Solaris' strengths, he said, like DTrace, a framework that enables users to troubleshoot system problems in real time, and Containers, a partitioning method that enables completely isolated virtual servers within a single operating system instance.
But Iams noted that the real issue with Project Indiana isn't software compatibility; it's support. "Sure, this will have a comparable [interface] and feature set to Linux, but which ISVs [independent software vendors] are going to go ahead and test and certify applications for it?" he wondered. "These ISVs are already doing the same thing with Red Hat and Novell's SUSE, and now Sun will have to convince them to do their testing again on its stack. Sun will have to prove to them that they have solved their compatibility issues."
Walking the Linux/Solaris tightrope
As it turns to Project Indiana, Sun faces another hurdle with its own customers. While Sun works to incorporate new Linux tools and features at the API level, it risks alienating existing Solaris users who like the OS the way it is. Haff and Iams agreed that Solaris devotees could push back in response to a perceived watered-down version of their OS.
"Conceptually, it's a good idea, but there are users who are going to say, 'Solaris is doing just fine' and then wonder why Sun is talking about ruining Solaris by contaminating it with Linux junk," Haff said.
And Linux components slated for Solaris won't necessarily contribute to a better operating system, said Haff. Indeed, Solaris already boasts a dedicated following of users. In the case of Project Indiana, Haff maintained that Sun would introduce Linux tools, APIs or interfaces to make Solaris more enticing to Linux users.
"It's an important distinction," Haff said. "This [project] is for people familiar with Linux."
Sun executives were not available for comment regarding this week's scheduled unveiling of Project Indiana.
Email Jack Loftus with your comments and suggestions.