[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ISN] To Fix Software Flaws, Microsoft Invites Attack

Forwarded from: William Knowles <wk@xxxxxxx>


September 29, 2003

Microsoft's Security Response Center in Redmond, Wash., is the 
computing equivalent of a hospital emergency ward. When a problem 
comes in the door the center's director, Kevin Kean, and his staff 
must swiftly make an assessment: Is the security weakness detected in 
a Microsoft software product only minor? Or is it possibly so serious 
that, if exploited by a vandal's malicious code (as happened last 
month with the Blaster worm) it might crash computers and networks 
around the world?

If the threat appears grave, the problem goes immediately into the 
center's emergency operating room, where it is attended to by a team 
of Microsoft engineers, working nearly round-the-clock to analyze the 
flawed code, anticipate paths of attack, devise a software patch to 
fix the defect and alert millions of customers of the problem and the 

"It's triage and emergency response - so it's a lot like an E.R. ward 
in that sense," Mr. Kean observed last week.

The race to protect the computing patient has begun again.

On Sept. 10, after Mr. Kean's team completed another E.R. mission, 
Microsoft issued an emergency warning of a critical vulnerability in 
its Windows operating systems and released a patch - its 39th so far 
this year. What particularly worries computer professionals about the 
warning is that the security hole in Windows is the same kind of flaw, 
in the same feature of the operating system, that was exploited in 
August by the notorious Blaster worm.

Those who monitor Internet crises know that once Microsoft raises the 
alarm and releases a patch, a curious race begins. Digital vandals - 
those who write worms, viruses and other rogue programs - eagerly 
download the patch and reverse-engineer, taking it apart to search for 
clues on how to exploit the very Microsoft security hole the patch was 
meant to cover.

Some portion of Microsoft customers, from corporations to home PC 
users, takes the time to download the patch, but most do not. 
Meanwhile, there is a scramble to write malicious code and spread it 
across the Internet.

The Blaster worm was sighted on the Internet 25 days after Microsoft 
warned of that security hole. The company issued the latest warning 19 
days ago. So if recent history is a guide, Blaster 2 may be coming 
soon to a computer near you.

The brand-name worms and viruses of the last couple of years - 
Blaster, SoBig, Slammer, Code Red, Nimda, ILoveYou and others - are 
simply the most virulent representatives of an alarming surge in 
attacks by malicious programmers. 

The CERT Coordination Center at Carnegie Mellon University, which 
monitors rogue computer programs, reported 76,404 attack incidents in 
the first half of this year, approaching the total of 82,094 for all 
of last year. And the 2002 incident count was nearly four times the 
total in 2000. If anything, the CERT statistics may understate the 
problem, because the organization counts all related attacks as a 
single incident. A worm or virus like Blaster or SoBig, a 
self-replicating program that can infect millions of computers, is but 
one event.

The security flaws Mr. Kean's team is scrambling to catch and patch 
are part of the larger problem with software today. The programs that 
people rely on for all manner of tasks - from writing reports and 
sending e-mail, to monitoring factory floors and managing electric 
power grids - are becoming increasingly large, complex and, all but 
inevitably, filled with bugs. The problem is magnified by the fact 
that most computers are now linked to the Internet, enabling programs 
to travel around the globe and mingle with other programs in 
unforeseen ways.

Most software bugs are a result of small oversights by a programmer. 
And most large software programs are combinations of newer code and 
old code, accumulated over time, almost as if in sedimentary layers. A 
programmer working years ago could not have foreseen the additional 
complexity and the interaction of software programs in the Internet 
era. Yet much of that old code lives on, sometimes causing unintended 

Security holes, computer experts say, are a manifestation of the 
fragile and often unreliable software foundation that underlies 
today's economy. "These worms and virus attacks are just the visible 
tip of a massive iceberg," said Peter G. Neumann, a computer scientist 
at SRI International, a research firm.

The major rogue programs all exploit vulnerabilities in Microsoft 
products, and Microsoft is the leading target of criticism by computer 
security experts. Indeed, Microsoft must shoulder a lot of the 
responsibility for the security woes suffered by its customers, 
analysts say. But the security weaknesses in Microsoft products, it 
seems, stem mainly from the company's success as the leader of the PC 
era of computing.

The PC business model has been to push products out the door fast, add 
features constantly and market each product version as a millennial 
event. Microsoft perfected the model and attracted millions of 
customers. But security experts note that the PC business model has 
not placed much value on building secure, well-engineered software.

The other reason Microsoft is the white whale for most digital vandals 
is that more than 90 percent of all desktop PC's run on the Windows 
operating system software. And the company's Office package of 
programs has more than 90 percent of the market for word processing, 
spreadsheet and presentation software.

Other operating systems like Linux, Unix and Macintosh, experts say, 
all have security vulnerabilities. "But they don't get the attention 
and the attacks because, unlike Microsoft, the other technologies are 
not deployed on 300 million computers," said Russ Cooper, a security 
expert at TruSecure, a computer security company. "This is not just 
Microsoft's problem."

The task of making software more reliable and secure will not be quick 
or easy. But computer scientists and industry analysts say that the 
goal is achievable, and that some encouraging steps have been taken. 
Improvements, they note, will depend largely on changing attitudes in 
the marketplace so that software makers have a greater incentive to 
invest in building better software.

"By and large, vendors build what people are willing to pay for," said 
Edward Lazowska, a professor of computer science at the University of 
Washington. "People have historically been willing to pay for features 
- not reliability or security."

There is evidence, though, that corporations and the federal 
government are placing a greater emphasis on obtaining secure 
software. Within the last two years, the government has pushed 
security initiatives in its technology policy, especially in the 
aftermath of the Sept. 11 terrorist attacks.

Recent moves by the government include placing greater emphasis during 
the purchasing process on software design and reliability standards 
like the Common Criteria and the National Security Telecommunications 
and Information Systems Security Policy No. 11, a Pentagon directive 
that went into effect 14 months ago.

Such standards now apply mainly to the Department of Defense and 
national security agencies, but Congress is looking to extend similar 
standards to other federal agencies. The federal government is the 
world's largest buyer of information technology, spending nearly $60 
billion a year.

"If the government made a serious commitment to buying better 
software, it would change the industry," said Mary Ann Davidson, chief 
security officer of Oracle, the big database software company.

Two weeks ago, the House Subcommittee on Technology, Information 
Policy, Intergovernmental Relations and the Census, which is under the 
Committee on Government Reform, held a hearing on the impact of the 
Pentagon's programs to link procurement to tighter security standards 
for software.

Representative Adam H. Putnam, the Florida Republican who is chairman 
of the subcommittee, said he saw great promise for adopting similar 
standards for civilian agencies. "The government can leverage its 
purchasing power," he said, "and can be a leader for the entire 
industry in setting rules and standards of engineering behavior."

A decisive step toward changing market incentives would be to expand 
product liability law to include software products. So far, software 
companies have sidestepped liability suits partly by selling customers 
licenses to use their programs, not own them, with a lengthy list of 
caveats and disclaimers.

The industry has resisted any suggestion that software should be held 
legally liable for bugs. The industry's argument is that software is a 
highly complex product, which users tend to misuse or modify, so 
trying to assign responsibility for a failure would be unfair to any 
single company.

Whether the software industry can continue to operate beyond the reach 
of product liability suits is uncertain.

A report last year by a panel of the National Academy of Sciences, 
"Cybersecurity Today and Tomorrow: Pay Now or Pay Later," included the 
recommendation that "policy makers should consider legislative 
responses to the failure of existing incentives to cause the market to 
respond adequately to the security challenge."

Professor Lazowska, a member of the panel who at times has advised 
Microsoft, explained, "You could draw an analogy to auto safety, where 
a set of government actions has caused automobiles to become far more 
safe over the course of the past 35 years."

Technology is giving programmers tools to build more reliable 
software. The Java programming language, created at Sun Microsystems, 
and C#, developed later by Microsoft, are technologies for creating 
"managed code," which sharply limits the damage that can be done by 
errant lines of programming. "You have to design it so that bad things 
don't happen when programmers make mistakes," said William Joy, the 
former chief scientist at Sun.

At Microsoft, much more time is now being set aside in the design 
cycle of products for security considerations, a mandate approved by 
senior management this spring. "There is a shift from mainly an 
emphasis on working features to an emphasis on trustworthy and secure 
computing," said Steven B. Lipner, director of security engineering 
strategy at Microsoft.

Some of the tougher security standards, Mr. Lipner said, have shown 
measurable improvement in Windows Server 2003, which shipped earlier 
this year. The number of security vulnerabilities detected so far is 
half as many as at this stage after the release of Windows Server 
2000, Mr. Lipner said.

Yet years of steady progress in the quality of software engineering 
will be needed for big gains in security and reliability to become 
apparent. And it starts with education, noted Shawn Hernan, a security 
specialist at CERT. He makes a game of seeing how quickly he can find 
security vulnerabilities in the programming examples used in college 
textbooks. It rarely takes him more than few minutes.

"The textbook examples are riddled with vulnerabilities," Mr. Hernan 
noted. "Computer science culture is based on, build it, get it working 
and fix it later. We need a culture change away from the cowboy and 
toward the engineer."

Even as his E.R. team scrambles to patch Microsoft's security holes, 
Mr. Kean agreed. "It's not just Microsoft," he said. "The world will 
commit itself to more secure computing. There will be a cultural 

"Communications without intelligence is noise;  Intelligence
without communications is irrelevant." Gen Alfred. M. Gray, USMC
C4I.org - Computer Security, & Intelligence - http://www.c4i.org
Help C4I.org with a donation: http://www.c4i.org/contribute.html

ISN is currently hosted by Attrition.org

To unsubscribe email majordomo@xxxxxxxxxxxxx with 'unsubscribe isn'
in the BODY of the mail.