I received an interesting call last week from Gary Read, the CEO of Nimsoft. He informed me that his company was buying the intellectual property assets of Cittio, a company with a rather storied history with respect to OpenNMS. With this acquisition, parts of the Cittio Watchtower product involving Layer 2 discovery, topology and root cause will be integrated with Nimsoft solutions and the Watchtower product will cease to be produced or supported.
Now, I’m a firm believer in karma. Not the formal concept as defined by many Indian religions (and I mean no disrespect by using the term casually) but the idea that if you do good things, good things will happen to you, and if you do bad things, well, bad things will happen. In my mind the failure of Cittio is karma in action.
If you search for “cittio” on my blog, you’ll see that since 2005 we have been rather curious as to how OpenNMS and Watchtower were related. The early releases of Watchtower seemed, feature for feature, to be remarkably similar to OpenNMS. I had a couple of conversations with Jamie Lerner, the CEO and founder of Cittio, where he assured me that any use of OpenNMS code was within the license requirements of the GPL. Information surfaced in 2008 that seemed to indicate that claim was inaccurate, so we hired the law firm of Moglen Ravicher to help us get to the bottom of this.
When our lawyers contacted Mr. Lerner, he reiterated his claim that Watchtower was not in violation of the GPL, and even if it was, any OpenNMS code being used by his application would be based on OpenNMS version 1.0.2. Since that code was copyright the Oculan corporation and not The OpenNMS Group, we had no right to enforce the copyright.
While we had received information that OpenNMS code post-version 1.0.2 was being used in Watchtower, we had no proof, and so without a long and expensive discovery process we were effectively stymied from getting to the bottom of this matter.
Thus it was with some surprise that we watched Cittio release a fork of OpenNMS in November of 2008 called Rooftop. If, as Mr. Lerner claimed, Cittio was always in compliance with the GPL, the fork should have been unnecessary, and I found it amusing that the press release went out of its way to specify it was a fork of OpenNMS 1.0.2.
As I write this, it appears that Rooftop has been downloaded around 60 times in the last 6 months or so. Compared to the 5000 downloads a month of OpenNMS from our repositories I would say that the fork hasn’t hurt our project at all. And for those who wish to test the claim that no OpenNMS code post version 1.0.2 was being used, I invite you to download the 1.0.2 source as well as the source of, say, 1.2.0, and draw your own conclusions. A search for the word “interfaceresolve” might provide a good starting point.
With the end of Cittio as a management company, I guess a lot of these points have become moot. Whatever the actual history of how Mr. Lerner came to use the OpenNMS code in Watchtower, the future should be much clearer. Mr. Read has stressed that he wants to make sure the OpenNMS project is satisfied that none of the OpenNMS code will be a part Nimsoft’s offerings, and he has offered, at his expense, to fly one of us out to inspect the code.
While I can’t (currently) see the Watchtower code, in looking through Rooftop there is no mention of linkd (our Layer-2 topology discovery process), alarmd (our event correlation engine) or our “drools-based” event correlator, so I assume that Nimsoft has done enough due diligence to be sure of what they are buying. While their emphasis on transparency is obviously driven by a wish to avoid any taint of license violations in Nimsoft products, it is refreshing that they took this seriously enough to contact me well before the press release and to make the blanket offer of showing us the code.
One might think this is strange behavior considering the press Nimsoft was getting in open source circles after the announcement that Hyperic was being acquired by SpringSource. Part of it was stupidity. A person at Nimsoft’s PR company apparently did a search on articles referencing the sale and sent a generic anti-open source article to Matt Asay. While Matt and I have had our differences, he was right to skewer the guy. I don’t mind the marketing effort, heck I spread my own FUD about commercial software with Cittio being a prime example (as in “what do I do now that this license I bought is to code no longer supported”), but just as with a lot of the near spam e-mails I get from potential “partners,” the PR guy should have done his homework before sending it to Matt. I do object to laziness.
Gary Read got his own bunch of replies on his blog post about the acquisition. Some people saw it as an attack on open source, but if you actually read his post it was an attack on the hybrid (what I call “fauxpen source“) model. When he called me, we had a nice long conversation and were pretty much in agreement that while pure open source has advantages, and commercial software has advantages, the hybrid model is not able to capitalize on either, except perhaps for a short-lived bounce in marketing exposure.
Moving forward, Cittio customers will obviously have the choice of migrating to Nimsoft’s solutions, but I also want to encourage them to check out OpenNMS. They might find it strangely familiar.