Over 90 percent of data breaches are the result of external attacks and almost 60 percent of organisations discovered them months or years later, Verizon claimed at the RSA security conference.
Called the Verizon 2011 Investigative Response Caseload Review, it compiles statistics from 90 data breach cases investigated by the company's incident response team last year, and provides a preview of Verizon's larger annual report that will contain data collected from additional sources like national CERTs and law enforcement agencies.
The report concludes that 92 percent of data breach incidents have had an external cause, which conflicts with the findings of other security vendors, according to whom most data breaches are the result of internal threats.
"I think that's a bit of a myth in the security community," said Wade Baker, director of risk intelligence for Verizon Business. "There's fewer people inside of an organisation than there are outside and I think it stands to reason that, by the numbers, we will have more external incidents."
However, Baker agreed that many insider actions resulting in data breaches get dealt with internally and aren't necessarily accounted for in Verizon's report, which only compiles information from incidents Verizon was called to investigate.
According to the report, the use of default or stolen credentials was one of the primary methods that attackers used to gain access to data in 2011.
Some organised crime groups have automated their attacks to scan for very specific ports, like those for remote desktop, pcAnywhere and similar products, and then they try to login with common or stolen passwords, said Chris Porter, principal for Verizon's RISK Team.
This problem is common with small businesses that outsource the administration of their IT systems to third parties who offer remote support. These organisations should implement some type of access control for remotely accessible systems, like restricting which IPs are allowed to connect to them, Porter said.
Web-based attacks like SQL injection have a lower frequency and didn't even make the top ten list on the annual report that will be published later this year, Baker said. The rate of SQL injection attacks is usually much higher for financial services organisations.
One problem that doesn't seem to improve from year to year has to do with breach discovery. It takes the majority of organisations months to discover a breach and some of them even take years, according to Verizon.
Furthermore, the discovery of these breaches is usually the result of external action, like law enforcement agencies or other organisations informing the affected companies, Baker said.
Restricting outbound connections to ports that are necessary for the organisation's communications is equally important as monitoring incoming traffic, Porter said. This is known as egress filtering and can help prevent breaches.
Whitelisting applications that are allowed to run on servers can prevent attackers from installing and executing malware on systems they've gained access to. It's much easier to allow only some applications and services to run and block everything else, than to build a list of malicious programs and continuously add to it, the Verizon researchers said.