RFID and the Middleman

Existing bank-card payment systems, such as EMV, have two serious vulnerabilities: the user does not have a trustworthy interface, and the protocols are vulnerable in a number of ways to man-in-themiddle attacks. Moving to RFID payments may, on the one hand, let bank customers use their mobile phones to make payments, which will go a fair way towards fixing the interface problem; on the other hand, protocol vulnerabilities may become worse. By 2011 the NFC vendors hope there will be 500,000,000 NFC-enabled mobile phones in the world. If these devices can act as cards or terminals, can be programmed by their users, and can communicate with each other, then they will provide a platform for deploying all manner of protocol attacks. Designing the security protocols to mitigate such attacks may be difficult. First, it will include most of the hot topics of IT policy over the last ten years (from key escrow through DRM to platform trust and accessory control) as subproblems. Second, the incentives may lead the many players to try to dump the liability on each other, leading to overall system security that is equivalent to the weakest link rather than to sum-of-efforts and