Identifying improper behaviour among the devices connected to their network is a critical tool for any organisation concerned about Advanced Persistent Threats (APTs). In light of the rapidly changing landscape of such targeted malware attacks, Fortinet lists the top five types of behaviour that might indicate that a device has been infected.
a) Bad Connection Attempts
Typical malware behavior often includes attempts to connect to hosts that dont exist on the Internet. While some bad connections may be due to user error or bad links, a series of bad connections could be a sign of malware infection.
b) Choice of Application
A host that installs a P2P file sharing application can be considered riskier than a host that installs a game. Some consider both actions problematic. The ability to add weights to each action allows each risk to be scored accordingly.
c) Geographic Location
Visits to hosts in certain countries can be categorised as risky behaviour, especially if there is a significant amount of traffic involved. Identifying such behaviour can be combined with a white list approach that identifies legitimate sites in such countries to help identify infected clients.
d) Session Information
When a device starts to listen on a port to receive a connection from the outside but does not initiate a connection, an APT infection could be the cause.
e) Destination Category
Visiting certain types of websites, such as gambling and adult sites as well as those known to contain malicious code can also be a predictor of APT infection.
Identifying risky user and application behavior represents the next step in protection against Advanced Persistent Threats. Signature-based protection is no longer enough. Its important to build a complete, evolving and up-to-date picture of the behaviour of network clients, said Vishak Raman, Senior Regional Director, Fortinet, India & SAARC. Client reputation and scoring is an essential component in ordering and understanding the enormous amount of security information available within organisations, and applying it to a dynamic, targeted security response.
Add new comment