A majority of IT security pros believe that continuous monitoring of the database network is the best approach to prevent large-scale breaches like the ones that occurred at retailers Target, Michaels and Neiman Marcus, a study showed.
Nearly two thirds of the 595 U.S. experts polled by the Ponemon Institute cited monitoring as the best form of database protection, a position other security experts challenged.
"Continuous monitoring, looking for unusual or anomalous type of behavior, becomes very important," Larry Ponemon, chairman of the Ponemon Institute, said. "The more you monitor, the more things you can see and the more things you can stop."
In general, tools described as database activity monitors capture and record activity related to the use of Structured Query Language (SQL), which is used by database administrators and others to perform a variety of tasks. Alerts are generated when activity violates policies.
While the tools have been around for years, many were too expensive and delivered too many false positives, which made them difficult to use, Ponemon said. Today, the tools are more mature and "they're not going to bankrupt the company."
For example, the tools are better at locating active databases that a company many not have known was in use, Ponemon said. They are also better at catching rogue software that might have been installed by a hacker to store stolen data before it is sent to a remote server.
Some technologies also monitor for unusual human activity on the network, as well as unauthorized applications connecting to a database.
Database monitoring vendor DB Networks sponsored the Ponemon study, which one security consultant said made the conclusions suspect.
"Database monitoring won't detect SQL injection (attacks)," Kevin Johnson, chief executive of Secure Ideas, which does penetration testing, said. "By the time the attack gets to where monitoring can see it, it looks like just another query from the application. This is a typical vendor solution."
Penetration testing of all business applications was more effective in preventing breaches, because "if you don't know the flaw is there, you can't prevent it from happening," Johnson said.
Paul Henry, a senior instructor at the SANS Institute, believed much more is needed beyond database monitoring for good security.
"I believe in a layered approach that perhaps should include a database firewall to mitigate the risk of SQL injection, combined with continuous monitoring of the database along with continuous monitoring of normalized network traffic flows," Henry said.
The study also found that almost six in 10 of the respondents believed the attacks against the retailers involved SQL injection as a component. Such an attack involves the insertion of malicious SQL statements into an entry field for execution. When successful, the technique can result in the database dumping its contents to the attacker.
The retailers have not disclosed whether SQL injection was involved in the attacks reported over the last eight months. The respondents' opinions were based on their own experience.
Nearly seven in 10 of the participants worked for organizations that must comply with Payment Card Industry Data Security Standard (PCI DSS). The standard is what retailers have to follow in order get approval to accept payment cards issued by banks.
"I would agree that SQL injection is likely involved because of its prevalence today, but I would also not yet draw a conclusion as we still do not have enough details," Henry said.
For example, Target has acknowledged that the credentials of a subcontractor that had access to the retailer's network were stolen. In addition, malware used to steal 40 million credit card numbers grabbed the data from the memory of the retailer's electronic cash register.
Target also had personal information taken from 70 million customer accounts.