Real Security Threats Ross Anderson Cambridge
Peer-to-peer networks (1) Early P2P proposals included the ‘Eternity Service’ (1996) – a widely distributed file store for censorship resistance Attack (1) – ‘kiddie porn’ Attack (2) – lack of motivation for participants Attack (3) – break the ring, or go for high-order nodes in the courts Attack (4) – spam out the content with trash
Fixes (1) Fix the motivation with a federation of clubs: Danezis/Anderson, ‘Economics of censorship resistance’, WEIS 2004 Or revolutionary cells: Nagaraja/Anderson, ‘Topology of Covert Conflict’, WEIS 2006 Instead of initial authentication, concentrate on recovery (Anderson/Chan/Perrig ‘Key Infection’ ICNP 2004; Anderson/Bond ‘Initial costs and maintenance costs of protocols’, Protocols 2005)
Peer-to-peer networks (2) HomePlug AV – v2 powerline networking Real problem: people connect to wrong network, set up large networks Simple connect mode – push buttons to make it work and ‘send the key in the clear’ Secure mode – type in the device’s AES key (on its label) into a network controller Can a public-key protocol do any more?
Why Homeplug has no PK mode Patent attorney using HomePlug as home LAN Attacker knocks out STB using jammer Net controller says ‘admit Sony STB type ABC123, cert hash = 4CA7 239C 210A 337F?’ Only safe if cert hash checked against label So better off copying key from label directly! See Newman, Gavette, Yonge, Anderson, ‘Protecting Domestic Power-line Communications’, SOUPS 2006
Future P2P networks We already have plenty home and personal networks (HomePlug, Bluetooth, … ) Plenty P2P apps for PCs too (Skype, … ) Sensor network apps? (see our paper later) Phone-based apps? (Haggle?) What are … incentives? … scalability? (Could Khayelitsha cope with success?)