For open-source projects, 2010 was an interesting year. While they weren’t the most popular open-source projects around even when Sun controlled them, Oracle made a lot of developers angry by dumping OpenSolaris and allegedly neglecting OpenOffice.org.

On the other side of that coin, the company managed to bring new contribution commitments for the OpenJDK from Apple and IBM, so the company cannot be called weak on open source.

The Apache Foundation also had a busy year, with Hadoop taking most of the spotlight. The open-source map/reduce implementation was the most popular project at Apache in 2010, whether you’re measuring by code contributions, sub-projects or discussion lists.

Hadoop’s diverse ecosystem grew in conjunction with a growing need for big data solutions. Subprojects of Hadoop addressed the need to make sense of all this big data. The Mahout project continued to expand its machine-learning development tools for use with Hadoop, and the NoSQL Cassandra project introduced the ability to transfer data into Hadoop directly from Cassandra.

Speaking of NoSQL, the hype cycle was at its absolute peak for these new-fangled databases in 2010. While NoSQL was the buzzword for developers looking to relieve database bottlenecks, by the end of the year, the bloom was already coming off of the rose. Dozens of NoSQL choices, coupled with the extreme immaturity of nearly all NoSQL projects, combined to make these faster data stores somewhat less appealing than they were at the beginning of the year.

For fans of the AMQP standard, SpringSource snatched up RabbitMQ, giving that open-source AMQP implementation corporate backing. Google purchased Instantiations, and while the Instantiations tool chain was not and is not open source, Google released WindowBuilder Pro, WindowTester Pro and CodePro Analytix for free.

Finally, Linux saw another year of growth, both on developer desktops and on the server. At this point, Microsoft is second fiddle in the data center, and the Linux community, led primarily by Red Hat, has been adding compelling features and reworking older systems. The new Completely Fair Scheduler, for example, smoothes out the wrinkles that made it tough to predict schedule execution in real-time systems.