Admin Project checkpoint #2 due tonight Keep letting us know of any - - PowerPoint PPT Presentation

▶
admin
SMART_READER_LITE
LIVE PREVIEW

Admin Project checkpoint #2 due tonight Keep letting us know of any - - PowerPoint PPT Presentation

Sep 16, 2023 211 likes •847 views

CSE 484 / CSE M 584: Computer Security and Privacy Mobile Platform Security [finish] Fall 2017 Franziska (Franzi) Roesner franzi@cs.washington.edu Thanks to Dan Boneh, Dieter Gollmann, Dan Halperin, Yoshi Kohno, Ada Lerner, John Manferdelli,

slide-1
SLIDE 1

CSE 484 / CSE M 584: Computer Security and Privacy

Mobile Platform Security [finish]

Fall 2017 Franziska (Franzi) Roesner franzi@cs.washington.edu

Thanks to Dan Boneh, Dieter Gollmann, Dan Halperin, Yoshi Kohno, Ada Lerner, John Manferdelli, John Mitchell, Vitaly Shmatikov, Bennet Yee, and many others for sample slides and materials ...

slide-2
SLIDE 2

Admin

  • Project checkpoint #2 due tonight
  • Keep letting us know of any fuzzing issues

with HW3

– Double check that you followed the instructions – Make sure to put things in C: instead of D: – It’s possible to run into issues on the MS side, so let us know and we’ll loop them in if needed (in the meantime you can relax J)

12/1/17 CSE 484 / CSE M 584 - Fall 2017 2

slide-3
SLIDE 3

This Week: Mac OS X High Sierra Issue

  • Given physical access, if root password not set, can login

(creating a root account?) without a password

  • Manual fix: set root password

12/1/17 CSE 484 / CSE M 584 - Fall 2017 3

slide-4
SLIDE 4

Back to Mobile Security: Challenges with Isolated Apps

So mobile platforms isolate applications for security, but…

  • 1. Permissions: How can applications access

sensitive resources?

  • 2. Communication: How can applications

communicate with each other?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 4

slide-5
SLIDE 5

State of the Art

Prompts (time-of-use)

12/1/17 CSE 484 / CSE M 584 - Fall 2017 5

Manifests (install-time)

slide-6
SLIDE 6

Are Manifests Usable?

Do users pay attention to permissions?

[Felt et al.]

… but 88% of users looked at reviews.

12/1/17 CSE 484 / CSE M 584 - Fall 2017 6

slide-7
SLIDE 7

Do users understand the warnings?

Are Manifests Usable?

[Felt et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 7

slide-8
SLIDE 8

Do users act on permission information?

“Have you ever not installed an app because of permissions?”

Are Manifests Usable?

[Felt et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 8

slide-9
SLIDE 9

Android 6.0: Prompts!

  • First-use prompts for sensitive permission (like iOS).
  • Big change! Now app developers need to check for

permissions or catch exceptions.

12/1/17 CSE 484 / CSE M 584 - Fall 2017 9

slide-10
SLIDE 10

Over-Permissioning

  • Android permissions are badly documented.
  • Researchers have mapped APIs à permissions.

www.android-permissions.org (Felt et al.), http://pscout.csl.toronto.edu (Au et al.)

[Felt et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 10

slide-11
SLIDE 11

Improving Permissions: AppFence

[Hornyack et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 11

slide-12
SLIDE 12

Let this application access my location no now. In Insigh ght: A user’s natural UI actions within an application implicitly carry permission-granting semantics.

12/1/17 CSE 484 / CSE M 584 - Fall 2017 12

Improving Permissions: User-Driven Access Control

[our work]

slide-13
SLIDE 13

Let this application access my location no now. In Insigh ght: A user’s natural UI actions within an application implicitly carry permission-granting semantics.

12/1/17 CSE 484 / CSE M 584 - Fall 2017 13

Ou Our r study shows ws: Many users already believe (52% of 186) – and/or desire (68%) – that resource access follows the user-driven access control model.

Improving Permissions: User-Driven Access Control

[our work]

slide-14
SLIDE 14

New OS Primitive: Access Control Gadgets (ACGs)

Approach: Make resource-related UI elements first-class

  • perating system objects (access control gadgets).
  • To receive resource access, applications must embed a

system-provided ACG.

  • ACGs allow the OS to capture the user’s permission

granting intent in application-agnostic way.

12/1/17 CSE 484 / CSE M 584 - Fall 2017 14

slide-15
SLIDE 15

(2) Inter-Process Communication

  • Primary mechanism in Android: Intents

– Sent between application components

  • e.g., with startActivity(intent)

– Explicit: specify component name

  • e.g., com.example.testApp.MainActivity

– Implicit: specify action (e.g., ACTION_VIEW) and/or data (URI and MIME type)

  • Apps specify Intent Filters for their components.

12/1/17 CSE 484 / CSE M 584 - Fall 2017 15

slide-16
SLIDE 16

Unauthorized Intent Receipt

Attack #1: Eavesdropping / Broadcast Theft

12/1/17 CSE 484 / CSE M 584 - Fall 2017 16

com.example.goodapp1 com.example.goodapp2 com.example.badapp

slide-17
SLIDE 17

Unauthorized Intent Receipt

Attack #2: Service/Activity Hijacking

12/1/17 CSE 484 / CSE M 584 - Fall 2017 17

“Caution: To ensure that your app is secure, always use an explicit intent when starting a Servier. Using an implicit intent to start a service is a security hazard because you can't be certain what service will respond to the intent, and the user can't see which service starts.”

slide-18
SLIDE 18

Unauthorized Intent Receipt

  • Attack #1: Eavesdropping / Broadcast Thefts

– Implicit intents make intra-app messages public.

  • Attack #2: Activity Hijacking

– May not always work:

  • Attack #3: Service Hijacking

– Android picks one at random upon conflict!

[Chin et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 18

slide-19
SLIDE 19

Intent Spoofing

  • Attack #1: General intent spoofing

– Receiving implicit intents makes component public. – Allows data injection.

  • Attack #2: System intent spoofing

– Can’t directly spoof, but victim apps often don’t check specific “action” in intent. [Chin et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 19

com.example.goodapp1 com.example.goodapp2 com.example.badapp

slide-20
SLIDE 20

Permission Re-Delegation

  • An application without a permission gains

additional privileges through another application.

  • Demo video
  • Settings application is

deputy: has permissions, and accidentally exposes APIs that use those permissions.

API Settings Demo malware toggleWifi() pressButton(0) Permission System toggleWifi()

[Felt et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 20

slide-21
SLIDE 21

More on Android…

12/1/17 CSE 484 / CSE M 584 - Fall 2017 21

slide-22
SLIDE 22

Incomplete Isolation

12/1/17 CSE 484 / CSE M 584 - Fall 2017 22

Embedded UIs and libraries always run with the host application’s permissions! (No same-origin policy here…)

[Shekhar et al.] Like us on Facebook!

Ad from ad library Social button from Facebook library Map from Google library

slide-23
SLIDE 23

Android Application Signing

  • Apps are signed

– Often with self-signed certificates – Signed application certificate defines which user ID is associated with which applications – Different apps run under different UIDs

  • Shared UID feature

– Shared Application Sandbox possible, where two or more apps signed with same developer key can declare a shared UID in their manifest

12/1/17 CSE 484 / CSE M 584 - Fall 2017 23

slide-24
SLIDE 24

Shared UIDs

  • App 1: Requests GPS / camera access
  • App 2: Requests Network capabilities
  • Generally:

– First app can’t exfiltrate information – Second app can’t exfiltrate anything interesting

  • With Shared UIDs (signed with same private key)

– Permissions are a superset of permissions for each app – App 1 can now exfiltrate; App 2 can now access GPS / camera

12/1/17 CSE 484 / CSE M 584 - Fall 2017 24

slide-25
SLIDE 25

File Permissions

  • Files written by one application cannot be

read by other applications

– Previously, this wasn’t true for files stored on the SD card (world readable!) – Android cracked down on this

  • It is possible to do full file system encryption

– Key = Password/PIN combined with salt, hashed

12/1/17 CSE 484 / CSE M 584 - Fall 2017 25

slide-26
SLIDE 26

Memory Management

  • Address Space Layout Randomization to

randomize addresses on stack

  • Hardware-based No eXecute (NX) to prevent code

execution on stack/heap

  • Stack guard derivative
  • Some defenses against double free bugs (based on

OpenBSD’s dmalloc() function)

  • etc.

[See http://source.android.com/tech/security/index.html]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 26

slide-27
SLIDE 27

Android Fragmentation

  • Many different variants of

Android (unlike iOS)

– Motorola, HTC, Samsung, …

  • Less secure ecosystem

– Inconsistent or incorrect implementations – Slow to propagate kernel updates and new versions

[https://developer.android.com/about/dashbo ards/index.html]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 27

slide-28
SLIDE 28

What about iOS?

  • Apps are sandboxed
  • Encrypted user data

– See recent news…

  • App Store review process is

(maybe) stricter

– But not infallible: e.g., see Wang et al. “Jekyll on iOS: When Benign Apps Become Evil” (USENIX Security 2013)

12/1/17 CSE 484 / CSE M 584 - Fall 2017 28

  • No “sideloading” apps

– Unless you jailbreak

slide-29
SLIDE 29

CSE 484 / CSE M 584: Computer Security and Privacy

Usable Security [start]

Fall 2017 Franziska (Franzi) Roesner franzi@cs.washington.edu

Thanks to Dan Boneh, Dieter Gollmann, Dan Halperin, Yoshi Kohno, Ada Lerner, John Manferdelli, John Mitchell, Vitaly Shmatikov, Bennet Yee, and many others for sample slides and materials ...

slide-30
SLIDE 30

Poor Usability Causes Problems

12/1/17 CSE 484 / CSE M 584 - Fall 2017 30

si.ed u

slide-31
SLIDE 31

Importance in Security

  • Why is usability important?

– People are the critical element of any computer system

  • People are the real reason computers exist in the first

place

– Even if it is possible for a system to protect against an adversary, people may use the system in other, less secure ways

12/1/17 CSE 484 / CSE M 584 - Fall 2017 31

slide-32
SLIDE 32

Usable Security Roadmap

  • 2 case studies

– Phishing – SSL warnings

  • Step back: root causes of usability problems,

and how to address

12/1/17 CSE 484 / CSE M 584 - Fall 2017 32

slide-33
SLIDE 33

Case Study #1: Phishing

  • Design question: How do you help users

avoid falling for phishing sites?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 33

slide-34
SLIDE 34

A Typical Phishing Page

12/1/17 CSE 484 / CSE M 584 - Fall 2017 34

Weird URL http instead of https

slide-35
SLIDE 35

Safe to Type Your Password?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 35

slide-36
SLIDE 36

Safe to Type Your Password?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 36

slide-37
SLIDE 37

Safe to Type Your Password?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 37

slide-38
SLIDE 38

Safe to Type Your Password?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 38

“Picture-in-picture attacks” Trained users are more likely to fall victim to this!

slide-39
SLIDE 39

Experiments at Indiana University

  • Reconstructed the social network by crawling sites

like Facebook, MySpace, LinkedIn and Friendster

  • Sent 921 Indiana University students a spoofed

email that appeared to come from their friend

  • Email redirected to a spoofed site inviting the user

to enter his/her secure university credentials

– Domain name clearly distinct from indiana.edu

  • 72% of students entered their real credentials into

the spoofed site

12/1/17 CSE 484 / CSE M 584 - Fall 2017 39

slide-40
SLIDE 40

More Details

  • Control group: 15 of 94 (16%) entered personal

information

  • Social group: 349 of 487 (72%) entered personal

information

  • 70% of responses within first 12 hours
  • Adversary wins by gaining users’ trust
  • Also: If a site looks “professional”, people likely to

believe that it is legitimate

12/1/17 CSE 484 / CSE M 584 - Fall 2017 40

slide-41
SLIDE 41

Phishing Warnings

12/1/17 CSE 484 / CSE M 584 - Fall 2017 41

Passive (IE) Active (IE) Active (Firefox)

slide-42
SLIDE 42

Are Phishing Warnings Effective?

  • CMU study of 60 users
  • Asked to make eBay and Amazon purchases
  • All were sent phishing messages in addition to the

real purchase confirmations

  • Goal: compare active and passive warnings

12/1/17 CSE 484 / CSE M 584 - Fall 2017 42

[Egelman et al.]

slide-43
SLIDE 43
  • Active warnings significantly more effective

– Passive (IE): 100% clicked, 90% phished – Active (IE): 95% clicked, 45% phished – Active (Firefox): 100% clicked, 0% phished

Active vs. Passive Warnings

Passive (IE) Active (IE) Active (Firefox)

12/1/17 CSE 484 / CSE M 584 - Fall 2017 43

[Egelman et al.]

slide-44
SLIDE 44
  • Some fail to notice warnings entirely

– Passive warning takes a couple of seconds to appear; if user starts typing, his keystrokes dismiss the warning

  • Some saw the warning, closed the window, went

back to email, clicked links again, were presented with the same warnings… repeated 4-5 times

– Conclusion: “website is not working” – Users never bothered to read the warnings, but were still prevented from visiting the phishing site – Active warnings work!

User Response to Warnings

[Egelman et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 44

slide-45
SLIDE 45
  • Don’t trust the warning

– “Since it gave me the option of still proceeding to the website, I figured it couldn’t be that bad”

  • Ignore warning because it’s familiar (IE users)

– “Oh, I always ignore those” – “Looked like warnings I see at work which I know to ignore” – “I thought that the warnings were some usual ones displayed by IE” – “My own PC constantly bombards me with similar messages”

Why Do Users Ignore Warnings?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 45

[Egelman et al.]

slide-46
SLIDE 46

Case Study #2: Browser SSL Warnings

  • Design question 1: How to indicate

encrypted connections to users?

  • Design question 2: How to alert the user if a

site’s SSL certificate is untrusted?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 46

slide-47
SLIDE 47

The Lock Icon

  • Goal: identify secure connection

– SSL/TLS is being used between client and server to protect against active network attacker

  • Lock icon should only be shown when the page is

secure against network attacker

– Semantics subtle and not widely understood by users – Whose certificate is it?? – Problem in user interface design

12/1/17 CSE 484 / CSE M 584 - Fall 2017 47

slide-48
SLIDE 48

Will You Notice?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 48

[Moxie Marlinspike]

Þ

Clever favicon inserted by network attacker

slide-49
SLIDE 49

Do These Indicators Help?

  • “The Emperor’s New Security Indicators”

– http://www.usablesecurity.org/emperor/emperor.pdf

Users don’t notice the absence of indicators!

12/1/17 CSE 484 / CSE M 584 - Fall 2017 49

slide-50
SLIDE 50

Aside (re: Phishing): Site Authentication Image (SiteKey)

12/1/17 CSE 484 / CSE M 584 - Fall 2017 50

If you don’t recognize your personalized SiteKey, don’t enter your Passcode

slide-51
SLIDE 51

Latest Design in Chrome

12/1/17 CSE 484 / CSE M 584 - Fall 2017 51

slide-52
SLIDE 52

Firefox vs. Chrome Warning

33% vs. 70% clickthrough rate

[Felt et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 52

slide-53
SLIDE 53

Experimenting w/ Warning Design

[Felt et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 53

slide-54
SLIDE 54

Experimenting w/ Warning Design

[Felt et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 54

slide-55
SLIDE 55

Experimenting w/ Warning Design

[Felt et al.]

12/1/17 CSE 484 / CSE M 584 - Fall 2017 55

slide-56
SLIDE 56

Experimenting w/ Warning Design

[Felt et al.]

12/1/17 56 CSE 484 / CSE M 584 - Fall 2017

slide-57
SLIDE 57

Experimenting w/ Warning Design

[Felt et al.]

12/1/17 57 CSE 484 / CSE M 584 - Fall 2017

slide-58
SLIDE 58

Opinionated Design Helps!

12/1/17 CSE 484 / CSE M 584 - Fall 2017 58

[Felt et al.]

Adherence N 30.9% 4,551

slide-59
SLIDE 59

Opinionated Design Helps!

12/1/17 CSE 484 / CSE M 584 - Fall 2017 59

Adherence N 30.9% 4,551 32.1% 4,075

[Felt et al.]

Adherence N 30.9% 4,551 32.1% 4,075 58.3% 4,644

slide-60
SLIDE 60

Challenge: Meaningful Warnings

12/1/17 CSE 484 / CSE M 584 - Fall 2017 60

[Felt et al.]

slide-61
SLIDE 61

Stepping Back: Root Causes?

  • Computer systems are complex; users lack intuition
  • Users in charge of managing own devices

– Unlike other complex systems, like healthcare or cars.

  • Hard to gauge risks

– “It won’t happen to me!”

  • Annoying, awkward, difficult
  • Social issues

– Send encrypted emails about lunch?...

12/1/17 CSE 484 / CSE M 584 - Fall 2017 61

slide-62
SLIDE 62

How to Improve?

  • Security education and training
  • Help users build accurate mental models
  • Make security invisible
  • Make security the least-resistance path
  • …?

12/1/17 CSE 484 / CSE M 584 - Fall 2017 62