A common mistake in the information security world is to use network security tools (like firewalls, IPS and extrusion prevention systems) as band-aids to fix root problems in software.
Bruce Schneier has been lobbying that software security should be a regulatory/compliance play where a software vendor will need to comply with government software security standards and if their bugs cause damage, they will be found accountable and be liable to payment of a fine.
I am skeptical of government regulation of software security ever happening (although if it did it would be great for business) however - a recent OASIS working meeting seems to suggest that there may be another way to tackle the problem from the top down - i.e. by mandating government use of vendor neutral security standards and open source software in order to grow public accountability through transparency.
OASIS - the Organization for the Advancement of Structured Information Standards, is a non-profit consortium focussed on developing e-business standards. They recently had a session in the EU where a draft resolution was proposed linking open source, software security,
and public accountability.
“The house proposes that within 10 years all European governments will have adopted systems based on open security standards for all external electronic communication. By then, governments will use open source software exclusively to implement those open standards in order to be accountable to citizens, business and other governments.”
The naive reader might conclude that OASIS is pushing an anti-Microsoft agenda. However, the converse is true. Open security standards are about being vendor-neutral and open to review in a similar way that open source encourages peer review.
There is absolutely no question at all, that security by obscurity is worthless and that the strongest security measures (whether they are in encryption or for software security) are those where the algorithm is held up to public scrutiny. The accumulated effect of eyeballs only results in improved quality of the software and results concomitantly in improved security.
When OASIS says Open Source software to implement open security standards, they are saying let's apply the peer review principle to both the algorithm (the standard) and the implementation (the software). This is a good thing - since in Europe (at the very least), transparency and free market may be stronger than government regulation for corporate software governance. (I am skeptical of this working in Middle Eastern or Eastern European countries).
The question still remains, are the Europeans ganging up on Microsoft or not ?
I did a search on sourceforge.net today and proved a thesis I've had for a long time - namely that more and more Open Source projects are being developed for Microsoft Windows operating systems. Today, December 2006, out of 12,000 projects, 5100 were on Windows and 6900 were for Linux; almost half of all Open Source projects on sourceforge are on the Windows platform.
It is no wonder - if you get a new PC or notebook - it will be running Windows XP, its dead simple to install MySQL on a Windows machine or Ruby on Rails or PHP, or Perl or Postgres - you name it. And - its free. And - it's open source.
Open Source applications are used for any operating system, not just Linux, thus making the OASIS proposal universal in its application.
This leaves us with a more fundamental question: Is anything a government do really transparent, and does this really improve software security?
Let's say that the French Health Ministry uses Squirrel Mail for its Web mail access, Squirrel mail is a well known FOSS product that works on either Windows or Linux platforms using Apache. There might be an XSS vulnerability in the product or the Ldap application they use for single sign on might be vulnerable, or the Web server itself might have weak passwords - any number of problems might exist.
In other words, it's a good first step, a necassary but not sufficient condition.