We consider truthful online auctions that aim at optimizing sellers? revenues, representing service contributors? satisfactions, as a general model for pricing peer-to-peer services under the assumption of individual service consumer?s rationality. For services that are in unlimited supply, we design a randomized truthful online auction with guaranteed revenue based on a randomized truthful offline auction. It is shown that the expected revenue extracted by our truthful online auction over all random factors achieves a Θ(1) approximation ratio relative to the optimal single-price revenue under some reasonable assumption about the input bids. Since a peer must serve others to earn sufficient revenue that can cover its payment for being served, we argue that our online truthful auctions can be suitable schemes for incentivizing peer nodes in peer-to-peer systems to share, and thereby addressing the "free-rider" problem in peer-to-peer service sharing.