Abstract: Via use of a shared data bus, a processor system offloads processing tasks. For example, a processor system communicates over a respective data bus with a data communication controller. After notifying the data communication controller of a particular block of data to retrieve, the processor system relinquishes control of the respective data bus so that the data communication controller can control the data bus and store a block of data in a specified memory location using direct memory access techniques. Upon receiving a notification of completion of storing the data block by the data communication controller, the processor system regains control of the respective data bus and notifies a data forwarding circuit to: i) partition the block of data into data packets, ii) apply respective headers to the data packets based on the header information received from the processor system, and iii) forward the data packets to the respective destination.

Abstract: An on-demand server configured to dynamically control and optimize memory utilization while minimizing access to disk storage by monitoring content addressable memory (CAM) usage so as to remove the content from resident CAM memory and page the same adaptively and to eliminate creating separate files for trick playback of content streamed to a customer. The method and system dynamically uses the current concurrency count of the contents to optimize memory utilization so as to swap content in and out of memory as the concurrency values change in conjunction with an on-demand memory based server to provided advantages of optimizing the usage of server resources.

Abstract: The present invention provides systems for improved quality of service and traffic management in network routers and other devices. This is achieved, according to one aspect of the invention, by coupling a plurality of queue processors to a plurality of input interfaces that receive data from one or more respective network connections. Each queue processor, in coordination with an associated scheduler that schedules dequeing of data from one or more queues, maintains the quality of service levels with respect to throughput, and delivers the data for a particular output context based on priority to a respective output interface. The output interface is coupled to the plurality of queue processors and transfers the data to one or more respective network connections. In addition, a plurality of output interfaces can be coupled to respective pluralities of queue processors for transferring data therefrom to one or more destination network connections.