Security Protocols with Isotropic Channels

We investigate the security properties of isotropic channels, broadcast media in which a receiver cannot reliably determine whether a message originated from any particular sender and a sender cannot reliably direct a message away from any particular receiver. We show that perfect isotropism implies perfect (informationtheoretic) secrecy, and that asymptotically close to perfect secrecy can be achieved on any channel that provides some (bounded) uncertainty as to sender identity. We give isotropic security protocols under both passive and active adversary models, and discuss the practicality of realizing isotropic channels over various media.