Abstract: Monitoring client traffic at frequencies suitable to infer playback information of applications executing at the client device is described herein. By way of example, network traffic transmitted to a client application can be analyzed to determine an amount of traffic consumed by the client. Based on the amount of traffic and optionally network conditions observed by the client, a suitable resource level can be determined for conveying the traffic to the client. Adjustments to allocated resources can be made, where requested, to improve consumption rates of allocated resources. The resource level can be determined at a frequency suitable to identify changes in playback state of the client application, to provide changes in allocated resource that closely respond to changes in traffic demand.

Abstract: Providing for prioritization of applications operating on a client device with respect to access to online content or services is described herein. By way of example, the prioritizing can be correlated with allocation of network resources for respective applications. An application having higher priority can be allocated a larger amount of resources, a guaranteed amount of resources, a guaranteed quality of service, first access to resources, or the like. Likewise, an application with lower priority can be allocated a lower amount of resources, have best effort resources, access to residual resources, and so forth. In various embodiments, applications can be prioritized based on a status of the application with respect to the client device. As one illustrative example, applications actively receiving content, responding to user commands, or maximized or displayed on a graphic display can be afforded higher priority, whereas inactive or minimized applications can be afforded a lower priority.

Abstract: Monitoring client traffic at frequencies suitable to infer playback information of applications executing at the client device is described herein. By way of example, network traffic transmitted to a client application can be analyzed to determine an amount of traffic consumed by the client. Based on the amount of traffic and optionally network conditions observed by the client, a suitable resource level can be determined for conveying the traffic to the client. Adjustments to allocated resources can be made, where requested, to improve consumption rates of allocated resources. The resource level can be determined at a frequency suitable to identify changes in playback state of the client application, to provide changes in allocated resource that closely respond to changes in traffic demand.

Abstract: Providing for prioritization of applications operating on a client device with respect to access to online content or services is described herein. By way of example, the prioritizing can be correlated with allocation of network resources for respective applications. An application having higher priority can be allocated a larger amount of resources, a guaranteed amount of resources, a guaranteed quality of service, first access to resources, or the like. Likewise, an application with lower priority can be allocated a lower amount of resources, have best effort resources, access to residual resources, and so forth. In various embodiments, applications can be prioritized based on a status of the application with respect to the client device. As one illustrative example, applications actively receiving content, responding to user commands, or maximized or displayed on a graphic display can be afforded higher priority, whereas inactive or minimized applications can be afforded a lower priority.

Abstract: Providing for dynamic allocation of network resources for delivery of network content is described herein. By way of example, network resources can be correlated with real-time network conditions, and monitoring of the network conditions can be employed to track availability of subsets of network resources. Where a subset of network resources allocated to support a particular application is available, the subset of resources can be allocated to one or more client devices, on demand. Where insufficient resource availability exists, resources can be freed up from other subsets of network resources to bolster availability of requested resources to support an application, or a demand for resources can be rejected. Billing for network resources can be in part or in total based on resources consumed, among other dynamic factors, in addition to or in lieu of static factors.