- Related Stories
-
Yahoo launches 100MB of free e-mail
June 15, 2004 -
Akamai glitch slows sites
May 24, 2004 -
Alarm growing over bot software
April 30, 2004 -
SCO waits for MyDoom attacks to end
February 12, 2004
The attack, which blocked nearly all access to Apple Computer, Google, Microsoft and Yahoo's Web sites for two hours on Tuesday, took aim at the key domain name system (DNS) servers run by Akamai. These servers translate word-based URLs, such as www.microsoft.com, into
![]()
![]()
![]()
![]()
![]()
![]()
Get Up to Speed on...
Enterprise security![]()
Get the latest headlines and
company-specific news in our
expanded GUTS section.
![]()
![]()
![]()
![]()
The deluge of data that hit the infrastructure provider was "so large that it (couldn't have) come from a couple of servers," said Tom Leighton, chief scientist and co-founder of Akamai. "Working with our network partners, we were able to identify a bot network that appeared to be operating and managed to shut it down, which resulted in stopping the attack."
Bot networks are collections of computers that have been compromised by software specifically designed to create a network of systems for attack. A bot--also known as remote-access Trojan horse program, or RAT--seeks out and places itself on vulnerable PCs. It then runs silently in the background, letting an attacker send commands to the system while its owner works, oblivious. The computers are essentially turned into zombies, controllable from afar.
|
The latest versions of bot software enable attackers to control and steal information from compromised computers via chat servers and peer-to-peer networks. These PCs can then be commanded to infect or attack other computers. Security experts have identified bot networks as a critical threat to the Internet.
A common use of a bot network is to order a compromised PC to send seemingly legitimate network information to a single destination, resulting in a torrent of data that overloads the target servers. Such a distributed denial-of-service, or DDoS, attack can block access to a Web site for several hours or even days.
A security professional who participated in investigating the attack confirmed that the DDoS attack apparently came from an extremely large botnet.
"If it was (a) bot, it was very well written and it was very large," the security expert said on condition of anonymity. "As far as we could tell...it all looked like real and legitimate traffic."
While Tuesday's attack was aimed at bringing down the four major Web sites, Akamai's Leighton said his company was the true target.
"At the high level, it was clear that this attack was focused on a subset of our customers," he said. "We assumed they were attacked as a way to get at Akamai."
What remains unclear is how the DDoS attack could be so selective as to focus on the main Yahoo, Google, Microsoft and Apple sites. Distributed attacks are typically blunt instruments rather than scalpels, as evidenced by the mass outages caused by this method in 2000.
Keynote Systems and other Internet performance companies said Web traffic actually dipped during the attack, raising questions about the volume of data sent to Akamai's servers. Typically, a large-scale DDoS would be observed as an increase in network traffic.
Nonetheless, DDoS attacks are getting sophisticated, especially in the variants of computer viruses that have recently surfaced. The Netsky virus used such a technique to target Kazaa and other file-sharing networks, disrupting service at some. Earlier this year, the main Web site of the SCO Group was crippled after attacks from computers infected by the MyDoom virus.
Akamai refused to provide greater detail about Tuesday's attacks, citing a need to keep mum on the details of the company's architecture and to avoid giving more publicity to the attackers.
"There was an extraordinary amount of traffic," Akamai's Leighton said.
See more CNET content tagged:
Akamai Technologies Inc., distributed denial of service, bot, zombie, attack





Most RAT?s can be detected by virus scanners, you have to make sure the scanner checks to Trojan?s, backdoors, etc and not just viruses and e-mail. Some versions of Norton don?t.
If we want this to be fixed then things need to change, we need the ISP?s to take some of the responsibility in that when things are reported don?t take a week to check on it. A week of an infected pc running is WAY to long. We need people to be understanding, not stubborn.
If everyone worked at the problem then it would be easier to battle, I don?t think we will win outright, but we can make a difference, if we try.
First...
1. Using currently available 'communications-mediums', ...IS A RIGHT. This HAS been ruled to be true by the Supreme Court of the United States (yes, they mentioned the 'Internet' specifically). If you don't like this fact, too bad...
Second...
2a. The REAL responsibility for such an incident falls FIRST upon the Virus-Writers themselves, ...if they even exist.
2b. The responsibility then falls squarely on "Microsoft", for producing a SERIOUSLY-FLAWED, and DANGEROUS, product which they sold to an unsuspecting public.
2c. "Granny" can no more be held accountable for the actions of criminals, or criminally-negligent companies, than "John Doe" could be if, for example, someone broke into his house, stole a flashlight, and then used the 'stolen' flashlight during the commission of another crime.
Such an assertion of culpability is simply ASININE.
Furthermore...
3. "Akamai's" claims are themselves currently quite suspect, since the actual evidence makes their assertions look like little more than a pathetic excuse from a company which has, in fact, had a string of embarrassing technical-problems.
4. The actual impact of this, "astonishing" event, was nothing more than yet another, all too common, 'website-access slowdown' (hardly the end of the Internet as we know it, or, a major threat to business).
And frankly...
5. The "threat" of, so-called, "Zombie-PCs" (along with the scourge of SPAM) currently seems to be a couple of the most popular "Cyber Boogie-men".
Of late, these "threats" are being used, mercilessly, as an excuse to limit all sorts of 'Freedoms' (by creating the technical-ability to monitor and control virtually all 'Internet access' and 'computer end-use').
Lets say I'm doing construction on my house. During the process, a pile of scap lumber has accumulated in my front yard. Under the legal doctrine of "attractive nuisance", I am responsible if a kid tresspasses onto my property, steals a 2x4, and then goes and whacks another kid upside the head.
You think net access is a right? Try not paying your ISP bill and see how long it takes for them turn off your connection. Have your rights been violated? How about sending spam? Is that a right too?
Freedom of the press is a right, but it only extends to those who own a printing press. You have no right to walk into your local newspaper and demand they print something for you. Newspaper have broad immunity from liability for what they print, but that immunity is not absolute. Papers are successfully sued to libel and slander.
It seems awfully selfish for a big corp. co. to w/hold info that helps the little man and I wouldn't be surprised if/when the starter of this DDoS see's the paragrapgh where they won't tell what it was (to not give them notariety) that the creators don't slam them once again.
And I have to say, as one of the little guys who just helped 3 friends computers go down after Tues attack, it would serve them right.
In order to protect all - you must share the info!
- It's a crime to stop the crime
- by June 22, 2004 2:46 PM PDT
- Say what, you ask?
- Like this Reply to this comment
-
(19 Comments)Yes, MS, Linux and other major OS's have security holes. Naive users are a major part of the problem, but the whole point of much of the modern internet is to make it available to naive users. Unfortunately we all will have a long wait for those security holes to be plugged, if they ever are completely plugged.
Yes, in many cases taking advantage of those security holes to form RAT packs is illegal. It should be illegal in all cases... but wait a minute! How do you bring the RAT commanders to justice?
The problem is, the internet is international! There will ALWAYS be places where this behavior is either legal or not prosecuted. So the RAT commanders send their commands to their zombies through other zombies in those jurisdictions.
The only way I can think of to get around the jurisdiction problems is to reverse RAT the RAT's, so that I could trace them back to their source.
But wait, that's illegal too!! So even if I penetrate all their layers of zombies commanding other zombies to find out who they are, I can't use any of that evidence against them because it was illegally obtained! At the very least, the law has to wink at those who do the very same thing to computers as the RAT commanders in order to find them.
So as long as the RAT commanders can send their commands through jurisdictions that can't, won't or don't care to persue them, they keep on happily messing up the internet for the rest of us. I and I imagine others like me won't persue them, even though we might know how, because it's illegal.
Thus my subject line: It's a crime to stop the crime.