First, some history. Red Hat Linux was one of the earliest created Linux distributions, invented by Marc Ewing in 1994. Ewing actually began working on the distribution, then known as Red Hat Software Linux, in December of 1992, soon after he graduated from Carnegie Mellon University. Over the next nine years, Red Hat Linux would continue to capitalize on these and other advantages, devoting much of its marketing and sales efforts to getting Linux into the workplace. Specifically, the enterprise, which are organizations with 500 or more computer users. By targeting this market, Red Hat was essentially aiming for the low-hanging fruit; there aren't a lot of enterprise-level customers out there, but it only takes a few to really build your revenue stream.
But this approach, while commercially successful, gave Red Hat a problem. While Red Hat Linux was becoming increasingly popular in the corporate world, the community-oriented developer and user bases that had helped to bring Red Hat Linux where it was technologically was feeling increasingly disenfranchised. What good was it to develop cutting-edge software, the community complained, when Red Hat, fearing any software instability for its well-heeled enterprise customers, would only include it after a laborious quality control process? This issue was serious enough to prompt Red Hat to launch several community outreach programs -- including two US bus tours -- because if the community began to vote with their feet and walk away from Red Hat, highly valuable development resources could be lost and Red Hat's community reputation would plummet. Some of this outreach helped, but in the end it took the birth of a new distribution and the death of Red Hat Linux to help save Red Hat. One of the community outreach projects started by Red Hat as it worked to keep community interest high was the Fedora Linux project, a collection of newer and more experimental software that, when installed on Red Hat Linux, would give users a chance to enhance their computers without introducing risky software to the main Red Hat Linux distribution. Fedora Linux was not a stand-alone version of Linux, but actually a repository of software. In September 2003, the split was made: Red Hat Linux would cease to be as a product. Instead, users who wanted to have commercial-grade support (such as those coveted enterprise customers) would now use and pay for Red Hat Enterprise Linux (RHEL). Users who wished to continue using Red Hat Linux for free, and without support, would use Fedora Core, a formal merging of Red Hat Linux and the software from the Fedora Linux project that would provide newer, cutting-edge software. Which brings us to 2010, with this week's release of Fedora 13 and last month's long-awaited release of RHEL 6, albeit in beta form. The history lesson, which can stand on its own, is also a long and convoluted way of explaining that I know exactly why RHEL 6 seems so far behind Fedora 13, particularly superficially. It's apparent as soon as you install the operating system: RHEL 6 Beta is running GNOME 2.28, Firefox 3.5.3, OpenOffice.org 3.1.1 -- each a couple of releases behind their Fedora 13 equivalents. New programs like Empathy are not even there -- RHEL 6 comes with Pidgin 2.6.5 instead. I get it. I really do. RHEL is stable, comfortable, no surprises here. And yet, it's still a little disappointing. The big reason for my let-down was found in the installation screen found in Figure 1.
Figure 1: RHEL 6 Beta offers few installation choices
As you can see, there were just five install options made available in this beta release: Basic, Web, Desktop, and Software Development Workstation. There's a fifth option, not visible here, for a Minimal Install. For a platform that purports to be the enterprise server platform, just five options seems sparse. My biggest question coming out of this beta review is: will Red Hat offer more packaged install choices? I certainly hope so. For instance, it would be great if, after all of the touting of the KVM virtualization software being supported in RHEL 6 over Xen, there wasn't a "Virtual Server" option that includes all of the software and kernel mods needed to run KVM. And, beyond that, what about pre-built stacks for application servers like Alfresco, Joomla, or JasperServer? It seems like Red Hat is missing a real opportunity to push itself into the cloud space. Ready-made stacks from the Red Hat release team would be a great value-add for anyone looking to turn a key and go. Of course, there is a case to be made for a "plain" multi-purpose platform that can be all things to all people. Many such stacks can be created during installation just by going into the customized install procedure and knowing which packages to pick. So it's not as if RHEL 6 is lacking for packages. Plus, let's face it, helping users put those stacks together is a nice reason to provide enterprise support. The install options that are there, of course, are top of the line. I looked at two: the Desktop and Web Server. The Desktop install is lean and functional, with much of the same look and feel of it's Fedora counterpart--though a few versions behind for many of the installed packages, as I mentioned earlier (see Figure 2). The Web Server is a little more advanced in the software department: Apache 2.2.14, Perl 5.10.1, PHP 5.3.1, and MySQL 5.1.42--none of these too far behind the latest and greatest versions of these packages.
Figure 2: RHEL 6 Beta's desktop interface
Of course, RHEL 6 is completely stodgy in its features. Besides the KVM toolset, the ext4 and XFS filesystems are both available. There's also enhanced multi-processor support. I didn't test this one myself, unfortunately, but Red hat reports one RHEL 6 server can run across 64,000 cores. RHEL 6 does exactly what it's advertised to do. It's a solid server platform with a wide variety of stable software available for installation. But if you need a stack beyond LAMP, it's do-it-yourself or call your local RHCE.