To be profitable, cybercriminal operations need to be both large scale and resilient to takedown attempts. In this talk, I show that both these requirements can be used against attackers to develop robust mitigation systems. Such systems are difficult for cybercriminals to evade, because they cannot avoid detection without suffering an economic loss. As examples of this research vision, I first present SpiderWeb, a system to detect web pages serving malware. SpiderWeb operates by analyzing the complex network of servers that real-world victims traverse before reaching malicious web pages. Then, I discuss a set of techniques that I developed to detect and block malicious activity on Online Social Networks. In particular, I present COMPA, a system that protects social network users from having their accounts compromised by cybercriminals. At the end of the talk, I discuss the change of mindset that is required to fight emerging advanced threats, such as targeted attacks.

Gianluca Stringhini is a PhD candidate in the Department of Computer Science at University of California, Santa Barbara. He works in the Computer Security Lab, advised by Professors Christopher Kruegel and Giovanni Vigna. In his research, Gianluca develops techniques to detect and block malicious activity on the Internet. He was awarded the Symantec Research Labs Graduate Fellowship in 2012, and his work on detecting malicious accounts on social networks received the Best Student Paper Award at ACSAC in 2010.