In 2004, I came across an empirical study published by the CERT/CC that indicated a diminishing correlation between the number of vendor-issued vulnerabilities and the number of reported security incidents. In the years prior to 2002, the number of reported security breaches had always been proportional to the number of vendor-published vulnerabilities. That corollary made sense, since attacks and worms followed vulnerabilities. However, in 2003 and beyond this was no longer the case. The number of incidents rose dramatically as compared to the number of published vulnerabilities.
In February of 2005, I researched a news article that reported a Florida businessman, Joe Lopez, had lost over $90,000 from his online bank account at the Bank of America. The compromise was attributed to a keystroke logging Trojan named Coreflood that was found on Mr. Lopez's computer during a U.S. Secret Service investigation. After reading that article, the big question seemed to be, "who was to blame, the customer, Joe Lopez, or the financial institution, the Bank of America?"
In May of 2005, my company was hired to investigate a forensics case similar to the one described above. The user had lost $50,000 in three separate fund transfers from her online account. The bank in question had hired experts to assess the security of their own networks and systems. The results indicated that the bank's systems were immune to any attack that may have resulted in this compromise -- and they believed they did their part. Forensic analysis of the compromised user's home computer provided the evidence of the compromise -- a keystroke-logging Trojan that was tailored to capture the user's credentials; electronically mail them to a server several thousands of miles away; and then delete itself. So who's to blame, the "unaware" end-user or the "negligent" bank?
Recently, I received a phone call from my cousin in London informing me that he had lost £8000 in a similar manner to the scenario described above. This conversation proved to be my moment of enlightenment. Eureka! It's the end user attacks that are superseding the conventional attacks we've seen against servers for years, ultimately increasing the total number of reported security incidents dramatically.
The pieces of the puzzle were coming together. With over 200 remotely exploitable vulnerabilities in Internet Explorer, the market leader in web browsers, and 49% annual growth in the number of broadband users; and 26% growth month-over-month in the number of phishing sites on the Internet (of which 78% targeted financial institutions), the rise in attacks against end users for their banking credentials was becoming common.