Abstract: Systems and methods, consistent with the present invention, provide a high-speed line interface for networking devices. Such an interface may be used in networking devices, such as routers and switches, for receiving data from, and transmitting data to, high-speed links, such as those lines carrying data at rates of 2.5 Gbit/sec, 10 Gbit/sec, and 40 Gbit/sec and more. In a preferred embodiment, the interface deserializes data from an incoming data stream onto a multi-line bus so that the data may be processed at a lower clock speed. Packets are extracted from the data on the multi-line bus and distributed among a plurality of switching/forwarding modules for processing.

Abstract: A bandwidth divider and method for allocating bandwidth between a plurality of packet processors. The bandwidth divider includes a plurality of counters for measuring the bandwidth of data packets transferred from the bandwidth divider to a respective packet processor; and a controller for analyzing the plurality of counters and transferring a data packet to a selected packet processor based on the contents of the counters. The method monitors the bandwidth consumed by the packet processors; determines, based on the bandwidth consumed by the packet processors, which packet processor has consumed the least amount of bandwidth; and allocates a next data packet to the packet processor which has consumed the least amount of bandwidth.

Abstract: A network device includes one or more sprayers, multiple packet processors, and one or more desprayers. The sprayers receive packets on at least one incoming packet stream and distribute the packets according to a load balancing scheme that balances the number of bytes of packet data that is given to each of the packet processors. The packet processors receive the packets from the sprayers and process the packets to determine routing information for the packets. The desprayers receive the processed packets from the packet processors and transmit the packets on at least one outgoing packet stream based on the routing information.

Abstract: A two wire serial bus is connected between different circuit boards in a complex electrical system. The two wire serial bus may be used to receive status information about each of the circuit boards in the system. A master control processor on one of the circuit boards controls which of the other circuit boards are active on the serial bus. Each of the non-master circuit boards includes a series of switches that electrically isolate or connect portions of the two wire serial bus from one another. Through the series of switches, both the master control processor and processors local to each of the other circuit boards may simultaneously access different portions of the serial bus.

Abstract: A network device includes redundant buses, redundant master controllers, and slave controllers. Each of the master controllers connects to a corresponding one of the buses. One of the master controllers acts as an active master and the other master controllers act as standby masters. The active master commences a bus cycle that includes an address interval and a data interval, provides a destination address on the corresponding bus during the address interval, and transmits or receives a command or data during the data interval. The slave controllers connect to the bus, detect commencement of the bus cycle, sample the destination address from the bus a predetermined amount of time after commencement of the address interval, and transmit or receive a command or data during the data interval.

Abstract: A debugging and diagnostic system allows a developer to receive low-level diagnostic information from multiple processors in a complex electrical system. A bus connects a master processor to the processors to be debugged via corresponding receiver/driver circuits. The receiver/driver circuits receive serial information from the processors and transmit it to the bus. The master processor controls the receiver/driver circuits through a control logic circuit.

Abstract: A system prevents processor faults during the configuring of a programmable device. The system sets the processor into a reset mode, configures the programmable device using first configuration data, and determines whether the configuring using the first configuration data is successful. When the programmable device is not successfully configured, the system configures the programmable device using second configuration data and determines whether the configuring using the second configuration data is successful. The system releases the processor from the reset mode when the programmable device is successfully configured using the first or second configuration data or when the configuration of the programmable device using the second configuration data fails.

Abstract: A traceless midplane contains substantially no traces, pins, or active components and includes a front portion and a back portion. The front portion includes first connectors. The back portion includes second connectors arranged in a grid pattern. Each of the second connectors includes electrically-conductive conduits that connect the second connector to a corresponding one of the first connectors through the midplane. The second connectors include data connection points, ground connection points, and clock connection points. At least some of the data connection points are separated from each other and from the clock connection points by the ground connection points.

Abstract: The invention provides forced-air cooling to components mounted on circuit boards oriented in a side-to-side direction in a system. Airflow may enter and exit the system through the front and back (or vice-versa), rather than the sides of the system. In one embodiment, airflow entering the front of the system is re-directed in an upward direction, then split to form airflow branches traversing in a side-to-side direction. The airflow branches traverse across the surfaces of circuit boards, then are directed in an upward direction and out the back (or front) of the system. The airflow branches preferably move substantially the same volume of air per unit of time.