An essential component of a botnet is the Command and Control (C2) channel (a network). The mechanics of C2 establishment often involve the use of structured overlay techniques which create a scaffolding for sophisticated coordinated activities. However, it can also be used as a point of detection because of their distinct communication patterns. Achieving this is a needle-in-a-haystack search problem across distributed vantage points. The search technique must be efficient given the high traffic throughput of modern core routers. In this paper, we focus on efficient algorithms for C2 channel detection. Experimental results on real Internet traffic traces from an ISP’s backbone network indicate that our techniques, (i) have time complexity linear in the volume of traffic, (ii) have high F-measure, and (iii) are robust to the partial visibility arising from partial deployment of monitoring systems, and measurement inaccuracies arising from partial visibility and dynamics of background traffic.