Carnegie Mellon University
Browse
Matsumoto_cmu_0041E_10354.pdf (2.5 MB)

Effective and Practical Improvements to the Web Public-Key Infrastructure

Download (2.5 MB)
thesis
posted on 2019-03-06, 18:23 authored by Stephanos MatsumotoStephanos Matsumoto
The Web public-key infrastructure (PKI) provides a mechanism to identify websites to end users for the purposes of encrypted communication. The security of the Web PKI primarily relies on certification authorities (CAs), trusted parties whose misbehavior can enable man-in-the-middle (MITM) attacks: the impersonation of websites to users, followed by the theft or modification of sensitive information. While many methods of addressing CA misbehavior have been proposed, no solution has been both effective and practical: able to protect websites users against C misbehavior and to be easily deployed and used by all parties involved. Thus, despite more than two decades of research advances, the Web PKI remains largely vulnerable to misbehaving CAs.
In this thesis, I argue that we can use minimal changes to existing technology to build deployable solutions that
reduce the rate of successful MITM attacks in the Web PKI. Specifically, I present three projects that exemplify
effective and practical approaches to improving the Web PKI. In IKP, I use the Ethereum cryptocurrency and smart
contract platform to build an insurance-like mechanism that disincentivizes CA misbehavior. In CAPS, I use two
global monitoring and logging systems, CT and Censys, to build a system that strengthens the existing PKI against
misbehaving CAs and enables the secure incremental deployment of new and improved PKIs for theWeb. In SAINT, I use the SCION future Internet architecture to propose a PKI that unifies public-key authentication for naming, routing, and end-entity public keys in a federated environment, and identify challenges and desired properties in such an environment. Through this work, I provide a first step towards making a more resilient Web PKI a reality.

History

Date

2019-02-19

Degree Type

  • Dissertation

Department

  • Electrical and Computer Engineering

Degree Name

  • Doctor of Philosophy (PhD)

Advisor(s)

Bryan Parno

Usage metrics

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC