While we position OpenNMS to compete with products from giants such as HP, IBM and CA, I had doubts that we were on their radar. But yesterday I saw the following in an article on Network World:

While not quoted, it appears that the CTO of CA dropped the name of our project, so one can only assume he is aware of what we are trying to do.

Cool.

I do agree with him that network management, especially at scale, is a “freaking hard problem”. Note that we are both using the term “network management” as an umbrella term for managing anything that is connected to the network, which ranges from traditional networking gear such as routers and switches all the way up the stack to applications and mobile devices – Internet of Things style.

It is the main reason we designed OpenNMS as a platform versus an application. It is a requirement that it be flexible enough to meet the unique needs of our users, and that can only be done by writing OpenNMS to be extensible while also automating as much of the work as possible. It is a complex problem.

It made my day that the CTO of a nearly US$14 billion company mentioned our effort. It means we are on the right track. We definitely don’t have the resources of CA but our team is talented and they understand the network management space.

It was also cool to see 451 Research mentioned in the article. I really like those folks, so much so that we just contracted with them. Perhaps we can get more people talking about us.

Susan Perschke has written up a fairly long review of four open source management tools, including OpenNMS. I thought it was pretty thorough for a Network World article, especially since they tend to follow the industry in general and not network management in particular. She reviewed OpenNMS as well as Zenoss Core, NetXMS and Nagios Core.

Note: You have to “register” to see the whole article, but all that seems necessary is to input an e-mail address. There is no confirmation e-mail, as far as I can tell, so you could enter in anything that looks like an e-mail address and access it.

I was a little nervous when I read that they tested OpenNMS on Windows. To be perfectly honest, our Windows implementation was done in part to settle a bet (Java should be write once, run anywhere, right?). It *does* work and work fine, but it is much harder to install and manage on the Windows platform, something that we are working to correct. So Susan gets some props for managing to get it running and not complaining at all about the install process on Windows. That fact alone made me look at her article more seriously.

Everyone is interested in the “winner” and the tagline of the article is “Zenoss Core edges Nagios Core, NetXMS and OpenNMS in four-product roundup”. She writes:

Zenoss is our top pick due primarily to its intuitive and professional-grade admin interface.

I really can’t fault that. GUIs are important, which is why we are working with the Vaadin team on our next generation user interface. But the underlying engine is important as well, since it is easier to add a pretty interface than it is to architect your application for performance. Apparently Zenoss is a resource hog:

We found out the hard way that Zenoss is not designed for lightweight infrastructure – in fact we were not able to get a clean-running machine going until we threw 6GB of memory at it.

My main complaint about the article is that she didn’t stress enough that Zenoss and Nagios both follow the fauxpen source model: the “Core” products are free but the whole is not. If you want certain features, be prepared to get out your checkbook and give up your open source freedoms. She does mention that with respect to Nagios:

You would need to purchase the ‘Professional’ or ‘Business’ version if you want to use features such as SNMP traps or the mobile application, plus the backend database option is only available in the ‘Business’ version.

Why is this distinction important? Well, as Brian Prentice at Gartner explains, when considering a solution like this with an commercial enterprise component, you have to treat the whole purchase decision in the same way you would any commercial software acquisition. You lose the open source value.

And we be all about adding value. In fact, we are actively targeting the features that people like the most about products like Zenoss and Solarwinds, first and foremost the GUI. Next, we’ll go after the second most popular feature. Little by little we hope to erode people’s dependency on commercial software by offering a truly amazing, totally free alternative.

I didn’t talk to much about NetXMS because, quite frankly, I had never heard of it. It seems like they have been around for awhile and it also seems like they are truly open source. I found it amusing that the two FOSS options in the article were both tested on Windows.

Anyway, the article is worth checking out if you are in to that sort of thing. I’ll be interested to see how OpenNMS ranks in a year when the GUI is finished.

For those of you who read German, OpenNMS gets a shout-out in the latest Didactum newsletter.

In the February edition of LinuxPro Magazine there is a nice article on OpenNMS.

Called “Big Time“, the author Kurt Seifried discusses using OpenNMS to monitor very large networks.

It is a pretty positive article. The only complaint I can raise is that it was against OpenNMS 1.9, which is our development release. In the section “What’s Wrong with OpenNMS” he quite fairly raises some issues with IPv6 support, but that code is definitely alpha and will be much better by the time 1.9 becomes 1.10.

Other than that, it is obvious that the author spent some time with the software and really “kicked the tires”. I am hoping that the article causes a few more people to check out our work.