Exposing Security Risks For Commercial Mobile Devices (CMDs)

Exposing Security Risks For Commercial Mobile Devices (CMDs)

Exposing Security Risks For Commercial Mobile Devices (CMDs) Jeffrey Voas, PhD, FIEEE [email protected] NIST Maturity of Technologies (source Gartner) CIO Business Priorities Mobile Device Manufacturer Sales Mobile OS trends Enterprise Security

Enterprise Security Enterprise Security Enterprise Security Devices Secure Verify Test Deploy Enterprise Security Device Provisioning

Risks in Mobile Security Supply Chain MDM/Middleware Providers Hardware Components of Interest CPU Display

Graphics Audio Microphone Wi-fi GPS Touchscreen Accelerometer Compass And more Application Vetting: Big Picture Progression of Testing Results An examiner will review the responses from the developer A human evaluation of trust in the responses is made

If questionable, developer may be asked for clarifications or app recommended for re-work If believed, app is ready for app store inclusion or for next two assurance approaches Questions 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16. 17. 18. Does the app that you offer contain any security issues that need to be mitigated or revealed? For example, does the app access phone, network (wifi and/or Bluetooth), and/or GPS resources?, does the app access camera or other recording hardware or software?, does the app access operating system resources or libraries?, and what information, if any, is transferred from the phone to a remote host (including during PC synchronization) by the app?. If so, please explain. If you believe that your app does not contain any security issues, please explain why, and then ignore remaining Questionnaire. If you answer no here, please give detail and good reasoning. If you answer yes please address as many of the following questions as are reasonable for your circumstances. Lacking information may trigger additional review of your application for app approval and thus delay adoption into the appstore. How long has the software source been available? Is there an active user community providing peer review and actively evolving the software? Does the license/contract restrict the licensee from discovering flaws or disclosing details about software

defects or weaknesses with others (e.g., is there a gag rule or limits on sharing information about discovered flaws)? Does software have a positive reputation? Does software have a positive reputation relative to security? Are there reviews that recommend it? What are the processes (e.g., ISO 9000, CMMI, etc.), methods, tools (e.g., IDEs, compilers), techniques, etc. used to produce and transform the software (brief summary response)? What security measurement practices and data does the company use to assist product planning? Describe the training the company offers related to defining security requirements, secure architecture and design, secure coding practices, and security testing. Explain. Are there some requirements for security that are structured as part of general release-ability of a product and others that are as needed or custom for a particular release? What process is utilized by the company to prioritize security-related enhancement requests? What review processes are implemented to ensure that nonfunctional security requirements are unambiguous, traceable and testable throughout the entire Software Development Life Cycle (SDLC)? Are security requirements developed independently of the rest of the requirements engineering activities, or are they integrated into the mainstream requirements activities? Are misuse/abuse cases derived from the application requirements? Are relevant attack patterns used to identify and document potential threats? What threat assumptions were made, if any, when designing protections for the software and information

assets processed? What security design and security architecture documents are prepared as part of the SDLC process? What threat modeling process, if any, is used when designing the software protections? How are confidentiality, availability, and integrity addressed in the software design? What are/were the languages and non-developmental components used to produce the software (brief summary response)? What secure development standards and/or guidelines are provided to developers? Questions (cont.) 19. 20. 21. 22. 23. 24. 25. 26. 27.

28. 29. 30. 31. 32. 33. 34. 35. 36. Are tools provided to help developers verify that the software they have produced has a minimal number of weaknesses that could lead to exploitable vulnerabilities? What are the tools, and how have they been qualified? What is the breadth of common software weaknesses covered (e.g., specific CWEs)? Does the company have formal coding standards for each language in use? If yes, how are they enforced? Does the software development plan include security peer reviews? Does the organization incorporate security risk management activities as part of the software development methodology? If yes, will a copy of the documentation of this methodology be available or information on how to obtain it from a publicly accessible source? Does the softwares exception-handling mechanism prevent all faults from leaving the software, its resources,

and its data (in memory and on disk) in a vulnerable state? Does the exception-handling mechanism provide more than one option for responding to a fault? If so, can the exception-handling options be configured by the administrator or overridden? Does the software validate (e.g., filter with white listing) inputs from potentially un-trusted sources before being used? Is a validation test suite or diagnostic available to validate that the application software is operating correctly and in a secure configuration following installation? Has the software been designed to execute within a constrained execution environment (e.g., virtual machine, sandbox, chroot jail, single-purpose pseudo-user, etc.) and is it designed to isolate and minimize the extent of damage possible by a successful attack? Does the documentation explain how to install, configure, and/or use the software securely? Does it identify options that should not normally be used because they create security weaknesses? Where applicable, does the program use run-time infrastructure defenses (such as address space randomization, stack overflow protection, preventing execution from data memory, and taint checking)? How does the company minimize the risk of reverse engineering of binaries? Are source code obfuscation techniques used? Are legal agreements in place to protect against potential liabilities of non-secure software? Does the software default to requiring the administrator (or user of a single-user software package) to expressly approve the automatic installation of patches/upgrades, downloading of files, execution of plug-ins or other helper applications, and downloading and execution of mobile code? Does the software have any security critical dependencies or need additional controls from other software (e.g.,

operating system, directory service, applications), firmware, or hardware? If yes, please describe. Does the software include content produced by suppliers other than the primary developer? If so, who? What are the policies and procedures for verifying the quality and security of non-developmental components used? What types of functional tests are/were performed on the software during its development (e.g., spot checking, component-level testing, integrated testing)? Who and when are security tests performed on the product? Are tests performed by an internal test team, by an independent third party, or by both? What degree of code coverage does testing provide? Are misuse test cases included to exercise potential abuse scenarios of the software? Questions (cont.) 37. 38. 39. 40. 41. 42.

43. 44. 45. 46. 47. 48. 49. 50. 51. 52. Are security-specific regression tests performed during the development process? If yes, how frequently are the tests performed? Are regression test scripts available? Does the companys defect classification scheme include security categories? During testing what proportion of identified defects relate to security? How has the software been measured/assessed for its resistance to identified, relevant attack patterns? Are Common Vulnerabilities & Exposures (CVEs) or Common Weakness Enumerations (CWEs) used? How have exploitable flaws been tracked and mitigated? Has the software been evaluated against the Common Criteria, FIPS 140-2, or other formal evaluation process?

What evaluation assurance was achieved? If the product claims conformance to a protection profile, which one(s)? Are the security target and evaluation report available? Are static or dynamic software security analysis tools used to identify weaknesses in the software that can lead to exploitable vulnerabilities? If yes, which tools are used? What classes of weaknesses are covered? When in the SDLC (e.g., unit level, subsystem, system, certification and accreditation) are these scans performed? Are SwA experts involved in the analysis of the scan results? Are there current publicly-known vulnerabilities in the software (e.g., an unrepaired CWE entry)? Has the software been certified and accredited? What release/version/configuration? When? By whom? What criteria or scheme was used to evaluate and accredit the software? Is there a Support Life cycle Policy within the organization for the software in question? Does it outline and establish a consistent and predictable support timeline? How will patches and/or Service Packs be distributed to the purchasing/using organization? How extensively are patches and Service Packs tested before they are released? How are reports of defects, vulnerabilities, and security incidents involving the software collected, tracked, and prioritized? What policies and processes does the company use to verify that software components do not contain unintended, dead, or malicious code? What tools are used? How frequently are major versions of the software released? Are configuration/change controls in place to prevent unauthorized modifications or additions to source code

and related documentation? Do these controls detect and report unexpected modifications/additions to source code? Do they aid in rolling back an affected artifact to a pre-modified version? Does the company perform background checks on members of the software development team? If so, are there any additional vetting checks done on people who work on critical application components, such as security? Explain. Please provide company names of all 3rd party entities (foreign and domestic) with whom you, the supplier, contracts software development for this app. Results Dynamic reliability and performance measurement of the product in the lab under assumed operational profiles. Static analysis of the source code using COTS and open source tools that search for programming errors such as buffer overflows. Top 25 Common Programming Weaknesses CWEs) [http://cwe.mitre.org/top25/#ProfileAutomatedManual] Capability to identify Application Bugs and Unwanted Functionality

Common Weakness Enumerations (CWEs) Detects hundreds of CWEs Detects 23 of the top 25 CWEs:

1. CWE-79 Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') 2. CWE-89 Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') 3. CWE-120 Buffer Copy without Checking Size of Input ('Classic Buffer Overflow') 4. CWE-352 Cross-Site Request Forgery (CSRF)

5. CWE-285 Improper Access Control (Authorization) 6. CWE-807 Reliance on Untrusted Inputs in a Security Decision 7. CWE-22 Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') 8. CWE-434 Unrestricted Upload of File with Dangerous Type 9. CWE-78 Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection') 10. CWE-311 Missing Encryption of Sensitive Data 11. CWE-798 Use of Hard-coded Credentials 12. CWE-805 Buffer Access with Incorrect Length Value 13. CWE-98 Improper Control of Filename for Include/Require Statement in PHP Program ('PHP File Inclusion') 14. CWE-129 Improper Validation of Array Index 15. CWE-754 Improper Check for Unusual or Exceptional Conditions 16. CWE-209 Information Exposure Through an Error Message 17. CWE-190 Integer Overflow or Wraparound 18. CWE-131 Incorrect Calculation of Buffer Size 19. CWE-306 Not supported 20. CWE-494 Not supported 21. CWE-732 Incorrect Permission Assignment for Critical Resource 22. CWE-770 Allocation of Resources Without Limits or Throttling 23. CWE-601 URL Redirection to Untrusted Site ('Open Redirect')

24. CWE-327 Use of a Broken or Risky Cryptographic Algorithm 25. CWE-362 Race Condition NISTs National Vulnerability Database (NVD) Free to Public Contains 176 Android Vulnerabilities Results Data collected may include: Amount of time an app is executed Type and amount of data transmitted Feature usage within an app Number of exception calls Benefits include: Usage data that can be used for billing, Reducing number of apps/functionality

Additional app testing Note: Instrumentation can be turned on and off easily, and done selectively as well. Also, instrumentation does incur performance and footprint hits. Application Testing Framework Application Static Analysis does not cover Program Functionality Fortify, Coverity, and other application testing tools cover regular, non-Android specific Bugs: No Security Analysis of the Code Functionality No Power Analysis of the Application components and code No Profiling of the resource consumption of individual applications Cannot Regulate/Deny the access and use of phone subsystems (Camera, Microphone, GPS..) Application Testing Framework

Android Specific Analysis includes analysis of the Application Security Manifest (not supported by third-party vendors) Tailored to the Android Permission Model Verify if the requested permissions are warranted by the submitted code Curtails excessive permissions and enforces a tighter security model Modifications on the Android Engine to enable dynamic policies Control the underlying Dalvik engine to report absence/depletion of resources instead of lack of permissions Regulate access to critical/restricted resources ATP Architecture ATP analyzes Android code bundles and returns

messages, analysis reports, and signed APKs Application Testing Portal Android Application Application Store Analysis /Reports & Signed APKs Android code bundle Developer App Manager Analyses Engine

Request Handler Pre-Processor Registration Handler Tool Invoker Submission Validator Post-Processor UI Handler APK Compiler/Signer API Handler

Result Handler Security assessor examines submissions that do not pass ATP analysis. ATP Repository Security Assessor Mobilize-ATP Workflow (PASS Use-Case) ATP applies Testing to Analyze Android code bundles App Store NIST Testing

Portal (ATP) 1. Submit Android code bundle 2. Register submission 3. Tool 1 analysis 4. Tool 1 status message & analysis report 6. Tool 2 status message & analysis report 8. Tool n status message & analysis report 5. Tool 2 analysis AVs and Testing 7. Tool n

analysis 9. Assess results PASS? 11. PASS message & APK Tools are invoked in parallel on received submissions APKs are generated and signed only if all security analyses pass. 10. Sign APK

24 Application Analysis Report ATP Monitor 26 Application Analysis Results Analyzed ~267,000 Applications from the Google Android Market Many Applications with: Access to Camera/GPS/Microphone Access to Sdcard Data/Contacts Network/Reach-back Functionality (Updates/???) Persistent Presence (Service) Deviation in Functionality from what was included in the Application Description

Some with remote access capabilities Defense in-Depth: Multiple Levels of Security Application Vetting & Testing Device Lock-down and Encryption of ALL Data and Communications Enforcement of Security Policies in the Android Framework Second-level Defenses placed in the Android Linux Kernel Prevent Attacks that bypass Android Security Framework Android has Inherited some (if not all) of the Linux Vulnerabilities Java Native Interface to Linux Libraries a potential Avenue for Exploitation Conclusions Assuring the Secure Operation of Smart

Devices has a wide-range of requirements Application Testing Static & Dynamic In-Field Instrumentation Power Behavior Metering & Policing Physical Device Security Lock-Down of the Device I/O (USB, WiFi, etc.) Encryption of Data both on the Phone & Network Power Metering Framework Design & Implement an accurate model for accounting and policing energy consumption Two-pronged approach Meter the per-process CPU & Device utilization over time Identify the relative impact of each device component on energy

consumption Design an Android kernel subsystem to estimate energy Meter energy consumption for each App/process Use for characterizing application behavior This behavior is Application dependent Sometimes the behavior is also User dependent Power Analysis & Modeling Goals Understand Power Consumption on Mobile Devices

How to accurately model power consumption on a smart-phone? Some tasks are delegated (e.g. wi-fi) There are many hardware subsystems (e.g. input sensors) Accounting for where and for what power is consumed in real-time can help the user make informed decisions Use the Power Profiling to Design Defenses Why is power profiling important? Market study: Up to 75% of total power consumption

spent by applications powering 3rd party advertisements[1] There is incentive for developers and users to use a proper energy accounting infrastructure to make more informed decisions about where to spend remaining device power [1] http://www.bbc.co.uk/news/technology-17431109 Contributions Measurements of each subsystem were performed by leveraging system hooks in Android kernel wake locks driver. Our power model is time-interval independent. Measurements are cumulative and real-time. No pre-defined set of states with fixed power ratings for each, in our model. In depth understanding of how power is consumed in subsystems: cpu, display, wifi, audio

PIXEL POWER ANALYSIS Display Subsystem (1) : Model Display uptime: main wakelock This measures the amount of time that the display was alive Each pixel consumed different amount of energy depending on its color. Display subsystem usage: Udisp = pixel strength x display uptime We calculate pixel strength from a snapshot of the framebuffer every fixed time interval. Display Subsystem (2) : pixel color Rate of Current Discharge (mAh/m) Display Brightness vs. Rate of Current Discharge 10

9.5 9 8.5 8 7.5 7 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 1

black (rgb 0,0,0) white (rgb 255,255,255) red (rgb 255,0,0) green (rgb 0,255,0) blue (rgb 0,0,255) magenta (rgb 255,0,255) cyan (rgb 0,255,255) orange (rgb 255,255,0) 0 10 20 30 40

50 60 Display Brightness (%) 70 80 90 100 Display Subsystem (3) : Pixel Consumption red, green, blue - different impact on current discharge. By regression, we found the relative impact to be: ( 0.4216 : 0.7469 : 1 ) for R : G : B respectively.

Remaining colors can be calculated using RGB values. Lets calculate Orange Vs Blue for 100% brightness: Orange: 0.4216*255 + 0.7469*255 + 1*0 = 297.968 (pixel power) Blue: 0.4216*0 + 0.7469*0 + 1*255 = 255.00 (pixel power) Verification: Comparing these values to area under the curves from the graph, we find the error = 0.87 % i.e. ( 255.00 : 297.968 ) :: ( 165.316 : 191.499 ) Conclusions Our measuring and power consumption analysis approach is accurate and has small power footprint. The results from our regression equations are obtained instantaneously and does not need a lot of CPU time. Power usage attributions to individual applications achieved with a low error of under 1%. Individual subsystems cpu, display, wi-fi, audio were analyzed for their power utilization. Paper in IEEE SERE12 available if requested.

High-Level Project Overview App Developers Banks outpost App Store Vetted apps ultimately go into an app store. Backflows of user feedback and in-field test data. If feedback is good, an app becomes app store accepted, and money is

deposited; otherwise, a new version from the developers needed.

Recently Viewed Presentations

  • Height Mod Focus Group - U.S. National Geodetic Survey

    Height Mod Focus Group - U.S. National Geodetic Survey

    The only practical option to collecting the near-shore ocean gravity data is through airborne based technologies. Dr. Dru A. Smith Chief Geodesist NOAA, National Geodetic Survey [email protected] 1/11/2006 Appendix B. Guidelines guarantee proper ties and provide acceptable relative accuracies. HARN...
  • Regional Policy - Europa

    Regional Policy - Europa

    Speed of (de-) urbanisation per decade Type of urbanisation: large city vs medium-sized Suburbanisation: city vs commuting zone Changes in rural areas, remote areas… Data can also be aggregated to: NUTS 2 and 3 regions, which allows an analysis by...
  • LaTeX - Unistra

    LaTeX - Unistra

    Most of the scientific editors provide the BibTeX entries for their publications, which means that you only need to browse their web sites, find the page of the appropriate article, and copy-paste the entry in your own BibTeX file, then...
  • Great Depression - MIT

    Great Depression - MIT

    Great Depression 21H.102 October 17, 2005 Great Depression 21H.102 October 17, 2005 WPA, Bronx bridge WPA mural art, "Construction of a Dam" (by William Gropper) Federal Theater Project, "One-Third of a Nation," 1938 Marian Anderson at the Lincoln Memorial, Easter...
  • A gyerekek és a digitális pénzek

    A gyerekek és a digitális pénzek

    A digitális pénzek közül a leghíresebb a 2009-ben létrehozott . Bitcoin. digitális, mert nincs fizikai megjelenési formája (bankjegy, érme) alternatív, mert a hagyományos, domináns pénzek alternatívája
  • Reverse Auction (Auction 1001) Initial Commitment Workshop

    Reverse Auction (Auction 1001) Initial Commitment Workshop

    The login process for the clock phase of the reverse auction is the same as the initial commitment login. Station Info As with the Initial Commitment module, bidders can view at any time static information about its stations, including the...
  • Occlusive vascular disorders of the retina

    Occlusive vascular disorders of the retina

    - Activated protein C resistance, lupus anticoagulant, anticardiolipin antibodies, protein C, protein S, antithrombin III . Raised. Intraocular pressure (IOP) Oral contraceptive use . CRVO Clinical presentation . Clinical entities. Non -ischemic CRVO (about 75% of cases)
  • How to Use The Borland C++ Compiler Version 5.02 to create ...

    How to Use The Borland C++ Compiler Version 5.02 to create ...

    How to Use The Borland C++ Compiler Version 5.02 to create C++ and C programs annedawson.com * * Double click on the Borland C++ icon on the desktop: If the icon is not on the desktop, click on the Start...