CNET también está disponible en español.

Ir a español

Don't show this again

Security

Mysterious Net traffic spurs code hunt

Worm? Trojan? Attack tool? Network administrators and security experts search for the cause of an increasing amount of odd data that has been detected on the Internet.

Worm? Trojan? Attack tool? Network administrators and security experts continue to search for the cause of an increasing amount of odd data that has been detected on the Internet.

Security software firm Internet Security Systems (ISS) on Thursday declared victory, saying that a new hacker tool that scans for paths into public networks was responsible. But many other security professionals--including those at Intrusec, the company that originally tracked down the hard-to-find code--believe that ISS jumped the gun.

The real culprit likely is still out there, said David J. Meltzer, founder and chief technology officer of Roswell, Ga.-based Intrusec.

"It is possible that (the tool's code) is causing some of this traffic," Meltzer said. However, he added, key differences between the data that was captured by security professionals and data created by the code suggest that the hacker tool isn't the original culprit. "That would leave us to believe that there is something out there that is creating the (data) packets that isn't this Trojan."

However, the security researcher hastened to add that the traffic seen by network administrators isn't ominous. It merely has piqued the curiosity of quite a few researchers.

"I don't think it is a serious threat because it's not self-replicating," Meltzer said. "And it hasn't caused serious disruptions to anyone."

Since mid-May, security researchers and network administrators have tried to track down the source of odd traffic that they have been seeing on their networks. The data frequently attempted to connect to nonexistent servers or to services not offered by existing servers. The only common thread seemed to be that the data packet had a window size of 55,808 bytes and, in many cases, came from a nonexistent Internet address.

The window size is a parameter used by TCP/IP networks to specify the amount of data, in bytes, that devices can send without receiving an acknowledgment from the receiver. The parameter, normally set by the operating system, can make a network more reliable, if set to a lower number of bytes, or faster when long delays are present, if set to a high number of bytes. Typically, when first connecting to another computer, a device on the Internet will use a lower window size--say, 1,024 bytes.

"None of the operating systems are going to start with a window size of that (55,808 bytes) size," said Joe Stewart, senior intrusion analyst with managed security service provider LURHQ (pronouced "lurk"), who also did a significant amount of analysis on the strange traffic.

In late May, Stewart detected a computer trying to connect to a nonexistent computer on the company's network. The only problem with the traffic was that its supposed source was from an Internet address that couldn't exist; the Internet Assigned Numbers Authority, or IANA, had reserved the block of addresses that contained the number, so no Internet router would have forwarded such data. While such an Internet address in the source field of a data packet will often be transmitted by less vigilant Internet service providers, almost no provider would forward data that had such an address in the destination field. So, while the analyst had seen more than 900 such attempts to connect to a server on his network, he could not track the data back.

"It could be a broken scanner or someone trying to map the address space," Stewart said. "I don't think people have really nailed down what was the source of this."

ISS believes the code given to the company by Intrusec is responsible.

"We feel that a significant portion of the traffic is caused by this code," said Dan Ingevaldson, engineering manager for ISS's incident response team. Previous traffic that doesn't match the current code's data patterns may have been caused by an earlier variant of the code, he said.

Theories abound about what the code does. Some people initially believed the data was sent by a worm that used the Internet relay chat (IRC) system, a precursor to the popular instant-messaging networks, to communicate. Others thought that the data had been a low-level denial-of-service attack. Security firm Network Associates classified the code as a worm, which it dubbed W32/Randex.c.

However, the most current theory based on analysis of the code seems to indicate that what Intrusec found is a scanner and attack tool that communicates by sending out information to random addresses hoping that another computer infected with the program would be at the other end. The code has so many bugs, however, that it hasn't worked properly.

"It is very buggy," Ingevaldson said. "It didn't even write information to its data file correctly."

Even if it didn't have flaws, the program would have a tough time working. While it hides the location of the sender and receiver, the technique of broadcasting out data to essentially anybody on the Internet is inefficient at best.

LURHQ's Stewart believes the code that has been found is a research effort by some anonymous would-be attacker. However, he's not so sure that something else isn't out there.

"Until someone comes up with analysis that matches what I am seeing, I'm going to wait," he said.