[ISN] Watchdogs on Way Out?

From: InfoSec News (isnat_private)
Date: Fri May 10 2002 - 01:37:55 PDT

  • Next message: InfoSec News: "Re: [ISN] Social Engineering: The Human Side Of Hacking"

    http://www.eweek.com/article/0,3658,s=701&a=26381,00.asp
    
    By Dennis Fisher 
    May 6, 2002 
    
    While much of the high-tech industry has spent the last several months
    focusing on security - for their information as well as their physical
    assets - a small but growing number of influential executives has been
    working toward the long-term goal of making the security industry
    obsolete.
    
    "The security industry as we know it today goes away in 10 years,"  
    said Chris Darby, CEO of @Stake Inc., a security consultancy and
    research company in Cambridge, Mass.
    
    As things stand now, the security market is a confusing and fractured
    mélange of technologies with exotic-sounding names such as firewall
    and IDS (intrusion detection system). Vendors in every segment tout
    their wares as the final piece of the puzzle, the magic potion that
    promises to make an IT manager's security headaches vanish.
    
    They play on customers' fears, telling them there are dozens, if not
    hundreds, of vulnerabilities in the software on which they're running
    their enterprises and that the only way to keep corporate data safe is
    to install yet another layer of security.
    
    However, the dirty little secret of the security industry is that if
    the big software vendors paid more attention to security, security
    hardware and software vendors would be out of business, according to
    experts.
    
    And that's exactly the scenario that companies such as @Stake,
    Microsoft Corp. and others are trying to bring about.
    
    Microsoft has long been a favorite target of crackers, much to the
    displeasure of customers that have been burned by vulnerabilities in
    the company's broad line of software. Microsoft has been quick to
    issue patches when someone identifies a new flaw in one of its
    products, but this spring, the company launched Trustworthy Computing,
    an all-out effort to improve the security of its products during the
    design and development phase, something its critics have suggested for
    years.
    
    The main focus of the effort is training developers to write secure
    code and eliminate common and easily exploitable vulnerabilities such
    as buffer overruns. @Stake, which offers secure-coding training
    services, has seen a lot of demand for those services in recent
    months.
    
    Such training and coding practices should lead to more secure products
    in the short term and in the long term, to a marked decrease in the
    number of vulnerabilities in corporate networks, which will mean fewer
    successful attacks, security insiders claim.
    
    And that, in turn, will mean less demand for security countermeasures
    such as firewalls and IDSes.
    
    "In the long term, over time, as we design more-secure products, what
    we should see - what we'd better see—is fewer successful attacks,
    better stability and better security," said Scott Charney, chief
    security strategist at Microsoft, in Redmond, Wash. "We should be able
    to measure and say vulnerabilities are going down. Things should
    improve."
    
    But "should" is far different from "will." Many of the common
    vulnerabilities that crackers use to invade corporate networks and Web
    servers have been around for years, if not decades. Buffer overruns,
    for example, were identified as early as the 1960s, and yet they
    continue to show up in new software packages such as Windows XP.
    
    This leads some security experts to challenge the notion that the
    security industry is on the verge of collapsing.
    
    "Today's simple-to-fix vulnerabilities, like buffer overflows, will
    likely be gone [in 10 years]," said Steven Bellovin, AT&T fellow in
    the Network Services Research Lab at AT&T Labs Research, in Florham
    Park, N.J., and a pioneer in network security. "But more complex
    semantic problems will remain. Most security holes are caused by buggy
    code. Buggy code is the oldest problem in computer science, and I see
    no reason to think that will change.
    
    "We'll make progress - but it's fundamentally a very hard, and
    possibly insoluble, problem," Bellovin said. "I also think that the
    right approach for systems architects is to design their systems
    differently. A lot more has to be done to understand what the
    security-sensitive modules are so that they can be made as small as
    possible and can be properly isolated from the rest of the system."
    
    But in the end, @Stake's Darby said he believes that the changes being
    made by companies such as Microsoft are moving the industry inexorably
    toward a fundamental shift.
    
    "The notion of overlaying security products on networks after the fact
    is inefficient," Darby said. "And technology in the long run is about
    efficiency."
    
    The changes that Darby, Charney and others envision for the security
    industry will not happen overnight. But if they do occur, users will
    eventually be better off, experts say. "I think that Microsoft will
    apply a lot of effort to fix their security problems because it seems
    that they have been ridiculed to the point that they have finally
    decided to get right with God," said Phil Zimmermann, chief
    cryptographer at Hush Communications Corp., based in Dublin, Ireland,
    and inventor of the PGP e-mail encryption program. "I think Linux,
    FreeBSD and Apple [Computer Inc.] will try to reach parity with
    OpenBSD in security discipline. These changes will take years but will
    eventually bear fruit. Firewalls and IDS as add-on products will
    become less needed."
    
    
    
    -
    ISN is currently hosted by Attrition.org
    
    To unsubscribe email majordomoat_private with 'unsubscribe isn'
    in the BODY of the mail.
    



    This archive was generated by hypermail 2b30 : Fri May 10 2002 - 04:55:54 PDT