Since its inception, LightCube Solutions has run on a custom-built Linux machine. Being a former LFS developer, I hail from the Linux world of ‘Do It Yourself’, and so I prefer to use self-configured servers, tuned and set exactly the way I like. This is no Fedora or Ubuntu where a host of unnecessary packages are forced on you and custom configuration files mask the generic and standard configuration files that come with the original software. This is ultimate flexibility.

But that flexibility does come at a price. Maintaining an LFS system can become a chore. Installing a new package always means compiling from source. Staying on top of security updates is entirely left to you. The system is only as good as your personal understanding of its internals. A balance somewhere in between would be ideal:

  1. A lightweight system that is known to be stable and secure.
  2. The possibility of complete configuration is given to the end user.
  3. The focus of the system is tight, and therefore higher quality (in terms of stability, functionality and reliability) can be achieved.
  4. All the while the system benefits from security updates and testing derived from a community of users and developers.

And so, having realized that I needed to move beyond my personal build scripts and start packaging the system (at the least, for my own sanity) it was decided to create a distribution based on our own needs for Linux-based web services. Voila! LightCube OS is born. The basic outline of the distro’s goals are this:

  • Provide a lightweight, fast, stable and secure LAMP application server.
  • As close as possible, adhere to the GNU principles of free software in the packaging and distribution of the system.
  • As nearly as possible, provide a ‘vanilla’ system. In other words, don’t create obscure custom configuration schemes. Allow as much manual configuration by the end users as possible.
  • Focus on packaging software that is reasonably used with production LAMP servers. (E.g., there’s no reason to build an X desktop environment for a server housed remotely and accessed mainly through ssh. Make the system geared towards advanced command line users. As much as a good GUI is nice, there’s no reason for a remote server to run one locally.)
  • Make the base system streamlined, optimized and small. While it is realistic to package a few variations of software (E.g., nano vs. vim, Exim vs. Postfix), the core system should focus on one basic set of core packages.

These are the main ideas behind LightCube OS. The build scripts and the core package specs are already under development. And the distro’s project site/infrastructure has been put in place: http://www.lightcubeos.org. Volunteers are welcome to join in the development.

In the meantime, what are your thoughts concerning the above? What advantages/disadvantages do you see to such a distribution? Do you have any comments or suggestions that will help improve its appeal or usability? I welcome your comments…

4 thoughts on “LightCube OS

  1. Hi, and thanks for sharing your idea.
    I find your approach quite interesting, we all tend to work in the same direction of getting the thing right, and I would be the first one to welcome the approach – If you / we can add something, I would also like to contribute.

    First and most important for me, where does it differ from Solutions like (Ubuntu JEUS and the good old Contribs.org / SME Server??). I have been working with the SME Server since the E-Smith days, and this project could sure need a helping hand, not in regards to stability – but in regards to innovation.

    My main thing these days are Virtualization – not as a host but as a virtual appliance, how well it plays with others. And I love the fact that the Contribs and Ubuntu does flow well with 256MB of ram. Not saying you should get down to less, but where are the innovation objective; what would be the advantage of “ditching the others” and proceed with yours.

    Reply
    • Hi Dion, thanks for the comment.

      For the moment, there are two main things I believe would set this distro apart from the rest:

      1. The tight focus on providing a LAMP application server. Given that the focus is so tight, the potential for perfecting the main system and its core functionality is greater, less time and effort is spent on non-essential software.

      2. Allowing for and even encouraging complete configuration by the end user.

      Let me explain the second one a bit further. In the development world we hear people praising ‘Convention over Configuration’. As long as your needs fit the convention, this is excellent since it can speed up development. The same principle however, is often applied to Linux distributions, as well. Again, this can be good, if it is your aim to appeal to desktop users or the sort of server admins that rely on GUIs to configure and control their machines.

      In my experience, far more flexibility and control can be achieved when using Linux from the command line and manually editing configuration files. Most advanced users I know will prefer to have their servers run without an installed GUI. As stated in my blog, ultimate flexibility and control comes from running systems like Linux From Scratch, since you compile and configure everything by hand. By keeping modifications to software configuration small or even non-existent, we can provide a packaged and well maintained system that offers the same sort of flexibility and control as a system built by hand from scratch.

      These are my main thoughts. However, I can appreciate that such a distro could benefit from other innovative ideas. Hence I welcome any suggestions from advanced Linux users as to what would make this distro even more appealing.

      Reply
  2. Pingback: Roy Schestowitz (schestowitz) 's status on Thursday, 17-Sep-09 22:21:28 UTC - Identi.ca

  3. Pingback: Links 17/09/2009: Citrix Enters Linux Foundation, Linux Mint 7 “Gloria” XCFE Praised | Boycott Novell

Leave a reply

required

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>