Share
Embed
Disclosure Samsara
Disclosure Samsara or "The Endless Responsible Vulnerability Disclosure Debate" (presented at NAISG and OWASP Boston, 2009)
Zach Lanier
October 21, 2009
Transcript
37 slides
-
Disclaimer Side effects of this presentation may include: anger; giddiness;
joy; introspection and navel gazing; debate; yelling; nausea and hallucination; patting me on the back; or punching me in the face -
“Long is the night to the wakeful, long is the
league to one who is exhausted. Long is samsara1 for the foolish who do not know saddhamma2.” -from the Dhammapada 1 Endless cycle of suffering caused by birth and death 2 The highest truth -
Quick “about me” I’m Zach Lanier and I'm... • a
hacker • a security consultant/researcher • co-founder of Midnight Research Labs Boston • manager of the Security Twits list -
Agenda Foundations WWYD? A dose of truth “Ancient” and Modern
History Current challenges Nirvana(-ish) -
Foundations Full disclosure Revealing all details of a security weakness,
flaw, or vulnerability Pro: More information for workarounds; can pressure vendors/providers; public scrutiny Con: Quicker time to exploit development; 0-day buffet! -
Foundations Partial disclosure Revealing only details that are deemed relevant
(minimal disclosure?) Pro: May reduce likelihood of exploit development; may increase time-to-publish Con: “Relevant” is totally subjective; may miss important details -
Foundations Zero disclosure Revealing no (or almost no) details of
a vulnerability; “security through obscurity” Pro: “Almost zero likelihood of exploit development” (well, it’ll probably happen anyway) Con: Less public scrutiny may reduce trust if/when disclosure does occur; “Many eyes make bugs shallow” -
You’re a software giant, known for your popular operating system.
Late last night, an independent security researcher released exploit code for a previously unknown buffer overflow that leads to remote code execution. What Would You Do? -
You visit the promotional website of a very popular beverage,
wherein you can find local tasting events and participate in a forum for fellow connoisseurs. You discover that the whole site is riddled with SQL injection and cross-site scripting vulnerabilities. What Would You Do? -
You’re a SaaS-based CRM provider. You help businesses succeed. You
just received an email from a small security consultancy suggesting that they’ve discovered a privilege escalation bug in the web services API of your flagship application. What Would You Do? -
Truth Vulnerability disclosure is... Sexy Helpful Ugly Loathed Destructive Appreciated
Stupid Awesome Heroic Criminal Painful Pleasant Rewarded Penalized -
Truth Vulnerability disclosure is... Sexy Helpful Ugly Loathed Destructive Appreciated
Stupid Awesome Heroic Criminal Painful Pleasant Rewarded Penalized -
Truth Vulnerability disclosure is... probably going to happen whether you
like it or not, so let’s be smarter about handling it. -
Truth Attacking researchers with gag orders, lawsuits, and other threats
doesn’t stop research and it doesn’t stop vulnerability disclosure -
Truth Vulnerability details generally yield corresponding exploits Vulnerability details help
administrators and security pros craft workarounds Exploits help test the effectiveness of compensating controls -
(Ancient) History Multics information disclosure vulnerability -- reported in 1965
• “Accident” led to password file being displayed in the MOTD • From OSVDB vuln #23257: • Multics CTSS on IBM 7094 contains a flaw that may disclose the contents of the password file...when multiple instances of the system text editor were invoked, causing the editor to create temporary files with a constant name...cause the contents of the system CTSS password file to display to any user logging into the system. -
(Modern) History Recognize these? PHOTO BORROWED FROM WIRED PHOTO CC
LICENSED FROM VFLOCK.COM -
(Modern) History Kryptonite bike lock vulnerability • Broken with a
ballpoint pen • Video circulated in 2004 • Kryptonite knew 12 years earlier; did nothing. Medeco M3 vulnerabilities • Picking and “bumping” • Disclosed in 2008 by Mark Tobias -
(Modern) History IMAGE BORROWED FROM ATTRITION.ORG -- I HOPE JERICHO
DOESN’T FIND ME. -
(Modern) History June 12, 2000: rain forest puppy publishes RFPolicy,
forebear to modern responsible full disclosure. Current incarnation says: • Vendor/provider must respond within five (5) days, and maintain contact no less than every five (5) days • Researcher and vendor/provider can agree upon delayed disclosure • Five (5) days w/o initial response from vendor/provider -> full, public disclosure -
(Modern) History October 2001: Microsoft Security Response Center manager, Steve
Culp, makes statements suggesting that full disclosure is like “following a practice that's best described as information anarchy.” On the heels of eEye’s disclosure of a vulnerability that may have benefited Code Red and Nimda authors Community counters that eEye disclosed details at least one month before Code Red emerged -
(Modern) History September 2003: Blackboard issues cease-and-desist to Interz0ne to
stop Billy Hoffman and Virgil Griffith’s presentation on vulnerabilities July 2005: Michael Lynn plans to present on Cisco IOS vulnerabilities at BlackHat; threatened with lawsuit, investigated by the FBI August 2008: MIT students prepare to talk about MBTA vulnerabilities at DEFCON; gag order issued (later lifted) -
(Modern) History September 7, 2009: Laurent Gaffié releases details regarding
remotely exploitable vulnerability in the SRV2.SYS driver in Windows Vista, Windows Server 2008, and Windows 7 Vulnerability initially leads to Blue Screen of Death; as of September 17, remote code execution is possible Microsoft says Gaffié was irresponsible for not notifying them; Gaffié counters, saying MS was irresponsible for bad QA -
Current Challenges BORROWED FROM...I DON’T EVEN REMEMBER WHERE. IMAGE BORROWED
FROM TENABLE. I HOPE RANUM DOESN’T FIND ME, EITHER. -
Current Challenges Black hats, white hats, gurus, and gawkers continue
to say the same things: “Disclosure is evil!” “Disclosure is vital!” “Screw the vendor! Disclose all the way!” “Screw the industry! Keep vulnerabilities for yourself! This happens like clockwork, year in, year out. -
Current Challenges We still can’t agree on disclosure timelines and
details • 5 days? 30 days? 90 days? Partial? Full? Vulnerability details are worth money • ZDI, IDEFENSE VCP, black markets -
Current Challenges Rapid emergence of and shift to web applications
• What do you do when the vulnerability and exploit are one in the same? (Think SQLi and XSS.) Process control and SCADA vulnerabilities • These things often run critical infrastructure -
Nirvana(-ish) Let’s accept that vulnerabilities will get disclosed • Stop
playing the “Bad researcher! Bad!” song Provide a facility for encouraging responsible disclosure Said facility would aim to give mutual benefit to research and vendor/provider -
Nirvana(-ish) Key requirements: Facilitate (or mediate) interaction between researcher and
vendor or service provider Acknowledges researcher's work Provide adequate (namely legal) protection for the security researcher -
Nirvana(-ish) Give guidance for • crafting a sensible timeline and
plan for a solution to the bug or flaw (and keep parties from stalling) • crafting a sensible timeline for public disclosure of the bug or flaw Document vendor-researcher interaction and disclosure process (for others to learn from) -
References/Reading RFPolicy http://www.wiretrip.net/rfp/policy-simple.html Attrition.org Legal Threats page http://attrition.org/errata/legal_threats/ EFF “Gray
Hat Guide” http://www.eff.org/issues/coders/grey-hat-guide EFF Coder’s Rights Project Vulnerability Reporting FAQ http://www.eff.org/issues/coders/vulnerability-reporting-faq Open Source Vulnerability Database http://osvdb.org/