Looking in on CentOS 7

With so many shared components — the Linux kernel, the GNU toolchain, X server, productivity applications, etc. — it can be challenging for a Linux distribution to stand out among the crowd of competitors. One way to do so is to have a substantial commercial interest develop the distribution, then support it with upgrades and security patches for a long time. Such a distribution may be attractive to casual and enterprise users alike, because the distributor is likely to avoid major updates of the packages that those users have become accustomed to.

That kind of stability is part of what makes CentOS particularly attractive. CentOS is a free Linux distribution that provides ten years of support, along with binary compatibility to the commercial Red Hat Enterprise Linux (RHEL) product. The latest version, CentOS 7, hit its general availability release milestone in early July, several months after the project joined forces with Red Hat. Given the many technical and organizational changes to the CentOS project in the past few months, it warrants a look at how these changes have manifested, as well as at the feasibility of CentOS 7 as a desktop solution for casual users.

To the non-technical user, a CentOS 7 desktop may simply look like a polished version of CentOS 6, with the default GNOME 3 Classic Mode GUI looking shinier and more modern. To a seasoned system administrator or developer, the changes are more radical. With XFS as the default filesystem, systemd as system and service manager, base images for Docker to run CentOS in a container, and (as with RHEL) no 32-bit release available, CentOS 7 reflects the major changes in the Linux world since RHEL 6 was released in 2010.

Previously, one would know what to expect from the CentOS project until the next major release of RHEL: updates, bug fixes and security patches would be regularly imported from upstream. For CentOS 6, there was some additional development for areas not catered to by RHEL—notably, Xen support. We can continue to expect this from CentOS going forward, but there's more happening with the distribution these days: special interest groups (SIGs) and official "variants" of CentOS that can be either moderate tweaks or radically different version of the "core", binary-compatible-with-RHEL-7 distribution.

SIGs

The SIGs are formal groups, approved by the CentOS Board, that put their attention and development resources to particular projects and problems related to CentOS. Some of these groups focus on particular development issues with the core CentOS distribution, similar to Fedora teams. For example, the Artwork SIG focuses on the visual design of CentOS, and the Promo SIG is a marketing and awareness-raising team.

While no variants have yet been released, the proposed and approved variant SIGs indicate that CentOS will cater to a variety of different use cases not otherwise addressed by the CentOS core. For example, the proposed Cloud Instance SIG would focus on developing "generic" CentOS images for cloud computing, as well as tweaked images that cater to the demands of particular platforms, such as Amazon Web Services and Google Compute Engine. Its group members include employees from Google, HP, and Cloudbase. The proposal is ambitious; if it is approved, there would be several images built to accommodate demand for OpenStack and Docker, among other popular "cloud" software options.

The recently approved Atomic Host SIG will be developing a tweaked CentOS specifically for running Docker containers, and will also develop resources for others to make their own CentOS Atomic Host variants that cater to their individual needs. There is some overlap with the proposed Cloud Instance SIG, as the Atomic Host SIG would also make images for particular cloud platforms and technologies like OpenStack. SIGs will collaborate to handle shared concerns.

Other proposed SIGs awaiting approval would offer other tweaks. For instance, the proposed Simplified Linux Server (SLS) SIG would make CentOS images catered to a variety of server use cases, with "a platform for pulling all of these disparate [server] applications together [like SpamAssassin, ClamAV, Dovecot or Cyrus IMAP, and Postfix], and then wrapping up the solution in a web/REST-based interface" for management by system administrators. This addresses the hurdles faced by system administrators to configure CentOS for the needs of particular server use cases. The proposal describes the process for configuring CentOS's core release for a mail server, noting that several packages would need to be installed, considerable configuration of those packages would have to be done, and then the system administrator would have to dig into the OpenSSL documentation "to figure out all the security certificate jargon." The SLS initiative would simplify the process of building and maintaining server images like a mail server.

The proposed VoIP SIG would provide a means to address the particular needs of VoIP software projects. The list of other possible SIGs reflects the considerable diversity of ISOs and images that the CentOS project could be offering in coming years. These include porting CentOS to other architectures, high performance computing, and supporting alternative desktop environments and window managers for desktop end users.

Desktop

Speaking of the CentOS desktop, testing the CentOS 7 core release reveals much to like, for technical and non-technical users alike. Just mentioning that CentOS has ten years of support may pique the interest of users of the abandoned Windows XP, which reached end of life earlier this year. While CentOS will have increasingly out-of-date software as time passes, that may not be a concern for many home users as long as they can "get things done" and it "just works" — things that CentOS certainly provides.

Many of the needs of basic, non-technical home users are met: web browsing and social media (Firefox), email (Evolution), and productivity (LibreOffice) are available out-of-the-box. Tools like GIMP for graphic design work, along with many other browsers, email clients, productivity tools, and more, are all available from the official repositories. With the Extra Packages for Enterprise Linux (EPEL) repository enabled, personal accounting (Gnucash) and the ability to run Windows applications or games (Wine) are also available. Toss in Steam for hundreds of Linux games, and one has a desktop usable by regular users after minimal configuration. And one supported until 2024, when CentOS 7 reaches end-of-life.

One might counter that a Chromebook is a more attractive solution for non-technical users; Chrome OS automatically updates without the need for user intervention, and only expects that users know how to use a web browser. However, those users might chafe at the limitations of a Chromebook. They won't be able to run iTunes, Microsoft Office, or their favorite Windows-only games — software disliked by the open-source community, but required by many other users. This is unlikely to change in the future; Wine is currently unavailable for Chrome OS, and Wine developers noted in a mailing list discussion earlier this year that Chrome OS's architecture and reliance on sandboxing may make porting Wine to Chrome OS impossible. The availability of Wine for CentOS may be enough to draw in former Windows XP users.

As the "Community ENTerprise Operating System", though, CentOS remains a project that targets the business world. With the additional resources of an S&P 500 company backing it, and the SIGs and variant proposals reflecting increased attention on new and popular enterprise use cases (including devops and containers), CentOS remains as a big player in the enterprise software world.