Information-Centric Networking (ICN) is a promising future network architecture where content instead of its location becomes the core of the communication model. Content is independent from location, which enables in-network caching and sharing. Consequently, the transport mode changes dramatically. In this paper, we develop an analytical model for content transfer performance in ICN. We derive expressions for the average content delivery time as functions of bandwidth/cache/buffer capacity, propagation delay, content popularity and other factors, and then we specifically analyze the typical linear topology and the tree topology. Our analytical results, supported by simulations, can be used as a reference for the design and evaluation of ICN architectures and protocols.