—In this work we study the problem of query privacy in large scale sensor networks. Motivated by a novel trust model in which clients query networks owned by trusted entities but operated by potentially untrusted adversaries, we consider several proposals for protecting the identities of the queried sensors. Our schemes do not rely on the use of public key cryptography nor do they make any assumptions on the topology of the network. Inspired by the data-centric communication model and the collaborative nature of sensor networks, our proposals distribute the data in random locations to be later retrieved by random direction queries, using only local computations and total absence of coordination. It is perhaps of interest to note that query privacy is achieved using only lightweight, symmetric cryptographic primitives. Extensive analytical and experimental results confirm that the proposed protocols can achieve their goals using only minimal communication and storage overhead.