Usability and security of out-of-band channels in secure device pairing protocols

  • Authors:
  • Ronald Kainda;Ivan Flechais;A. W. Roscoe

  • Affiliations:
  • Oxford University Computing Laboratory, UK;Oxford University Computing Laboratory, UK;Oxford University Computing Laboratory, UK

  • Venue:
  • Proceedings of the 5th Symposium on Usable Privacy and Security
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Initiating and bootstrapping secure, yet low-cost, ad-hoc transactions is an important challenge that needs to be overcome if the promise of mobile and pervasive computing is to be fulfilled. For example, mobile payment applications would benefit from the ability to pair devices securely without resorting to conventional mechanisms such as shared secrets, a Public Key Infrastructure (PKI), or trusted third parties. A number of methods have been proposed for doing this based on the use of a secondary out-of-band (OOB) channel that either authenticates information passed over the normal communication channel or otherwise establishes an authenticated shared secret which can be used for subsequent secure communication. A key element of the success of these methods is dependent on the performance and effectiveness of the OOB channel, which usually depends on people performing certain critical tasks correctly. In this paper, we present the results of a comparative usability study on methods that propose using humans to implement the OOB channel and argue that most of these proposals fail to take into account factors that may seriously harm the security and usability of a protocol. Our work builds on previous research in the usability of pairing methods and the accompanying recommendations for designing user interfaces that minimise human mistakes. Our findings show that the traditional methods of comparing and typing short strings into mobile devices are still preferable despite claims that new methods are more usable and secure, and that user interface design alone is not sufficient in mitigating human mistakes in OOB channels.