Open-Source Database Security

A recent article on Dark Reading underscores a growing concern in IT: how to secure open-source databases.

Jun 21, 2010 | 08:53 PM |

By Adrian Lane
Dark Reading

A recent article on Dark Reading underscores a growing concern in IT: how to secure open-source databases.According to the article, "Open-Source Databases Pose Unique Security Challenges," by Ericka Chickowski, Ingres and MySQL are being used for large-scale commercial applications. They support Web commerce sites and are growing in use as fast as their commercial counterparts. They are used as a replacement for Oracle, DB2, and SQL Server when cost or flexibility is an issue. The issue Ericka raises is not that these platforms are less secure by nature, but that they lack the supporting security knowledge and tools available on other platforms.

Until recently, the major database vendors did not offer a full suite of database security options. Now they do, with activity monitoring, assessment, and transparent encryption to go along with access controls and authorization schemes. Rummage around the open-source Web sites and you will discover deployment guides and basic security tips -- but these center around access controls and secure communication (i.e., SSL). Security research and tools designed to help secure these platforms is generally absent. Third-party vendors have not ported their monitoring, masking, auditing, and assessment tools because customer demand has not been high enough to justify the costs.

To get a better idea of why this is important, let's look at vulnerability assessment. There are no formalized assessment products, and beyond a smattering of policies for MySQL, no research teams perform policy development for open-source databases.

1. Research for threats, remediation and workarounds: DBAs for the most part are not security professionals, they don't spend their days researching security threats, and they don't know what to look for. What does the attack look like? What is a suitable resolution? Are patches available? If not, is there a workaround/resolve the issue?

2. Automation for the collection of user permissions, manual comparison of settings in relation to established policies, and keeping up to date with patches are all time-consuming, mistake-prone process.

3. Writing the queries to find the information is time-consuming. One of the aspects I hated most about creating assessment policies was writing the queries as a simple check, which could take a day to write and optimize, then test it across all revisions of the database. It might be cost-effective for very large firms to do, but small security firms have trouble justifying the cost of this development.

Apache Derby, Open Ingres, MySQL, and Postgres (or PostgreSQL for some of you) have millions of databases in use. Look at the vendor best practices guides and you will see a preponderance of the discussion around access control and use of SSL. Advanced security usually falls onto the shoulders of the application developers to thwart SQL injection and memory exploits. There are few generic tools over and above file/OS encryption, and that only addresses media protection. Only MySQL has support for monitoring, assessment, third-party auditing, blocking, advanced configuration guides and the like. The other databases? Not so much.

I don't bring this up as some form of advertisement for tools and tool vendors because, frankly, the security vendors don't make solutions for these platforms. My point is to illustrate that security goes largely unaddressed, offering something of a greenfield for attackers. And while we don't hear about these attacks, they do happen.

Is it time for the open-source community to build security policies? Probably. Will it happen? I doubt it. Heck, I can't even find good tuning and performance guide for Postrges: It took years of speaking to other DBAs, along with trial and error, before I found what worked. And believe me, performance questions are far more common than security questions. There is not enough customer demand to motivate third parties or vendors to invest in development. As use of these platforms increases, especially as low-cost, cloud-based services, attacks will increase in frequency, much as they have with Oracle and SQL Server. Unless the open-source community steps up, the platforms will likely suffer a black eye before things change.

Adrian Lane is an analyst/CTO with Securosis LLC, an independent security consulting practice. Special to Dark Reading.

To upload an avatar photo, first complete your Disqus profile. | View the list of supported HTML tags you can use to style comments. | Please read our commenting policy.

Subscribe to RSS