Tuesday, January 29, 2008

Spring Source merges with Covalent



I wish we (JBoss inc.) had acquired Covalent back when they split into Covalent and Hyperic back in 2004. JBoss inc had just raised the money from the investors and Covalent was going out of business and was being sold for a mere $500k. I jumped on that deal. But I was late to that party. Both executives, Mark Brewer and Javier Soltero, knew what each independent branch was worth and were hell bent on pursuing those paths. I was mad with my investors and myself for missing the deal.

4 years later, Covalent, the HTTPD/Tomcat support branch, is merging with Spring Source, they are still not making much money but they are merging. From a distance the deal feels like a natural fit, since they are complementary stacks and Tomcat is the preferred runtime with the Spring development framework.

Spring was perceived by JBoss as a competitor, on the development side, in the early days. Indeed the EE programming models were competing with the POJO/IOC approaches. We took care of the quirkiness of the programming models in 04/05 and that resulted in EE5/EJB3 and then the programming models were all on par. However, on our core business at JBoss, which was monetizing the run-time and not the development time, we quickly grew to realize they were not competition. People running their stuff still paid us for our runtime, the EE stack. Spring is not a JBoss competitor were it counts. So our competition was the vendors that played in runtime: the BEA/IBM/Oracle's of the world.

On further analysis here are the salient points for me, you will see that I am mostly negative with a bit of positive.

1/ ASF runtime support is very far from the Spring core competency which is development help. These guys are going from consulting on development with POJO methodologies and a proprietary API (Spring) to supporting runtimes they don't own.
That is a pretty big jump. Like going from owning a driving school to supporting oil and tires replacement on cars through a garage. Spring's main problem is that people don't keep on paying driving school once they have graduated. They will however pay on an ongoing basis for oil change and tires change, i.e. from training to maintenance.

2/ Are we to understand that they are having a tough time monetizing their core competency? People download Spring, use it and say Thank You and that's it? I am not surprised! In some cases, millions of downloads ***DO NOT*** translate into hundreds of customers paying in the 6 figures. I think the ASF whole 15 years of history is a living testimony to that. Sometimes usage does not translate into money. That is one of the "mirages" of OSS.

3/ I know from experience that monetizing Tomcat is a difficult thing. Covalent is a fine business but it's revenues, after close to 10 years in business are still puny. I mean puny. Does Spring Source provide the basis to grow this? I mean in the sense that a SUN hopes to grow MySQL by throwing its systemic weight? JBoss has been developing Tomcat for the past 5 years. As a result you would think that JBoss/Red Hat is the legitimate purveyor of support. So far supporting Tomcat has not been a huge money maker. The BSD license is part of the problem but the bigger problem is the ASF brand. The Apache and ASF brands obfuscates every other brand, it obfuscates the individuals brands and more damningly their associations to companies.

4/ With support of runtime (Tomcat) what happens to the nice "vendor-agnostic" positioning and the existing relationships? Maybe they think they can do without it or that the vendors just don't care? possibly. Reality is they are now trying to compete on a narrow (albeit ubiquitous) slice of the runtime. They will have to compete with everyone else that has been doing support on Tomcat for so many years. It is a small and diluted market.

5/ One thing in their favor: they become an even juicier target for acquisition. I am sure they will be making a lot of noise and arm-waving about staying independent but they remind me of this Gary Larson cartoon with a cat looking out at a truck wreck that has a sign on it that says "Assorted Rodents and Flightless Birds". So close, the cat can even taste it! and Spring Source is starting to look like a fat flightless bird.

Overall, to me this smacks of a semi-desperate attempt to leap into the support business to drive product revenue. As I said many times, a services business I could get excited about, but as a product business it is difficult to monetize a development tool and it is difficult to monetize ASF products.

It seems to me they are hoping out of the frying pan into the fire.

I could be entirely wrong on everything of course.

MF

6 comments:

Pardee Terran said...

Hi Marc. I think you are one of the true geniuses of the open source world. So why is Red hat continuing to mess about with JBoss, continuing to underinvest in core enterprise functionality, and continuing to lose customers? I would really like to know from you, I trust you to be honest about such things. They are squandering a huge opportunity to be an open source leader, and are instead limping along like an also ran, just imitating competition. why? I am totally confused. Will you return to Red Hat?

The Rabbit said...

What does MF mean ?

Marcf said...

Rabbit,

MF means: Marc Fleury... what is that you are smoking?

Marcf said...

Pardee,

Thanks for the props but honestly I will not touch that one with a 10 foot pole.

Truth is the development team is intact and that is both a miracle and a opportunity. There are exciting products being developed and last I heard from the earnings call, JBoss hadn't lost a single customer but Oracle (which was a customer of JGroups).

As to returning to Red Hat... I don't know, when I was there, there was really no room for me in the executive team. With the new CEO, I don't know, no one has approached me :)

Bill said...

I'm gonna steal this from what Emmanuel said internally a few weeks ago...

For a company that is supposedly "dead" the past 2 years, JBoss released JBoss Developer Studio, all the Drools funky-monkey rule flow. Varies Seam versions. Hibernate Search. As well as the EAP 4.2 was is a productization of all the important EE 5 functionality. I think that's a pretty hefty amount of things.

Marcf said...

Bill,

now i know it is Burke and not Gates :) Good to hear from you my friend.

Look in all seriousness the only ones that claim JBoss is dead or at least "losing momentum" are Roy Russo and Rod Johnson. Roy? I don't know what is wrong with him, maybe he figures he will get traffic to LoopFuse that way? As for Small Assorted Rods he is now running around claiming he wants to be the "Dominant Player In Enterprise Java", he reminds me of someone, ah yes, me 4 years ago... Good for him, but take it from me: ignore the grandstanding.

BTW for all the still at RHT employees from the JBoss division. It is with pride that I have followed the announcements coming out. It seems there is a drum beat picking steam on the product side. I hear rumblings that some pretty interesting stuff may be on the way? keep it up guys!

PS: hummmm do get the EE5 out...
PS2: I am truly looking forward to seeing all of you in Orlando

I keep monitoring the work that comes