Virulent new worms that exploit vulnerable instant messaging (IM) clients and could infect hundreds of thousands of computers in seconds are a real threat for Internet users worldwide, according to security researchers from Symantec.

A small but growing number of documented IM security holes and the rapid adoption of IM technology within corporations was combining to pose significant risks of infection and information theft, the two researchers said at the recent Virus Bulletin conference in Toronto.

Currently, there were about 60 published IM vulnerabilities, according to Eric Chien, chief researcher at Symantec Security Response in Dublin, Ireland. Those range from security holes that could be used to crash IM clients in denial of service (DOS) attacks to those that allow attackers to remotely install and run malicious code on computers running the vulnerable IM clients.

Such vulnerabilities were already being used by hackers to compromise individual machines, Chien said.

However, the IM vulnerabilities become particularly dangerous when they are combined with hacker applications written using documented APIs (application programming interfaces) from the major vendors and used to create an IM worm, according to Neal Hindocha of Symantec Security Response.

The APIs can be used to make tools that silently sent virus files or Trojan horse programs to IM users or captured a remote user's list of IM correspondents or "buddies," Hindocha said.

The ability of hackers to grab a user's buddy list also gave IM worms the potential to be more virulent than predecessors such as Code Red, Slammer or Blaster, which spread over the Internet rather than over IM networks, Chien said.

Unlike those worms, IM worms do not need to scan the Internet for the IP (Internet Protocol) addresses of vulnerable systems to infect, a process that greatly slows the spread of traditional worms. Instead, IM worms simply use the infected user's buddy list to find new targets.

Even with a scenario in which the buddy lists of infected and target machines were identical except for just one IM user, an IM worm could infect 500,000 machines in just 31 seconds, Chien and Hindocha showed.

Most popular IM networks, including America Online's AOL Instant Messenger, Microsoft's MSN Messenger and Yahoo's Messenger are designed to be fast and support hundreds of thousands of users, rather than be secure, the researchers found.

User passwords and other sensitive data was not encrypted and could easily be obtained from Windows systems or sniffed from IM traffic, Hindocha said.

While about 30 IM worms had been identified, there could be barriers to widespread IM worm outbreaks of the magnitude of Slammer or the recent Blaster worm, which quickly spread worldwide, Chien said.

Because communication on IM networks often relies on centralized servers, IM companies can quickly filter attack traffic for all users once new threats are identified.

In addition, IM companies can deny unpatched users access to the network, forcing them to upgrade their client software once patches were available, Chien said.

Finally, massive traffic generated by even 10,000 or 50,000 infected hosts would likely knock IM servers offline, halting the spread of the worm, he said.

Companies worried about the security threats posed by IM worms should seriously consider blocking communications ports used by common IM clients.

Full inspection firewalls were also useful for IM clients that communicated using essential ports such as those used for Web HTTP (Hypertext Transfer Protocol) traffic, he said.

Organisations should also assess the business need for IM, which Chien said was often not used for essential business communications.

Organisations that absolutely need to use IM should consider investing in enterprise-class IM products that use encryption to protect sensitive data and allow employees to communicate within a corporate intranet without exposing IM traffic to the public Internet, Chien said.