Be fair to OSS – lay off the sauce

Open source software has come on by leaps and bounds over the last ten years or so.

There are more and more examples of credible OSS (Open Source Software) solutions available today, including well-known brands such as Firefox and LINUX. However, OSS seems to me to face a major enemy – itself.

The problem is that, deep down, the OSS community can sometimes see itself as a bastion of freedom against the evil commercial ISVs – certainly some of the more vocal supporters of OSS are guilty of making more and more outlandish claims about their own projects, really pouring on the sauce. Some of this is to be expected, of course, but there comes a point when fanatical over-statement of the facts can so offend credibility that prospective OSS users are turned away.

I was reminded of this problem by a recent article by Dennis Byron, writing for ebizQ. While I found the article both interesting and informative, the section on enterprise service buses (ESBs) illustrated my point. In the article, Dennis states

ESBs are especially interesting because they may turn out to be the first category of software code that was OSS from the get go.” And later, “The OSS movement in turn blocked those early non-OSS ESB market leaders before they could gain a lot of traction.

The ‘early non-OSS ESBs’ phrase refers to companies such as Progress, Cape Clear and Fiorano, according to the article. This is the type of hyperbole that in my view causes the OSS movement to weaken its own credibility. The implication is that OSS has ruled the ESB market from the start, apart from some minor incursions by early ESB vendors such as those listed.

Now, I will agree that OSS ESBs are maturing nicely – the Mule OSS ESB, for example, already lists a substantial number of production implementations on its website. I think that as time goes on, these OSS offerings will become more and more competitive to the commercial offerings. However, to state that ESBs were ‘OSS from the get go’ is just ridiculous. For the first three years of the ESB market’s life, the only serious usage was with commercial offerings. While it may be true that some of the early players failed to gain much traction, there are now commercial ESBs available from all the major vendors such as IBM, Oracle, SUN and BEA, and some of these are performing strongly in the market.

I believe the OSS cause is best served by a more realistic assessment, without all the sauce. We are now at a stage where OSS ESBs are a viable choice alongside commercial offerings. There are references for successful usage of OSS ESBs, and the future appears to be bright – for example, OSS solutions with a supportive community could result in pools of less common adapters being available to satisfy a wide range of needs, something less likely to happen in a commercial world. But to state that commercial ESBs never made it, and OSS ESBs ruled the roost from the start, is in my opinion a wild exaggeration.

To be fair to Dennis, later in the article, while discussing message-oriented middleware and the impact of OSS offerings here, he states

But it will be well into the period 2011-2020 before commodity MOMs/ESBs significantly displace IBM MQSeries and like products.

This comes across as more realistic, and hence makes a far more credible claim. MOM (Message-Oriented Middleware) products like MQSeries are much more extensively deployed that ESBs, with many thousands of users, and therefore even if there were attractive OSS alternatives (and this is definitely arguable today) it would take some years for any significant displacement to happen.


Posted in enterprise service bus, esb, Imported, Industry trends, Open Source, SOA.


  1. Steve, thanks for mentioning my article. I never thought I’d be accused of having drunk the OSS Kool-aid because I am the leading proponent of “open choice.” But if that’s the way I came across, let me clarify.
    In my 1500-word research article, the sentence “ESBs are especially interesting because they may turn out to be the first category of software code that was OSS from the get go” was immediately followed by the words “Progress, Cape Clear and Fiorano would argue with that characterization but my point is that OSS communities and sponsoring suppliers jumped in quickly with OSS ESBs.” You left off the explanation (and implied the references to Progress and the others came later).
    I am a market researcher and what I look at is sales/market dynamics trends. I believe the ESB is going to trend like the web server did, bundled into stacks, with very little chance for a separate market to develop. The large-supplier ESBs you mention are almost all sold already as part of the supplier’s stack (just as the Apache web server is also in IBM’s and Oracle’s stacks). BEA and IBM may have separate ESB prices on their price lists but that is not how they sell their ESBs. The ESB on Oracle’s price list is bundled with the Integration Server (where it has been for years, only it was called Interconnect). And of course Sun’s is OSS already.

  2. Dennis,
    Thanks for the clarification – you are right, I did not include the list of early vendors in your quote. However, my point still stands. The implication of your article is that the bulk of production implementations of ESBs are with OSS solutions. This is incorrect, as I stated.
    I agree with everything else though! So we are not far apart here…

Comments are closed.