What to Look for in Open Source Systems Management Products
You could call them the "Little 4," a play on the "Big 4" systems management companies: BMC, CA Technologies, HP, and IBM.
The Little 4 may be smaller in terms of their bottom lines, but in terms of systems management capabilities, there's nothing little about the open source offerings from Zenoss, Hyperic, GroundWork Open Source, and The OpenNMS Group.
This is a biosphere within the open source ecosystem that has seen a lot of flux in the past few years. In 2006, Zenoss was one of six vendors (the others being Ayamon, Emu Software, Qlusters, Symbiot, and Webmin) that launched the Open Management Consortium (OMC), "an effort to promote the adoption, development, and integration of systems and network management software based on open source and open standards technologies."
Now, the OMC is long gone, faded into the mists of oblivion, while other firms and technologies have stepped up to fill the systems management sector, which encompasses monitoring, systems, and network management tools -- and generates about $4 billion in sales. It is projected that this will eventually expand to around $12-$14 billion in annual sales, with just four companies taking up to half of those sales numbers.
There is not much linking these companies together, other than a very healthy sense of coopetition. For the most part, the atmosphere in systems management land is friendly, though there are some skirmishes every once in a while.
Take, for instance, the recent launch of MonitoringForge, a centralized gateway for the monitoring community. The site is sponsored by GroundWork Open Source (GWOS), and has been criticized by Tarus Balog, the CEO of OpenNMS, as being simply a marketing front for GWOS. GWOS has taken the high road on this one, continuing to assert that there needs to be some sort of a system management community in play. (This community stance is a bit on the ironic side, given that GWOS did not join the OMC in 2006, instead forming an Open Source Council that same year.)
Looking at all of these systems management tools, it can be difficult to parse any real advantage between these companies. Zenoss, perhaps the best known of this group, but their product offerings seem little different from those of, say, Hyperic's. To add to the confusion, products in this space can even include each other: the GWOS and Hyperic offerings incorporate Nagios for event tracking, and both Zenoss' and GWOS' products make use of the Multi Router Traffic Grapher for resource graphing.
Also in play is the actual definition of what these firms mean by "open source." All of them, with the exception of OpenNMS, use variants of what's known as the "open core" model.
Open core is often confused with dual licensing, the classic example being MySQL, which has two licenses: the free version under the GNU General Public License (GPL) and the commercial licensed version. With the commercial license, customers can sell a MySQL-based product without making the product open source.
Open core doesn't describe the licensing of software, but rather the features of the software itself.
In the case of the open core systems management companies, they typically offer a decent open source core product that can do an adequate job of monitoring and management, alongside one or more commercial offerings that offer more features.
Critics of open core -- and there are a lot of them -- usually come after this business model as exploitive of the open source technology from which it came. They also decry the core products as actually limiting innovation.
Not unpredictably, OpenNMS' Balog is one of the most vocal critics of open core. Of all of the Little 4 firms, only the OpenNMS product is released under the GPL, operating on a service and support revenue model.
Unless license evangelism is a part of your decision-making process, choosing one of these monitoring and management products will be a matter of determining how the Little 4's products handle the load of management and monitoring on your network. All of these products feature low network and resource footprints, which is key to a successful systems management platform--you don't want to the management tools themselves to introduce latency into your server farms. But there are nuances between these products that you should watch.
Another important criterion to use is scalability of the system. This is not just scaling up as your datacenter grows, but also scaling out: What framework is in place to let you manage and monitor new apps as they get added to your network? One tip in looking at the Little 4: lurk around their communities a bit and see which have the most activity. A healthy community means more flexibility and adaptability should you need some work done later.
Another thing to check, if you plan to modify any of these products to accommodate your application space: Find out up front how collaborative the open core offerings are. There have been instances of open core vendors rejecting contributions from customers that the customer needs because to do so would bring the core product too close to the commercial add-on capabilities. To date, this has not been reported within the systems management sector, but it remains a possibility. If you plan to simply consume a systems management product without any such collaboration, then this should not be a concern.
Open source systems management products offer a robust, less expensive solution to administration and monitoring your IT systems. The good news is, any of them can be a good fit for your organization--but that's the bad news as well, so be prepared to do the homework to research them thoroughly.
Brian Proffitt is a Linux and Open Source expert who writes for a number of publications. Formerly the Community Manager for Linux.com and the Linux Foundation, he is the author of 19 Linux and Open Source works, including the new Introducing Fedora: Desktop Linux.