Abstract:
An edge server communicates with a client in a local network. The client displays a remote desktop image representative of a virtual desk image generated by a virtual desktop server (VDS). The edge server receives from the VDS a desktop image to be displayed on the remote desktop. The desktop image includes an incomplete window in which graphic media, such as video, is to be displayed. The edge server determines communication link conditions in the local network and a corresponding transmit data rate at which to transmit data to the client device and a suitable encoding rate at which the graphic media is encoded. The edge server downloads the graphic media from the content server, merges the downloaded graphic media into the incomplete area to produce a completed desktop image, and transmits the completed desktop image at the determined transmit data rate for display on the remote desktop.
Abstract:
A precaching system identifies an object, such as a media file, that a user accesses and then analyzes a social graph of the user to identify social graph contacts that may be interested in the object. Based on the content of the object—and the interests and connections of contacts in the social graph—the precaching system determines whether a particular contact in the user's social graph is likely also to access the object. For example, the precaching system may determine a hit score corresponding to the object and a likelihood that the particular contact in the social graph will access the object. If the precaching system determines that the likelihood that the particular contact will access the object meets or exceeds a threshold probability level for precaching the object, the precaching system precaches the object near the contact in anticipation that the contact will access the object.
Abstract:
A precaching system identifies an object, such as a media file, that a user accesses and then analyzes a social graph of the user to identify social graph contacts that may be interested in the object. Based on the content of the object—and the interests and connections of contacts in the social graph—the precaching system determines whether a particular contact in the user's social graph is likely also to access the object. For example, the precaching system may determine a hit score corresponding to the object and a likelihood that the particular contact in the social graph will access the object. If the precaching system determines that the likelihood that the particular contact will access the object meets or exceeds a threshold probability level for precaching the object, the precaching system precaches the object near the contact in anticipation that the contact will access the object.
Abstract:
An edge server communicates with a client in a local network. The client displays a remote desktop image representative of a virtual desk image generated by a virtual desktop server (VDS). The edge server receives from the VDS a desktop image to be displayed on the remote desktop. The desktop image includes an incomplete window in which graphic media, such as video, is to be displayed. The edge server determines communication link conditions in the local network and a corresponding transmit data rate at which to transmit data to the client device and a suitable encoding rate at which the graphic media is encoded. The edge server downloads the graphic media from the content server, merges the downloaded graphic media into the incomplete area to produce a completed desktop image, and transmits the completed desktop image at the determined transmit data rate for display on the remote desktop.
Abstract:
Weighted fair sharing of video bandwidth may be provided, for example, over TCP. First, a number of connections with a receiver may be opened. The number of active connections may be proportional to a weighting parameter associated with a stream. Next, the stream may be transmitted over the number of active connections and feedback information may be obtained corresponding to the stream received at the receiver. A quality adaptation may then be made based on the obtained feedback. The process may be repeated over time with the number of active connections changing along with the weighting parameters.