"Client-side attacks are where the paradigm is going," Smith said. "Monolithic security systems no longer work."

Hackers use Web browsers as exploitation tools to spread malware and collect sensitive information. Smith used examples from clients of his company, which analyzes and researches computer attacks, to demonstrate the threat posed by blog spam and SQL attacks.

Attackers targeted high-traffic sites with blog spam and posted comments on blogs, he said. The comments looked odd and tended to have non-English phrases placed in large blocks of text with random words hyperlinked, he said. Clicking on such links took users to sites that seemed like blogs but were pages loaded with malware, Smith said.

A Chinese bank owned the domains for each malware site, but the IP (Internet Protocol) addresses traced to Germany. Studying the links revealed that each one contained words in Russian or Romanian, said Smith. By placing an international spin on their nefarious activities, the hackers hoped to confuse anyone investigating their work, he said.

"How are you going to track these back to the bad guys?" he said, noting that tracking is complicated by language barriers, working with foreign law organizations and dealing with countries "that just may not want to talk to us."

While the goals of blog spam attacks remain unclear, Smith said financial incentives serve as motivation. Adware installed after a user visits an infected site nets a hacker money, as does clicking on an advertisement on the page. Other hackers are looking to expand their botnets, or networks of compromised machines used for malevolent purposes.

Smith's investigation traced the attacks to a home DSL account in Russia. The international nature of the incident made prosecution unlikely, he said.

The SQL injection attack Smith discussed originated in China and attempted to steal information on the businesses that visited the Web site of the company, which was Smith's client.

Hackers first launched a SQL injection and uploaded a back door that allowed them to take control of the system.

Additional SQL injections failed, so the hackers searched the system for another exploit. They found a library application that allows images to be uploaded. Hackers uploaded a GIF file with a line of code contained in the image. The computer system read the GIF tag and uploaded the photo and automatically executed the code.