Business models for open source it management companies
As it is in many market segments, companies in IT management industry are looking for an open source business model that works. The prevailing approach seems to be the hybrid model. In this model, the companies offer both open source and proprietary versions of their products and proprietary versions typically provide additional high end features that is not available in the open source versions of their products.
This business model is only available to the companies that have the IP ownership of the open source code and not an option for more traditional open source projects where IP rights are distributed among many people and there is no single holder.
The model is a win win for both the vendors and the customers provided that the open source versions of the products are viable solutions that stand on their own and not just a ramp that requires every user to upgrade to the paid versions. Customers get access to open source software that solves their problems and vendors get well established benefits of having an open source community.
Not everyone agrees with the above statement. There is a discussion that started with cote naming the open source IT management companies “little 4″ as contrast to the proprietary “big 4″ (IBM,CA,HP,BMC), and heated up again with QClusters exit from the openQRM project, hence little 4 becoming “Little 3″ .
openQRM was not really in the same category with the other 3 in little 4 anyway, and its exit provided an opportunity to reevaluate the open source IT management companies. John Willis took a more minimalist approach, coming up with his own name, calling Hyperic and Zenoss the “mighty two“, others suggested Groundwork and OpenNMS should round up the new Little 4.
John states that Hyperic and Zenoss has a better chance to succeed in the enterprise as they are software companies with significant funding and solid infrastructure. Not surprisingly opposing view comes primarily from the OpenNMS camp (read the comments in John’s post). Tarus is not a fan of the hybrid model. He has criticized the hybrid model (and hence Hyperic and Zenoss), claiming it to be flawed, several times in his posts and comments. He advocates the OpenNMS group’s professional services only model as the superior (and even the only viable) model and states the success of the OpenNMS project as the evidence.
The hybrid model is not an option for OpenNMS group as (unlike hyperic/zenoss) they don’t own the IP rights for the code. OpenNMS apparently has an active developer community that contributes code (surprisingly rare as more and more open source code is developed by few individuals or companies) and OpenNMS group has one or two people who may be dedicated to development rather than services. OpenNMS folks seem to be content with their position, but I wonder why they don’t aspire for more.
Despite being compared to Hyperic and Zenoss, OpenNMS is different product. IMHO, it is the only true (open source) “network monitoring” product available. Hyperic/Zenoss/Nagios are primarily for server monitoring with some capabilities to monitor network devices, yet still not an option for serious network monitoring. The paste of opennms development is slow which is understandable given the fact that there is no dedicated development team. Let’s imagine for a second that OpenNMS group also used the hybrid model and external funding to staff developers to work on opennms. Could they have developed a discovery engine that auto discovers the network layers (layer 2, 3, routing, etc.)? Visual maps to represent these layers? What is OpenNMS offered this additional functionality only in the paid version but all the functionality available now was still available. What if on top of what is currently available, foss version of opennms had gotten a package that can be installed and run on windows easily 3-4 years ago? Would having the option to pay for this additional functionality be a bad thing for the community? Would removing entry barriers such as extensive documentation and easy installation packages not help increase opennms user base?
Open source companies need to continue to innovate in the business model level to find ways to fund the projects that work both for the companies and the communities. Hybrid model is what has emerged so far as one viable option. It is not perfect but it offers an alternative that is in many ways better than pure professional services model. There does not have to be “one true way”. The alternative may be getting squeezed out of the market. Tarus likes to compare OpenNMS with HP and the likes, suggesting that OpenNMS is a viable alternative to them. It is. but it is playing the catch up game instead of leading the field. The tougher competition for opennms (and other open source projects) is not coming from the large proprietary companies like HP and IBM, but from small, agile companies like solarwinds and adventnet with nicely packaged, easy to use products with often lower prices than the open source companies.
If open source companies cannot innovate both business model and technologically, they may get squeezed by these companies in the lower end of the market and the larger proprietary companies with massive sales forces in the higher end of the market.
Subscribe to feed
Add New Comment
Viewing 6 Comments
Thanks. Your comment is awaiting approval by a moderator.
Do you already have an account? Log in and claim this comment.
Do you already have an account? Log in and claim this comment.
There is a huge difference between what OpenNMS attempts and the hybrid open source models. To choose an easy target, Microsoft publishes a lot of code under OSI approved licenses, but no one thinks of Microsoft as an open source company. Enterprises are tired of paying high per node prices for proprietary software licenses, be they fully closed or partially open, and in the long run the free and open solution will win. This doesn't mean that a hybrid company can't make money or provide a decent solution, but without a totally free and open application there is little difference between a partially-GPL'd application and a proprietary application with a decent API. Of course, you can always "fork" a GPL'd application, but when the vast majority of the code is controlled within the corporate sponsor, the chance of that happening is slim (although Red Hat's fork of the Hyperic agent is one example of what can happen).
At OpenNMS we do aspire to greater things, but since we run the commercial side as a profitable company it is taking time to build the business versus those companies with millions in VC. Our view is much more long term, whereas VC-backed firms are looking at a 5 year window. I think Groundwork provides an interesting case study. They've been around the longest of the hybrid management firms and they have raised about as much money as Hyperic and Zenoss combined. Yet they haven't really done as well as one might expect, and with the recent departure of Ranga (the CEO) they must be trying to figure out how to change that. It will be interesting to see how Hyperic and Zenoss avoid the same fate.
We don't see much competition from AdventNet, but we lost a number of opportunities to SolarWinds in the past. We have recently won several, so perhaps that is changing.
Our goal is for OpenNMS to become as powerful a framework as OpenView or Tivoli, with more agility and, of course, considerably less cost. Thus we focus mainly on scalability and flexibility, often at the expense of ease of use. As a services company, we want the most powerful tool, and our knowledge of how to best use that tool drives the services business. Most of our commercial customers are large enterprises and carriers who are looking at a spending hundreds of thousands of dollars (and more) on commercial frameworks. Our smaller users don't tend to buy support (although it does happen).
My point is that Solarwinds will probably be more successful than the hybrid companies because, as you state, they have "nicely packaged, easy to use products with often lower prices than the open source companies". What's Up Gold is considerably less expensive that the cheapest OpenNMS support contract, so if it works for you, great. Solarwinds is a well run software company that doesn't muddy the water with and open vs. closed version. It's easier for people to understand and to buy.
However, the larger you get the more custom your management needs become. Prepackaged solutions don't fit, and thus you need something more flexible. That's the niche we at OpenNMS are aiming for.
Anyway, sorry to ramble. Great post.
Do you already have an account? Log in and claim this comment.
Thanks for sharing your thoughts, ramble away, it's all welcome :)
Looks like you're going to be in Geneva in couple of weeks. How long will you be here? Love to get together if it works out.
PS: your blog's password reset system does not seem to work, just FYI
Do you already have an account? Log in and claim this comment.
"Despite being compared to Hyperic and Zenoss, OpenNMS is different product. IMHO, it is the only true open source “network monitoring” product available. "
After carefully rereading I see you are advocating the use case for OpenNMS as network monitoring. I guess I have a heightened sensitivity when I see the mixture of company business models and opne source software. I would contend that Zenoss Core offers a very capable network management and monitoring capability as well. Thanks for the clarification.
Do you already have an account? Log in and claim this comment.
indeed the "only true" was not referring to the open source but refers to network monitoring aspect of OpenNMS which is my background.
Zenoss pretty much seems to do everything :) I'll have to learn more about the network monitoring capabilities. thnx!
I'll see whether I can slightly modify the post to shift the emphasize
Do you already have an account? Log in and claim this comment.
Thanks for taking time to comment. I'm a bit puzzled. May be my post reads different than what I was trying to express. I certainly don't think that corporate sponsorship dilutes open source and also agree that it is just FUD. In fact, it is partly what I was trying to convey (clearly failed big time:-)
I see dual license model (GPL and proprietary) embraced by Zenoss and Hyperic as a business model innovation that enhances the options for customers, not as being not pure or as a negative. I also think that the innovation in the business model needs to continue to find even better models.
I used OpenNMS as an example of why we may need better models since the dual model is not an option for them (even if they wanted to embrace that model which they don't). IMHO, pure services model is limited and more resources are needed for companies like OpenNMS to be able to invest more into their products. Currently we seem to be stuck between the dual license and pure services models. I simply think alternatives are needed. What are they?
Any feedback you may be able to provide where I botched the post (that implied zenoss/openms/hyperic are not open source companies etc.) would be much appreciated as it is polar opposite of what I was trying to say.
Do you already have an account? Log in and claim this comment.
Let's take the companies out of the equation. I think all three products have merit and they all three have users who use them for a variety of reasons. Corporate sponsorship diluting the "purity" of an open source is just FUD. The code is there and obtainable from third parties like SF.net if the users don't like the direction of the project they can fork it.
Also I think you should delve a little deeper into this great distributed IP ownership hypothesis for some projects looking at the commits for development beyond the core teams and they are pretty limited. Read some copyright's for ALL three projects and you will see copyright notices from privately and publicly held companies.
I do think what might be interesting is to look at the forum activity and source code repositories and see what's really going in these projects. Just because companies charge for a relatively small amount of additional code and a high level of service doesn't necessarily mean that the open source offering isn't a very robust solution.
Add New Comment
Trackbacks