- 12 iPhones Apps That Will Make You a Networking Star
- 10 Careers Robots Are Taking From You
- Big Data Gold Isn't Always Where You Would Expect It
- 6 Tips to Build Your Social Media Strategy
Network World - Asked whether you should trust your network to open source, you may interpret the word "trust" in two ways. Trust could refer to robustness: Can open source applications be counted on to work when it matters? Or trust could refer to security: With the source code open to everyone, is open source safe to use?
For years, companies have based their businesses on such open source software as the Linux operating system, the Apache Web server and databases such as MySQL. It makes sense that the open source management tools that grew out of that environment are starting to see widespread adoption. Tools such as OpenLDAP and OpenVPN are used to administer servers and network access. Webmin simplifies server management via a Web browser, and such network-management tools as Nagios and OpenNMS are quickly gaining ground on established commercial applications.
This is surprising, considering that most open source projects have not been around nearly as long as their commercial counterparts. How is it that they can approach commercial functionality faster and for less cost? Part of the reason is the open source nature of the code. Features that get created first are precisely the ones most needed, or desired, by the creators. Functionality is not driven by marketing.
What about robustness? The sad fact is that all software has bugs. With open source, however, the identification and correction of those bugs happens at a much faster pace. In one example with OpenNMS, a network vendor was sending an incorrectly formatted SNMP trap. Confronted with a network trace (using another open source tool, Ethereal) and the RFC demonstrating the mistake, the vendor replied, "Well it works with [HP] OpenView." Instead of having to argue with the vendor, the code was simply modified to handle the formatting error. The whole process, from problem identification to workaround, took less than four hours.