When I look back at all the major bad security news in 2004, I wonder how the worlds computing infrastructure still manages to get up in the morning. It looks a lot worse in retrospect than it felt while it was going on.
Year of the worm
The year began with several new and successful worms, starting with Bagle. MyDoom and Netsky followed soon after. Most of these worms include a “backdoor” program allowing a remote attacker to take control of the infected system. Its generally assumed that the real purpose of these worms, as with earlier worms such as last years Sobig, is to provide a platform for spammers and other sociopaths to do their dirty work. Nevertheless, all three brought us some innovation.
Later versions of Bagle were the first not only to send their payload in a ZIP file in order to evade some blocking methods, but to password-protect the ZIP file and include the password in the body of the message. MyDoom achieved fame for its built-in distributed-denial-of-service attacks against Microsofts and SCOs Web sites. Microsoft was able to dodge the bullets, but SCO had some tough times.
Netsky was not a sophisticated worm, but it was extremely successful and began a bizarre war with the Bagle worm, in which each worm attempted to interfere with the execution of the other. Netsky was winning the battle until its author was arrested.
It wasnt the end of worm innovation in 2004. A new worm known variously as Bofra and as a MyDoom variant actually sets up a Web server on the infected system and mails out links to it to spread itself.
Spam takes over, authentication standards process begins
By most measures now spam is at least two-thirds of all mail on the Internet,, and is occasionally as high as 90 percent. At this rate, by the end of 2005 finding your real e-mail will be like panning for gold at the end of the rush.
In addition to the many great products that appeared in 2004 to help users and organizations control spam, a consensus emerged to support efforts, begun in late 2003, to fix the lax Internet e-mail standards that make spam so unstoppable. The general idea was to add authentication to the process, so that when someone sent an e-mail purporting to be from a particular address, the recipient could have some confidence that it actually came from that address.
The IETF fast-tracked a group working on such an authentication scheme. MARID (MTA Authorization Records In DNS) seemed to be marching towards a new standard, based on work from a little company named Pobox and from the somewhat larger Microsoft, when discussions collapsed in a heap of contention over Microsofts intellectual property claims to their work and technical objections to the proposal. Seeing their consensus and momentum going down the drain, the chairmen of MARID dissolved the group, leaving the authentication movement in a temporary shambles.
The groups doing real work toward a standard didnt close up shop, though. By Thanksgiving, the standard on which Microsoft and POBox had worked had actually progressed to the point where many large ISPs, including AOL, agreed to give it a try. Yahoo integrated its A separate approach called DomainKeys, which relies on cryptographic signatures, was integrated into developer Yahoos mail system and a few others, including Googles Gmail.
Next page: Phishing takes off.
Page Two
Phishing takes off
Attacking perfect strangers for the sake of attacking them must have become blasé in 2004, because the hackers all seem to have gone phishing. My first column of the year was about phishing attacks, but I am surprised at how common it is now to receive messages from support@somebankIdontdealwith.com.
The efforts to create an e-mail authentication system are not just motivated by stopping spam. The same efforts would make it difficult to send phishing attacks—or at least most of them would. The real motivation behind adding the Microsoft Caller ID parts to the Sender ID specification was to fight phishing, because SPF, the other part of Sender ID, did nothing to stop spoofing of the From: address, the one the user sees (potentially as billing@paypal.com).
Just when you think phishers cant get any more immoral, they set the bar even lower. This year we saw a fake attempt to solicit contributions for the Kerry campaign. The campaigns Internet people seem to have been pretty savvy about it and were able to sabotage the effort.
Phishing attacks were increasingly used exploit vulnerabilities and install Trojan horses, but social engineering is still the bread-and-butter technique. Specifically, a link to one Web page disguised to look like a link to another Web page; the true target of the link is a fake page. But weve also begun to see anti-phishing systems that specifically look for this.
A vulnerable year
There were a number of significant security vulnerabilities in a variety of products in 2004, but as usual the ones with the biggest impact were in Windows and Internet Explorer.
Well consider three of this years major problems. Perhaps the most widely-exploited vulnerability of the year, used by several prominent worms, was a bug in the Windows LSASS (Local Security Authority Subsystem Service). The bug was remotely exploitable without any user intervention on the computer—a true nightmare of an exploit. To make matters worse, Microsofts first attempt to patch this problem was itself buggy.
The most prominent worm exploiting this bug, Sasser, is still endemic, even though its author, also the principal author of the Netsky worms, was arrested in May.
The ASN.1 bug was scary not just because it allowed for potential complete compromise of the system, but because the vulnerable component is so pervasive throughout Windows. The ASN.1 system, which describes the encoding of data used by the system, was also the source of vulnerabilities in other systems, including Kerberos and a VPN. But this bug, despite potential for widespread damage, was not the source of any serious wave of attacks.
Later in the year, Microsoft revealed that the code in Internet Explorer and other programs that parsed JPEG graphics files had a bug that could allow an attacker to take over a system simply by getting a user to view the graphic. The potential for malicious JPEG files is particularly concerning since most anti-virus scanners dont scan graphics. The fallout from this was also not as bad as it might have been.
SP2 to the rescue
A new, stealth version of Windows was released in 2004, but Microsoft called it a service pack. Service Pack 2 for Windows XP made profound changes in the inner workings of the operating system, intentionally breaking existing third-party products in the name of improving the security of the system. It has been successful in both regards.
Since SP2 was released in August, researchers have discovered only a few security problems in spite of unprecedented scrutiny. Its record compared with earlier versions of Windows is stellar, but it does have its drawbacks. Users have reported a large number of applications that dont work on it, and some have complained that their systems wouldnt come back up after they upgraded to SP2.
Still, SP2 users are more resistant to attack than their counterparts. As they grow through new system shipments, especially in the consumer space, the opportunities for attack will decline.
Next page: The appliance-ization of security.
Page Three
The appliance-ization of security
Security vendors moved heavily toward implementing their products in network appliances. This happened most aggressively in the enterprise space, but its happening even in small and midsize businesses. This is part of the good news this year: Security solutions are getting easier, cheaper and more powerful. These devices are typically Linux PCs in disguise, running an application that could run on a regular PC.
Appliances for firewall protection and other general network perimeter security have been common for years. Such devices with sophisticated functionality broke the $1000 barrier this year, putting them in reach of SMBs. We also found appliances focused on spam and virus detection, single-sign on and encryption acceleration.
Some of the innovative developments included an appliance from Juniper that secures Web-based conferences. The device itself is not without problems, but its a good example of how application-focused appliances are becoming.
Firefox: The first decent IE alternative
The Mozilla browser has been available for some time and has had fans, but the noise it generated in the browser “market” (if you can call it that) was as a mere firecracker next to the thermonuclear device that is Microsofts Internet Explorer.
But in November, the Mozilla Foundation released Firefox 1.0, a simpler version of the Mozilla browser without all the ancillary features that were also made available separately. Firefox was well-received, and it has started to make a dent in Internet Explorers market share.
Many people were clamoring for a good replacement for IE, given how many of Windows vulnerabilities are, in fact, Internet Explorer vulnerabilities. But the jury is still out. Firefox has had its own vulnerabilities.
If Firefox maintains a better record than IE, it might become respectable enough for corporations to adopt it, and maybe even for OEMs to preinstall it and set it as the default browser. Firefox advocates dont dwell on the fact, but the browser goes to great pains to duplicate many IE user interface elements, so at least part of what makes Firefox acceptable is that it doesnt shock people who like IE. Perhaps if theyre shocked enough by the next IE security disaster, theyll rush to Firefox.
Nonstories of the year
I also like following the stories that were supposed to happen but didnt, and we had some big ones this year.
None was bigger than the great e-voting disaster that just had to happen. For a variety of reasons—some plausible, some paranoid—many people felt that increased use of electronic voting in this election guaranteed increased voting fraud through subversion of flaws in the e-voting systems.
And indeed there were problems, as just about anyone could have predicted. An underground conspiracy theory continues to bubble—though it has settled down lately—that “evidence” such as the exit poll numbers indicate the vote was “hacked.” But those of us who took the precaution of wearing our tinfoil hats realize that if there were really something to it the Democratic party would be complaining, too.
A few months ago, we got a warning of a hacker effort to bring down the Net. “E-Jihad” had the potential, we were told, to cripple the Internet for at least several hours. Of course, nothing happened. There was a similar nonevent in 2003—a Web page defacement contest that fizzled.
Remember how after Microsoft bought RAV AntiVirus in 2003, it was supposed to leverage its monopoly to crush the anti-virus industry? It didnt happen. I suspect that they dont want to put out a loser, and they think that the technology they bought isnt good enough to simply put a “Microsoft” label on and sell. And thats because Microsoft, unlike many of their observers, knows that it doesnt always win, monopoly or not.
Windows source code leaks
In February, source code from Windows 2000 was leaked to the Web by someone with access to it. It turned out that the code came from Mainsoft, a development company with a license to it. Their systems had been compromised by an attacker who was recently caught.
But conventional wisdom immediately leapt to the conclusion that hackers would scour the source code and find large collections of security holes that they wouldnt have without the source. This hasnt happened for reasons that were foreseeable at the time.
Do you think 2004 was a rough year? Were the problems serious, or just annoying? Let us know in the TalkBack section below.
Security Center Editor Larry Seltzer has worked in and written about the computer industry since 1983.
More from Larry Seltzer
Check out eWEEK.coms for the latest security news, reviews and analysis. And for insights on security coverage around the Web, take a look at eWEEK.com Security Center Editor Larry Seltzers Weblog.