Blogs

RedMonk

Skip to content

RHX: Package Management Fully Realized?

In an interview with Red Herring yesterday, I claimed to be “outspoken” on the subject of a Linux “repository.” I’ve already gotten one question on that rather cryptic quote, so let me explain.

For several years now, I’ve been making the argument that package/application management is a significantly differentiating feature for the operating system. More specifically, it differentiates Linux – which has both officially sanctioned package management applications and community repositories behind them – from Solaris and Windows. It’s a significant advantage, in my view.

Linux users can install and maintain literally thousands of applications without even leaving the command line, while Solaris (unless they’ve gone the Blastwave route) and Windows users are more or less on their own for application procurement, patching and management.

Despite this seemingly impressive feature, you’ll rarely if ever hear this touted by Linux advocates; perhaps they simply take it for granted given that many if not most Linux admins have never known a distro without it. As a result, my recommendations that Linux distributions press their advantage in this area by leveraging existing infrastructures to connect to both commercial and community oriented repositories more or less fell on deaf ears.

In the summer of 2005, as an example, I recommended that Red Hat integrate their Network offering with yum to begin the convergence of commercial and non-commercial libraries. While the Red Hat folks were polite, they also seemed rather unimpressed with the idea.

They were not alone. Generally speaking, when I’ve spoken to commerical Linux providers on the subject, I’ve been told that a.) customers don’t have interest in a wider network of applications, or b.) community offerings threaten our commercial network. I tackled the latter point in the piece linked to in the above paragraph, saying:

Q: But isn’t there the possibility that these application management facilities will compete with commercial versions, such as the Red Hat network?
A: Great question. The answer, in my opinion, is no. Here’s why:
  1. I accept as a given that community maintained libraries will always be more comprehensive and long tail-ish than commercial equivalents
  2. I further accept as a given that big commercial entities would prefer to pull their version of, say, MySQL from a commercial network versus a library maintained by individuals without commercial needs in mind
  3. I accept as likely that commercial ISVs would prefer to work with commercial networks

In short, I think there’s an excellent opportunity for commercial and community networks to function seamlessly side-by-side, each serving different needs and having different strengths…Either network by itself is incomplete, in my view. Commercial networks will always lag non-commercial ones in terms of application breadth and depth, and non-commercial networks are unlikely to be perceived as a viable basis for higher end enterprise needs.

That commercial and non-commercial networks can operate side-by-side in a complimentary rather than zero sum fashion I think is better understood and appreciated. But what about the first objection – customer demand? Are customers interested in a network of applications to be supported and offered from the operating system vendor?

If the Red Hat Exchange (rhx) is anything to judge by, the answer is yes. I have yet to speak directly with the folks from Red Hat on the subject, but from the outside looking in this appears to be an excellent first step towards realizing the potential advantage Linux has always enjoyed in the package/application management space.

Microsoft is living proof that a platform ecosystem and the inertia therein is the best defense against would-be rivals, but Red Hat seems poised to exploit precisely that strategy at the expense of Windows, not to mention Solaris. This is obviously important for Red Hat; like Eclipse, it has moderate immunity to would-be challengers simply by virtue of its ISV commitments, and this could make that already strong story stronger. What is perhaps less obvious is how significant a development this could be for the greater open source ecosystem: simplified entry-points to Red Hat’s installed base? As Matt says, not a bad opportunity.

Anyhow, will be checking in with the Red Hat folks ASAP to see whether or not my expectations are born out or merely wishful thinking.

Categories: Open Source.

  • http://www.michaeldolan.com Mike Dolan

    RHX is a monstrous +1 from what I’ve been briefed on. Let me know if you need a contact to talk to.

  • http://www.michaeldolan.com Mike Dolan

    I should have also cautioned that I think it’s early stage still, and will not happen overnight, but is a great strategic vision from a company that has in the past focused more on simply repackaging and shipping CDs.

  • http://tomayko.com Ryan Tomayko

    I’m not sure what developers you’re talking about but I’ve always said that package management was one of the most important features Linux and BSD have over other options. I don’t consider an OS modern unless it comes with some decent form of package management.

    Also, speaking of Yum, Seth Vidal just posted a note that he is leaving Duke. Anyone working in this space should be advised to get in touch with him.

  • http://www.mulesource.com Dave Rosenberg

    I think RHX is a good idea (and one they should have done long ago) but question how feasible it is for RH to support these apps they have minimal experience with.

    For that matter, how do they expect to sell other people’s apps when they haven’t proven they can sell JBoss yet.

  • stephen hood

    I couldn’t agree more Stephen. Here was my experience and it was a good one for a relative newbie..

    While I have a technical background my experience was sooo pleasant because I DIDN’T need to have it..

    I was testing out one of the vmware appliances based on Ubuntu. It was missing some stuff..so I went to add appplications..checked them off and a minute or so later they were installed and running..no urls to track and go to..no re-boot etc..etc. Just worked.

    Just amazing ease of use for your average consumer.
    Dead simpler than anything.

    This is in fact THE KILLER FEATURE of Linux…more so than anything else. You are dead on. Absolutely dead on about this.

    My dad could do it …if Dell would sell him one :)

  • http://redmonk.com/sogrady sogrady

    Mike: agreed on the vision, obviously. i’m definitely bullish on the opportunity here.

    Ryan: well, you’re smarter than the average bear ;) but in all seriousness, it’s not that the feature is unappreciated by devs and users, it’s that it’s not marketed as such. witness some of the folks that publicly maintain that Linux is “harder to update,” where nothing could be further from the truth.

    thx for the note on Seth, btw.

    Dave: excellent question. i think i’ll pose that to Red Hat when i speak with them ;)

    Stephen: amen. i’m reminded by your post of Federico’s Mom who was able to update her kernel via apt/Synaptic. it’s absolutely a killer feature.

  • Pingback: tecosystems » Beyond Support/Service: Making More Money from Open Source

  • Pingback: tecosystems » Launchpad and the Future of Package Management

  • Pingback: tecosystems » Launchpad and the Future of Package Management