Methods and systems for delivering content to differentiated client devices

摘要

Methods and systems are disclosed for delivery of tailored content to differentiated devices, such as desktop, mobile, and tablet devices, over a computer network. In one embodiment, a proxy cache server has a content cache for storing previously retrieved objects like web pages or multimedia files. For at least some objects, several versions are stored, each version representing an object suited for a given set of client device characteristics. A device-equivalency data structure maintained at the proxy facilitates a determination of whether such cached versions can be used to service a current request. The versions might represent, for example, modified versions created using, e.g., mobile device transcoding techniques, in response to prior requests. They may also represent a set of alternate content created by a content provider and available from an origin server. Such methods and systems may be implemented in a distributed computing networks, e.g., a content delivery network.

主权项

1. A method operable in at least one content server for delivering content to a client over a computer network, the at least one content server having at least one hardware processor and memory holding instructions for execution by the at least one hardware processor, the method comprising:
receiving a first request for first web content from a particular client device, the first web content being identified by a particular universal resource locator (URL); selecting, based at least in part on the particular URL, a particular set of classes from amongst a plurality of such sets, wherein the particular set has at least two classes, and wherein each class in a given set (i) is defined by a plurality of stored characteristics for client devices that the at least one content server matches against a given client device to determine membership in the class and (ii) is associated with a different version of web content, the at least one content server holding data that specifies, on a URL by URL basis, which set of classes to select based on a given URL; selecting a particular class from amongst the at least two classes in the particular set, at least in part because the particular client device has characteristics that match the stored characteristics of the particular class; obtaining the version of the first web content that is associated with the particular class; serving the obtained version of the first web content to the particular client device, in response to the request for first web content; receiving a second request for second web content from a second client device, the second web content being identified by a second universal resource locator (URL); selecting, based at least in part on the second URL, a second set of classes from amongst the plurality of such sets, wherein the second set has a different number of classes than the particular set of classes selected for the particular URL.