Security assets in e-voting

Abstract: As discussed in the literature [PrMü01; Rub04; Phi02] e-voting faces a lot of threats. The purpose of this paper is to give a systematically ordered overview of attacks against e-voting and to show one solution to the issues. The challenge is to provide identification and anonymity at the same time and to exclude the possibility of fraudulent manipulations by the server administration, the voter, and any third party. 1 Protocol Issues 1.1 Two-Stage Versus One-Stage Voting Protocols In a fundamental contribution, Nurmi et al. [NSS91] identified two building blocks in an electronic voting system: (i) Voter identification and registration for e-voting and (ii) vote casting. These steps can be provided in one Internet session (one-step protocol); but here the identification may be used to trace the identity of the vote via the IP address or temporary files. This issue is avoided by a two-stage procedure, which strictly separates voter identification and vote-casting. But the advantage comes at a price, as the result of successful identification (voting token) has to be stored at the voter to be used later to cast a vote. Figures 1a and 1b provide an overview of the two stages. Registration phase: The voter applies for a voting token. The system performs a check of his credentials and a check for multiple application. If this is his first attempt, the voter will receive a voting token which he can use anonymously to cast a vote later. If not, the system performs a restart procedure, which always issues the same token to the applicant, which is stored in the database of the regis-tration service. At the end of the process, the voter checks the authenticity and integrity of the token and stores it either on a smart card or on another media, e.g. a USB token