CYBERCRIME & VULNERABILITY ISSUES: WHAT EMERGENCY MANAGERS NEED TO KNOW Jim Duncan, Security Engineer Juniper Networks, Secure Development Lifecycle Initiative.

CYBERCRIME & VULNERABILITY ISSUES: WHAT EMERGENCY MANAGERS NEED TO KNOWJim Duncan, Security EngineerJuniper Networks, Secure Development Lifecycle InitiativeTuesday, 2014-10-14North Carolina Emergency Management Association Fall Conference

2 Copyright © 2014 Juniper Networks, Inc. www.juniper.net

BRIEF BIOGRAPHY

Subject-Matter Expert on software vulnerabilities

Currently working on prevention of flaws (Juniper SDL program)

Previously, product-security and cyber-security incident responder Juniper, BB&T, Cisco, Penn State University, Old Dominion University TRANSITS Instructor – helping National CSIRTs in emerging economies Participant in multiple IRT and cybercrime-fighting forums (FIRST, ICASI) Critical Infrastructure Protection evangelist (NIAC VDF, CVSS, ISACs) Ideal candidate for the exit row!

(Also soccer referee, parliamentarian, piano technician, pistolsmith, etc., can’t keep a job!)


WHY AM I HERE TODAY? Cybercrime and vulnerabilities are here to stay

You know that already; this will not be yet another trend report

Technology Complexity – the Internet of Things – growing without bound

Implications for interactions with other disciplines both exciting and scary Security, if any, is frequently low priority, or omitted from consideration entirely Definitely no security in Version 0.1, which is what is deployed to first responders!

Technology is just a tool, you should not need to be an expert in it How many of you are well-versed in internal combustion?

My goal is impart observations, rules, etc., for thinking about cyber systems and understanding the larger threats and countermeasures


PROBLEMS


NATURE OF CYBERCRIME

“Cybercriminals are business people, too.” Amazing parallels to counterfeiting of old: front office, back office, etc. Well-financed, distributed, smart, not greedy (mostly) Misalignment of cultural expectations is a complicating factor

Definitions of “crimes” vary from place to place, hard to get support sometimes Resourceful: example of CAPTCHA workaround Well-researched: example of bank phishing aimed at small church officials Follow the money and/or spirit: motivations explain a lot All of the above apply to nation-state and populist activities, too


CONFIDENTIALITY/PRIVACY/REPUTATIONAL THREATS

SWATting and EAS hijacks; not much help here except the obvious

D0Xing of staff and officials – Internet-based embarrassment is deadly Teach staff how to protect themselves online if you expect them to protect

other people’s stuff online. Consider reputation-monitoring services

Monitor and prevent exfiltration of data in your stewardship Don’t assume data was erased – it can never be completely erased Use full-disk encryption and test it Consider reputation-monitoring services for this as well


TELECOMMUNICATIONS THREATS

DoS can’t be prevented, but often it can be managed Various services for ensuring against DoS or mitigating once underway Work with your ISP (maybe more than one ISP) Make sure you have experts involved

Telephony DoS is old, but new again Multiple efforts in multiple countries to improve technological response “Honeypots” deployed to look for TDoS, do-not-call violations, other errors VoIP is exciting, isn’t it? Yeehaw! Fundamental flaw: circuit-switched v. packet-switched security models


TRANSPORTATION THREATS

GPS spoofing and jamming Documented that thieves are using spoofing to hide stolen vehicles Florida(?) motorist operated a cellphone jammer from his car during his

daily commutes to force other motorists to put down their cellphones Easy to imagine similar stunts to fraudulently redirect consumers away from

competitors’ gas pumps or (pick a retail industry)

How do you know your GPS is receiving correct data? Anyone?

Highway sign hijacks are clever, but what if they are subtle? Instead of “zombie” alerts, consider believable “Detour via…” instructions


ENVIRONMENTAL THREATS

EMP and solar flares Really naïve in this area, despite decades of study Recent work very revealing and alarming, but seems to be ignored

Structural HVAC, building & power controls, SCADA systems Never underestimate the potential for someone to inadvertently connect

these systems to something they shouldn’t be connected to And never underestimate the ability of criminals to find them (e.g., Target) What do you do when your EOC gets too hot? Too cold? Too wet? Dry?

Example of first World Trade Center attack in the early 1990s


CASE STUDY NOT YET PUBLISHED

Analysis of pre-hospital information system used by EMTs It was in another state, not North Carolina. Relax! Resume breathing…

Criminals’ delight: No AUP, no password policies Ruggedized laptop running unpatched XP, plans to upgrade to Win7 No full-disk encryption Brand-name software vendor truly did not keep PII on device, but… Helpful cache was uncovered, unencrypted, with PII for 13,500 patients!

THIS HAPPENED IN 2014!!! This is all too believable, unfortunately


SOLUTIONS


OCCAM’S RAZOR, HANLON’S RAZOR, ETC.

Occam’s Razor: “When considering multiple possible causes, select the cause requiring the least complexity”

Not guaranteed to be correct, but likely

Hanlon’s Razor: “Never attribute to malice that which is adequately explained by stupidity.”

Duncan’s Corollary: “Never attribute to an attack that which is adequately explained by negligence.”

“Negligence” can be misconfiguration, software flaw, or user error Example of inadvertent internally-sourced “attack”


AVOID BYSTANDER EFFECT/DIFFUSION OF DUTY

First responders would never do this in the real world, but they fall prey to it in the cyber world: Don’t assume someone else will respond!

Ask questions. Lots of questions. Recipients of questions: be professional and answer appropriately Consider documenting individual findings in “security observation reports”

Advocate for proper brainstorming practices In the first round, get the ideas out there; no vetting whatsoever Second round, go back and evaluate the first-round responses Disciplined facilitator is sometimes needed for this to be effective


REPLACE BLACKLISTING WITH WHITELISTING

Blacklisting: “That, which is not expressly denied, is permitted.” Far too many systems start out this way Painful to go back and close up unnecessary ports/services/features

Whitelisting: “That, which is not expressly permitted, is denied.” Much safer Start with all services disabled, then enable only those that are needed

Example: Instead of allowing browsing everywhere, and then blocking access to a few pages, block all pages except for a selected few.


GET SMART AND STAY SMART ON CRYPTO

“Gosh, crypto is hard!”

Doesn’t have to be difficult to understand the basics Key length is important: long, but not too long (time is an issue, too) Key space should be as large as possible (or reasonably pragmatic) Don’t keep plaintext and encrypted text around, close by Repetition means something failed; algorithm selection is important Watch out for so-called “security improvement trade-offs”

Example of password-typing alternate-left-right scheme (“key space”, above)

Full-disk encryption is worth mentioning again, at this point


UNDERSTAND SPHERE OF ACTION

Expectations and assumptions creep into our thought processes, distort our reasoning, and cause us to produce incorrect results

Cyber threats are global but not the typical disasters most of you handle Example of NRP and Lori Bush, “There’s the hurricane/forest fire/flood!”

Cultural & linguistic differences affect results Example of CAPTCHA workaround, earlier Mismatch of importance regarding Asia/Pacific “loss of face” Example of encipher/decipher v. encrypt/decrypt Time and date formats (ISO-8601), ICS phonetic alphabet


POLICIES AND PROCEDURES

No excuses for not having Acceptable Use Policies, Password Policies, Data Retention Policies, and so forth; write’em down, publicize them

Don’t expect staff to pick good password management schemes; research apps, make recommendations (working group for NCEMA?)

Consider implementing two-factor access schemes

Remember that policies and guidelines should be viewed primarily as tools for education; enforcement comes only when education fails


FIGURE OUT WHAT HAPPENED LATER

“Accountability is the price of openness.” [Daniel E. Geer, Sc.D.]

No one builds a perfect system, so institute appropriate logging and auditing mechanisms so that after something goes wrong, you can backtrack to figure out what happened

Study Ken Thompson’s “Reflections on Trusting Trust” 1984 ACM Turing Award lecture Brilliant, short (3 pages) explanation on how all systems are flawed

because humans are involved, and cannot be separated Completely destroys the “Many eyes makes good security” argument


DON’T ATTEMPT TO BUILD PERFECT SYSTEMS

“The perfect is the enemy of the good enough” (or something like that)

Lots of unnecessary effort is expended on lofty conceptions of the really cool and awesomely beautiful solution to a basic problem

Don’t build seamless systems, especially in an emergency

“Make them seamful, but with beautiful seams.” [Mark Weiser] Example from ruggedized telecom-in-a-box in Hurricane Katrina


BE PART OF THE SOLUTION, NOT THE PRECIPITATE

Encourage proper brainstorming Need sector-specific experts like you to think up interesting problems We don’t know the stuff that you don’t even know you already know

Roll up the results into tabletop exercises

Collaborate with cybersecurity incident responders We both learn from each other We can help with cross-sector exercises We’ll know who to call when we find something important


ANYTHING ELSE?

Q&A

Contact Information:

Jim Duncan, [email protected], +1 919-608-0748

mailto:[email protected]

CYBERCRIME & VULNERABILITY ISSUES: WHAT EMERGENCY MANAGERS NEED TO KNOW Jim Duncan, Security Engineer Juniper Networks, Secure Development Lifecycle Initiative.

Documents

cyber threats

internet of things

larger threats

deadlyteach staff

obviousd0xing of staff

exfiltration of data

cyber systems

managedvarious services