Peer-to-peer networking is a popular topic for network researchers. A fundamental step before digging into understanding peer-to-peer network behavior is to collect a sufficient amount of peer-to-peer traffic traces. However, it is not really easy to obtain traces. Due to privacy concerns, researchers usually have to either sign non-disclosure agreements or choose to use anonymous/obfuscated traces, which may hide important details of the traces. In this paper, we explore the possibility of observing peer-to-peer networks at home. A virtual machine based trace collection system is developed. We then show that with a limited resource, i.e., equipments and network nodes, it is possible to sample a number of peer-to-peer traffic traces and then make further analysis. Based on the traces we collected, a preliminary analysis is also made to show what can be observed in the traces. We believe it is helpful for researchers whom are interested in observing peer-to-peer activities and Internet dynamics.