Brocade Communications Systems, Inc.

Nishan Systems, Inc.

Advanced Micro Devices Corporation

Avago Technologies International Sales Pte. Limited

Anil Mehta

Stuart F Oberman

Rodney Norman Mullendore

Kamran Malik

Keith R Schakel

Scott Kipp

Amos Ben-Meir

Reading G Maley

Jeffrey A Slavick

Link speed negotiation for 64 Gbps is done at 32 Gbps to allow only two speeds to be used during link state negotiation. The desire for 64 Gbps operation is indicated in a field shared during link state negotiation. After link speed negotiation is completed at 32 Gbps, a determination is made whether 32 or 64 Gbps operation is desired. If 32 Gbps operation is desired, procedures continue as in the prior operations. If 64 Gbps operation is desired, a new procedure is performed. The new procedure provides time for the optical transceiver to changeover from the PAM2 (pulse amplitude modulation) or binary operation used in 32 Gbps operation to the PAM4 multi-level operation used in 64 Gbps operation. After determining that the optical transceiver is ready to transmit, transmitter training is performed, with increased handshaking to provide improved granularity. After transmitter training is complete, conventional link initialization is performed.

64 gigabit fibre channel speed negotiation and transmitter training

The PCS and FEC layers are combined into a single layer and the number of lanes is set at four lanes. The combination allows removal of many modules as compared to a serial arrangement of a PCS layer and an FEC layer. The reduction in the number of lanes, as compared to 100 Gbps Ethernet, provides a further simplification or cost reduction by further reducing the needed gates of an ASIC to perform the functions. Changing the lanes in the FEC layer necessitates changing the alignment marker structure. In the preferred embodiment a lane zero marker is used as the first alignment marker in each lane to allow rapid sync. A second alignment marker indicating the particular lane follows the first alignment marker.

128 gigabit fibre channel physical architecture

The FC port state machine enhanced by determining if various lanes are configured to be operated in 128 Gbps mode by operating as parallel lanes as indicated by using a reserved bit in a link training field. If so and if all of the ports are 32 Gbps ports and pass training, then four lanes can be combined to form a 128 Gbps link. If the ports are configured for 128 Gbps only operation and at least one lane does not negotiate to 32 Gbps or fails training, the link is not activated and none of the lanes are activated. If the ports are configured to do either 128 Gbps or independent operation and at least one lane cannot operate at 128 Gbps, then the lanes operate independently at the negotiated and trained speed. If the lanes are configured for only independent operation the transceiver develops independent links as negotiated.

128 gigabit fibre channel speed negotiation

128 Gigabit fibre channel physical architecture

A system and method for converting low-jitter, interleaved frame traffic, such as that generated in an IP network, to high jitter traffic to improve the utilization of bandwidth on arbitrated loops such as Fibre Channel Arbitrated Loops. Embodiments of a high jitter scheduling algorithm may be used in devices such as network switches that interface an arbitrated loop with an IP network that carries low-jitter traffic. The high jitter algorithm may use a separate queue for each device on the arbitrated loop, or alternatively may use one queue for two or more devices. Incoming frames are distributed among the queues based upon each frame's destination device. The scheduling algorithm may then service the queues and forward queued frames to the devices from the queues. In one embodiment, the queues are serviced in a round-robin fashion. In one embodiment, each queue may be serviced for a programmed limit.

High jitter scheduling of interleaved frames in an arbitrated loop

A system and method for late-dropping packets in a network switch. A network switch may include multiple input ports, multiple output ports, and a shared random access memory coupled to the input ports and output ports by data transport logic. Packets entering the switch may be subject to input thresholding, and may be assigned to a flow within a group. A portion of a packet subject to input thresholding may be accepted into the switch and assigned to a group and flow even if, at the time of arrival of the portion, there are not enough resources available to receive the remainder of the packet. This partial receipt of the packet is allowed because of the possibility of additional resources becoming available between the time of receipt of and resource allocation for the portion of the packet and receipt of subsequent portions of the packet.

System and method for late-dropping packets in a network switch

A system and method for managing the allocation of Time Division Multiplexing (TDM) timeslots in a network switch. The network switch may use a TDM cycle comprising multiple timeslots to manage shared resources and to schedule data ingress and egress through the ports of the current configuration, wherein each port is assigned one or more timeslots. The network switch may be reprogrammed to support one of multiple timeslot assignment schemes for one of multiple port configurations. The network switch may support configurations with varying numbers of ports, e.g. 8- and 16-port configurations. A network switch may also support configurations where two or more ports are combined to form one port, for example, a 2 Gbs Fibre Channel port. To meet the requirements of the various configurations, the timeslot assignment scheme may be reprogrammed to meet the scheduling requirements of each of the possible port configurations.

Method and system for managing time division multiplexing (TDM) timeslots in a network switch

A system and method for input thresholding packets in a network switch. A network switch may include multiple input ports, multiple output ports, and a shared random access memory coupled to the input ports and output ports by data transport logic. Packets entering the network switch may be assigned to one of a plurality of threshold groups and to one of a plurality of flows within the threshold group. In one embodiment, each threshold group may be divided into a plurality of levels of operation. As resources are allocated or freed by the threshold group, the threshold group may dynamically move up or down in the levels of operation. Within each level, one or more different values may be used as level boundaries and resource limits for flows within the threshold group. In one embodiment, programmable registers may be used to store these values.

Packet input thresholding for resource distribution in a network switch

A system and method for enabling a network switch to transmit queued packets to a device when opened by the device, and thus to utilize the Fibre Channel Arbitrated Loop (FC-AL) in full-duplex mode when possible. The switch may include a plurality of queues each associated with a device on the FC-AL for queuing incoming packets for the device. The switch may determine a next non-empty queue, open the device associated with the queue, and send packets to the device. The device may send packets to the switch concurrently with receiving packets from the switch, thus utilizing the FC-AL in full-duplex mode. When a device opens the switch to transmit packets to the switch, the switch may determine if there are packets for the device in the queue and, if so, send packets to the device concurrently with receiving packets from the device, thus utilizing the FC-AL in full-duplex mode.

Method and apparatus for scheduling packet flow on a fibre channel arbitrated loop

A system and method for low latency switching of data packets in a network switch. A network switch may include multiple input ports, multiple output ports, and a shared random access memory coupled to the input ports and output ports by data transport logic. Under normal operation, the data transport logic stores packet data into the memory. Later, the packet data is read from the memory and output to a destination output port. To reduce latency when the switch is not congested, the switching logic may be configured to perform a cut-through operation by routing packets directly from input ports to output ports without storing any portion of the packet in the memory. Alternatively, the switch may begin forwarding the stored packet data to the output port before the entire packet has been received or stored in the memory.

Dynamic selection of lowest latency path in a network switch

A processor includes a system bus interface that permits short set-up and hold times for bus signals including loop-back signals. Loop-back signals are transferred from an input cell in the interface to a target I/O cell in the interface without resynchronizing the loop-back signal with the processor clock. Accordingly, set-up and hold times for the loop-back signal need only be sufficient to allow for jitter or uncompensated delay in the bus clock signal at the target I/O cell. The processing core provides valid signals that might be required for generating an output signal from the target cell. The core avoids changing those signals near triggering edges of the bus clock signal to prevent the signals from changing before the target I/O cell uses the required signals. Typically, the loop-back signal determines whether I/O cell is enabled for output and is also used at the edge of the bus clock signal.

Processor with short set-up and hold times for bus signals

A processor includes a processing core that is operable at a frequency that is an odd half-integer multiple of a bus clock frequency. Signals on a system bus are synchronized with a selected edge, e.g., the rising edge, of a bus clock signal, but the processing core requires signals synchronized with a processor clock signal. Signal crossing between the clock domain of the processing core and the clock domain of the system bus pass through a storage element that selectably latches a value of the signal either at a rising edge or a falling edge of the processor clock signal. A control circuit selects either rising-edge or falling-edge latching depending on which edge (rising or falling) is closest to being synchronized with the selected edge of the bus clock signal.

Growing community of inventors

Milpitas, CA, United States of America

Anil Mehta