EvoSec 2025W Lecture 12: Difference between revisions

From Soma-notes
 
Line 12: Line 12:


==Class Notes==
==Class Notes==
<pre>
Lecture 12
----------
Plan for the rest of the semester is to read (some) my old papers
- lots to think about trust
Remember that you can and should critique these papers!
G1
- confining JS goes against economic model of the web, maybe?
- discussed what happened with Honey
- rethinking the web as a "wild west"
G2
- new tech defends against injection
- increasing demands for ads, so increasing need for security
  - large platforms must enforce this security
- browser extensions, particularly ad block, why do we trust them more than the ad itself?
G3
- new tech prevents injection but not content-provider conflict, issues is JS environment
- large providers increase risks of conflict
- increased awareness of security issues with ads
G4
- what do we do when legit businesses go bad?
- centralization out-competed regular malware, you're already being spied on!
- can't stop people being taken advantage of through tech
- risk/reward impacts willingness for bad actors to be bad
Terri later did ViSP, SOMA, and SSS (security style sheets)
The real security issue in content-provider conflict is the DOM, not JS per se.
The security solutions that get implemented on the web are the ones that serve the major platforms
- not regular users
page-level isolation mechanisms inhibit ad monitoring and ad security
- avoiding click fraud
</pre>

Latest revision as of 18:06, 25 February 2025

Readings

Discussion Questions

While these are questions to consider, please focus your discussion on what your group finds interesting related to the paper.

  • How do modern security technologies like CORS, CSP, and CORP impact the problems identified in this paper?
  • How does the rise of large platforms impact content-provider conflict?
  • Did this paper change how you thought about the web?

Class Notes

Lecture 12
----------

Plan for the rest of the semester is to read (some) my old papers
 - lots to think about trust

Remember that you can and should critique these papers!

G1
 - confining JS goes against economic model of the web, maybe?
 - discussed what happened with Honey
 - rethinking the web as a "wild west"

G2
 - new tech defends against injection
 - increasing demands for ads, so increasing need for security
   - large platforms must enforce this security
 - browser extensions, particularly ad block, why do we trust them more than the ad itself?

G3
 - new tech prevents injection but not content-provider conflict, issues is JS environment
 - large providers increase risks of conflict
 - increased awareness of security issues with ads

G4
 - what do we do when legit businesses go bad?
 - centralization out-competed regular malware, you're already being spied on!
 - can't stop people being taken advantage of through tech
 - risk/reward impacts willingness for bad actors to be bad



Terri later did ViSP, SOMA, and SSS (security style sheets)

The real security issue in content-provider conflict is the DOM, not JS per se.

The security solutions that get implemented on the web are the ones that serve the major platforms
 - not regular users

page-level isolation mechanisms inhibit ad monitoring and ad security
 - avoiding click fraud