Many midsized organizations have done without this type of sophisticated network-management tool because of the high costs. OpenNMS puts sophisticated network monitoring into the toolkit of IT departments that would not otherwise have the budget for it.
The project started in 1999 and was released under the GPL. It uses Apache Tomcat for the Web server and Java servlet engine, and PostgreSQL for the database. OpenNMS is written in Java and is tested to run on most major Linux distributions (Red Hat, SuSE, Debian, Mandriva) as well as Solaris and Mac OSX. Users have also reported being able to run the product on FreeBSD, OpenBSD, Gentoo and Ubuntu. Binary packages are available from the Web site for most operating systems.
If there are outages, OpenNMS will try to isolate the problem to help you respond. For example, it will try to determine whether the problem is a single application or an entire section of the network. Once the problem has been isolated, it can record the event, run any preconfigured script, create a trouble ticket, send e-mails or pages, and create a report. Administrators can optionally create custom reports.
The OpenNMS project's stated goal is to be a truly distributed, scalable platform for all aspects of the FCAPS (fault-management, configuration, accounting, performance and security) network-management model. Because OpenNMS is released under the GPL, this platform is available to any application. Administrators can use OpenNMS to plug into outside services such as Nessus, an external security scanner, to add functionality.
- Free open source network-management system
- Polling intervals dynamically change, based on outage event
- Allows polling of outside sources such as vulnerability information from Nessus (external security scanner), tailed log files, and /proc-based monitors
- Simple Network Management Protocol (SNMP) hooks
- Isolates problems for more meaningful reporting
OpenNMS users are typically experienced network administrators comfortable with the command line and at least somewhat familiar with network monitoring concepts. The installation complexity is average, but anyone experienced in package management (apt, yum, yast, up2date, etc.) will have no trouble. Once OpenNMS is installed, using it requires climbing a learning curve. It is a very powerful tool and, as such, can be daunting to configure at first. But the time spent learning OpenNMS will result in significant rewards because it can handle a wide variety of complex management tasks.
Customers derive a lot of value from OpenNMS compared with the proprietary options. Rackspace Managed Hosting uses OpenNMS for monitoring about 200 internal servers. The $200,000 to $250,000 price tag of a commercial product (Rackspace was looking at HP OpenView) gave the company sticker shock, but it found that OpenNMS met its needs with a little elbow grease and a few thousand dollars per year in commercial support. [Note to editor: I did a case study on this. It's at http://windows-linux.com/articles/OpenNMSCaseStudy.html if you want to link to it.] OpenNMS is being used in a wide range of settings. Administrators report monitoring anything from groups of 25 servers to 80,000 devices.
Commercial support is provided by the OpenNMS Group Inc., which employs the major contributors to the project. Prices are listed online for a range of service offerings, including custom installation and configuration, training, and ongoing support contracts. The OpenNMS project mailing lists are also helpful in getting started.
Maria Winslow is Open Source Practice Leader with Virtuas, where she assists clients in understanding the technical and budgetary impact that open source software will have on their computing environments. She is the author of The Practical Manager's Guide to Open Source and a site expert on SearchOpenSource.com.
This was first published in May 2006