Recently at the Cloudera Analyst Conference, open source was front and center. In fact, it was suggested that they stressed open source so much that their unique value seemed to be lost. What stood out the most was not how they embraced open source so much as how Cloudera understood the fundamental nature of it. Given how many companies claim to be open source or contributing to it, one would think their particular stance on open source wasn’t unique. Sadly, it is.
When is Open Source Not Open?
Open source is as much a philosophy as a business model. The assumption is that by opening up software source code to community scrutiny and modification, the code will evolve more quickly, be better quality, and adoption will happen faster as well. Unfortunately, openness of intellectual property (IP) is not in the DNA of most vendors. It’s hard to blame software vendors. For most of them, their IP is their lifeblood. Control of that IP is essential to their survival.
To adopt an open source strategy is to put aside sole reliance on IP. It doesn’t mean that there is no proprietary intellectual property – Cloudera is pursuing a hybrid model of some open source and some proprietary software. Instead, it is a commitment to openness and the community approach to core development. This only works when open governance is embraced as well as open source.
Open governance means that decisions about the software are made by communities not companies. Typically control of the software rests in the hands of a foundation or other not-for-profit. The Apache Software Foundation (ASF) and Mozilla Foundation are examples of this type of organizations that ensure that the community is in charge. To be clear, there is nothing wrong with proprietary software. There is something wrong with proprietary software masquerading as open software. When a single company controls the decision and licensing of the software, it can hardly be called open.
Cloudera Gets What Open Source Means
Cloudera is a company that understands this. As a company that leverages open source projects such as Apache Hadoop and Apache Spark (and a host of others in the big data space) their core software is a distribution of software that is managed and licensed by an independent community. This means they can influence but not control the direction of this software. They add value through packaging, service, and their own software that they don’t call open source such as Cloudera Navigator. Sometimes they do create something they want the entire community to have, such as Impala and Kudu. In that case, they turn it over to the Apache Software Foundation, giving up the IP and control. What they don’t do is call the proprietary software open source while trying to maintain control of the license. In other words, they don’t try to create open source without open governance.
As long as a software vendor maintains control of the license they can force their will on the community, rendering the community powerless. In no way can software in that state be construed as truly open. Proprietary software in open source clothes is inherently dishonest. It’s refreshing to see companies such as Cloudera willing to be open about their openness. They get the open part of open source.