In this paper we present clustering protocol (CP) for sensor networks. Clustering techniques are used by different protocols and applications to increase scalability and reduce delays in sensor networks. Examples include routing protocols, and applications requiring efficient data aggregation. Our approach which is based on the covering problem that aims at covering an area with minimum possible circular disks assuming ideal conditions. CP is a lightweight protocol that does not require any neighborhood information and imposes lesser communication overhead. We present simulation results to show the efficiency of CP in both ideal cases and randomly distributed networks. Moreover, CP is scalable with respect to density and network size t.