Oracle has popped out a white paper that may well turn some heads, because it contains robust criticism of open source software.
Titled “The Department of Defense (DoD) and Open Source Software” and available here as a PDF to those with Oracle accounts or here in Dropbox, the document's premise is that folks in the USA's Department of Defense (DoD) could think it is possible to save money if they “... avoid buying commercial software products simply by starting with open source software and developing their own applications.”
The paper goes on to explain why that's a bad idea and why paying Oracle for commercial software is a much more sensible thing to do.
The foundation of the arguments is that developing applications based on open source software has hidden cost, mostly in labour.
It also warns that open source software may not scale. “Commercial software companies have developed highly refined methodologies to perform these tasks,” the document suggests. “Don’t underestimate the difficulties associated with testing open source software and incorporating required changes into the main development stream, especially when it comes to testing for robustness and reliability under load”.
Oracle also argues that the DoD need not concern itself with integrating software into the world's myriad hardware ecosystems, again because it is hard to do and Big Red's already doing the heavy lifting. Another argument suggests it will be harder to certify open-source-derived projects as secure than it will be to buy a certified app from … guess who … Oracle. Or other software vendors.
Another skein of the argument asks whether it is appropriate for a government-funded organisation to work on open source software when private outfits like Oracle do it already, do it efficiently and do it in ways that meets the needs of many industries.
The paper's not blind to Oracle's own reliance on open source, noting the company's investments but insisting open source only makes sense when someone like Oracle takes the time to integrate it into wider hardware and software stacks. Big Red also asserts it offers better support than a DoD user could hope to provide itself or acquire elsewhere.
Overall, the paper is decently balanced, raising real risks associated with software development even if its tone, narrative and a section recounting problems with an open-source-based health records program open source all suggest strongly that development based on open source software is a silly idea for the DoD to contemplate.
The paper also contains more than a few passages The Reg imagines might lead open source advocates to take issue.
For example, a concluding section titled “The Proper Use of Open Source” offers the following advice:
“Oracle helps ensure that open source software fits well within the surrounding infrastructure and provides a route to enterprise grade production. However, for the intensive, mission-critical capabilities required by most DoD projects, Oracle recommends its flagship commercial software products.”
It should come as no surprise that Oracle reaches such a conclusion or that it makes the effort to put such a view into the public domain and the minds of DoD workers. The paper is also not terrible advice.
Of course it is not hard to think up rebuttals to the paper, not least the NHS' recent decision to ditch Oracle in favour of open source. ®