Stealth virus warning sounded again

Kaspersky Labs has warned that malware authors have worked out that there is more money to be made from causing many low-key virus infestations than single, massive outbreaks

Virus authors are choosing not to create global epidemics — such as Melissa or Blaster — because that distracts them from their core business of creating and selling botnets, according to antivirus experts.

Botnets are groups of computers that have been infected by malware that allows the author to control the infected PCs, and then typically use them to send spam or launch DDoS attacks.

Speaking at the AusCERT conference on Australia's Gold Coast on Tuesday, Eugene Kaspersky, founder of Kaspersky Labs, said that the influence of organised crime on the malware industry has led to a change of tactics, echoing comments made in March of this year by Mikko Hyppönen of F-Secure. Instead of trying to create viruses and worms that infect as many computers as possible, malware authors are instead trying to infect 5,000 or 10,000 computers at a time to create personalised zombie armies.

"Do I need a million computers to send spam? No. To do a DDoS attack, 5,000 or 10,000 PCs is more than enough. That is why virus writers and hackers have changed their tactics of infection — they don't need a global epidemic," said Kaspersky.

According to Kaspersky, organised criminals are advertising networks of zombie computers for rent on underground newsgroups and Web pages. When they receive an order for a botnet of a certain size, they set about trying to infect computers using infected email attachments or socially-engineered spam with links to malicious Web pages. As soon as they infect enough computers to fulfil the order, they stop using that particular piece of malware.

"It seems that if, say, the virus author needs 5,000 infected computers, they put the Trojan on a Web page and wait for 5,000 machines to be infected. Then they remove the Trojan because that is enough. When they get a new request for another zombie network, they release a new Trojan — they are able to control the number of infected computers," said Kaspersky.

Adam Biviano, senior systems engineer at antivirus firm Trend Micro, agrees. He said that by only infecting a relatively small number of computers, the malware has a better chance of flying 'under the radar' and not being spotted by antivirus companies.

"It makes sense to have a discreet number of PCs under your control and be able to sell that on," said Biviano, who added: "With 5,000 PCs under your control — none of which are being destroyed or showing actual qualifiable damage as a result — you will fit under the radar, probably make some money and you probably won't get arrested".

Kaspersky said that to fight this new tactic antivirus companies have to be more thorough by scouring Web pages and email attachments for new and obscure pieces of malware — to ensure as few Trojans as possible get through to users.

"Before releasing the new infected code they test it using antivirus scanners and they don't release the new Trojan or worm if it is detected. I believe that if only 1,000 machines are infected, anti-virus companies will never receive the infected file. That is why antivirus companies have to collect data reactively and get samples as quickly as possible," said Kaspersky.

Vincent Gullotto, vice-president of McAfee's antivirus emergency response team, told ZDNet Australia that antivirus companies are responding to the new threat by proactively seeking out new forms of malware.

"It is standard for us, Kaspersky, Symantec and some of the other prominent antivirus companies scour the Web in many different ways. We go out looking for [malware] with a very aggressive search and we do passive searches where we have machines that are just sitting around waiting to get attacked. When we see a machine getting attacked we grab a sample rather quickly so we can add it to our database," said Gullotto.

Munir Kotadia reported from Sydney for ZDNet Australia. For more ZDNet Australia stories, click here.