Unisphere conducted the research project, surveying 761 SQL Server database professionals, asking questions about environment, budget, security issues, and drivers for security testing. In essence, the study finds that budget for database security is on the rise, that companies are aware that sensitive information is scattered around their organizations without much in the way of controls, and that some 20 percent think they will suffer a data breach in the coming 12 months.
Having tracked this market closely for a decade now, I had a couple of observations on the study:
* It's interesting that database administrators were only involved in security 75 percent of the time (Figure 4). As it stands, most database security tasks involve the database administrator in some way. The setup of auditing, patch management, vulnerability assessment and many other actions must be performed by the DBA. The fact that DBA involvement was _down_ to 75 percent is indicative of the impact data privacy and regulatory efforts, which require separation of duties, are having on the industry.
* The percentage of respondents with monitoring in place was 41 percent (Figure 26). While this is SQL Server specifically, and by my informal count SQL Server databases are more commonly monitored that DB2, MySQL, Postgres, Sybase or just about any database not called Oracle, this percentage seems high to me. I would have expected less than 30 percent using DAM. When you look at which activities are being monitored (Fig 24), the top three -- or 5 of the top 6 for that matter -- are actions normally gathered by native auditing. If I was to venture a guess I would say that these 'monitoring' functions are not Database Activity Monitoring (DAM), rather auditing provided by native audit features such as tracing or system event logging.
* Those reporting no breaches in the last 12 months were 74 percent (Fig 7). Small firms and even small enterprises are not usually targeted the way large enterprises have been. Even if they were susceptible to 0-day attacks or "drive-by malware," odds are they would not have adequate detection and forensics in place to know that they were breached. The number of actual breaches or incidents of information leakage is likely far higher than being reported during the last 12 months.
* Local state data laws was the biggest driver for security (Fig 29), followed by Sarbanes-Oxley. I am not surprised by this, but I imagine that many people are. If I was taking bets I would have said SOX was number one, with state privacy laws at number two. Activity monitoring and assessment has proven incredibly effective at implementing financial controls in complex accounting systems, and can be done in such a way that the auditor is not totally dependent upon IT or accounting to gather information. Similarly, activity monitoring is effective at identifying data misuse, a key for various state data privacy laws. There is a lot of hoopla around PCI and HIPPA, but the customers I speak with use database encryption for compliance, and are far less likely to invest in monitoring and assessment.
* The report implies that most firms are using generalized "configuration management" tools determine database patch levels, or Microsoft's native revision checker. Specialized database assessment tools that look at patching as well as vulnerabilities, configuration defects and user authorization maps. Maybe it's an issue of budget, but most configuration management tools miss the bigger security picture. This is a serious omissions from my perspective as DB assessment is the single greatest tool for database security as it covers all of the basic database security issues. The report is lengthy, but there is a lot of good information here. And 761 respondents, with a nice distribution of company sizes, is a great cross section of the database security market. Thanks to Application Security for funding this project and publishing the results. Definitely worth your time to review.
Adrian Lane is an analyst/CTO with Securosis LLC, an independent security consulting practice. Special to Dark Reading..