The present application discloses a method for caching data. The method includes: a controller calculates a local caching gain of first content for each forwarding device on the first content delivery path, wherein a local caching gain of the first content of an ith forwarding device is generated by calculation according to a first parameter and a popularity of the first content, and the first parameter is used for identifying bandwidth consumed by transmitting the first content between the ith forwarding device and the server. The controller obtains a first caching gain, wherein the first caching gain is a maximum value of local caching gains of all forwarding devices on the first content delivery path. The controller sends the first caching gain to the server and sends the local caching gain of the first content to the ith forwarding device.