Barefoot Networks, Inc.

Telefonaktiebolaget Lm Ericsson (publ)

Michael Feng

Anurag Agrawal

Yi Li

Edward Ho

Remy Chang

Changhoon Kim

Jeongkeun Lee

Edmund C Chen

Thomas C Yip

Patrick Wang

Robert G Hathaway

Ranjit Joseph Rozario

Brian Alleyne

Srivathsa Dhruvanarayan

Sun Den Chen

Stephen Chow

Yan Wang

Stephan G Meier

Ramanathan Lakshmikanthan

Edmund G Chen

Sun-den Chen

Sunden Chen

Robert Li

Peter Klikovits

Jayaram Beladakere

Jeffrey Hu

Srivi Dhruvanarayan

Some embodiments provide a method for an ingress packet processing pipeline of a network forwarding integrated circuit (IC). The ingress packet processing pipeline is for receiving packets from a port of the network forwarding IC and processing the packets to assign different packets to different queues of a traffic management unit of the network forwarding IC. The method receives state data from the traffic management unit. The method stores the state data in a stateful table. The method assigns a particular packet to a particular queue based on the state data received from the traffic management unit and stored in the stateful table.

Using stateful traffic management data to perform packet processing

In a method for allocating physical queues of a network forwarding element, a request is received at the network forwarding element, the network forwarding element including a plurality of physical queues, where each physical queue of the plurality of physical queues has a fixed bandwidth, the request identifying an allocation of a plurality of virtual queues at the network forwarding element. Based at least in part on the request, a configuration of the plurality of physical queues to the plurality of virtual queues is determined. The plurality of physical queues is configured according to the configuration, wherein the configuring includes allocating at least two physical queues to a virtual queue.

Allocation of virtual queues of a network forwarding element

Some embodiments provide a method for a traffic management circuit of a data plane forwarding circuit. The traffic management circuit receives data messages from a set of ingress pipelines and provides the data messages to a set of egress pipelines. The method identifies a flow control event. The method provides metadata regarding the flow control event to a message generation circuit of the data plane forwarding circuit via a bus between the traffic management circuit and the message generation circuit.

Flow control visibility

Some embodiments provide a method for a hardware forwarding element that includes multiple queues. The method receives a packet at a multi-stage processing pipeline of the hardware forwarding element. The method determines, at one of the stages of the processing pipeline, to modify a setting of a particular one of the queues. The method stores an identifier for the particular queue and instructions to modify the queue setting with data passed through the processing pipeline for the packet. The stored information is subsequently used by the hardware forwarding element to modify the queue setting.

Queue scheduler control via packet data

Some embodiments provide a method for a hardware forwarding element. The method receives a packet to add to a buffer. The packet is assigned a packet class. The method determines an amount of buffer space available for the assigned packet class. Different packet classes have different amounts of buffer space available in the buffer. When the available buffer space for the assigned packet class is large enough for the received packet, the method adds the packet to the buffer.

Buffer space availability for different packet classes

Some embodiments provide a method for a traffic management unit of a network forwarding integrated circuit (IC). The traffic management unit includes multiple queues for storing packets. Each stored packet is (i) received by the traffic management unit from one of multiple ingress packet processing pipelines and (ii) for processing by an egress packet processing pipeline after being released from the queue storing the packet. The method determines that a particular one of the queues has crossed a threshold amount of stored packet data. The method provides queue state data including an identifier of the particular queue and a current amount of data stored in the particular queue to at least a subset of the multiple ingress pipelines. The ingress pipelines use the provided data to process subsequent packets.

Transmission of traffic management data to processing pipeline

Some embodiments provide a network forwarding integrated circuit (IC) including multiple configurable ingress pipelines, multiple configurable egress pipelines, a traffic management unit, and a statistics bus. The configurable ingress pipelines are for processing packets received from ports of the network forwarding IC. The configurable egress pipelines are for processing packets to be transmitted out the ports of the network forwarding IC. The traffic management unit includes multiple queues, each of which corresponds to one of the egress pipelines, and is for receiving a packet from an ingress pipeline and enqueuing the packet into one of the queues. The statistics bus connects the traffic management unit to at least a subset of the ingress pipelines, and is for providing the ingress pipelines with state information regarding the queues.

Bus for providing traffic management statistics to processing pipeline

Some embodiments provide a method for scheduling a packet to be dequeued to a processing pipeline of a hardware forwarding element. The method selects a node representing one of multiple ports associated with the processing pipeline. For each of one or more layers of logical queues, the method (i) identifies a set of logical queue nodes associated with a previously selected node based on a configuration that maps multiple physical queues to the multiple ports via the one or more layers of logical queues and (ii) selects one of the identified logical queue nodes based on properties of the identified logical queue nodes. The method selects one of a set of physical queues associated with a selected logical queue node of the last layer of logical queues. The method dequeues a next packet from the selected physical queue for processing by the processing pipeline.

Hierarchical queue scheduler

A method for packet job scheduler in data processing based on workload self-learning is disclosed. In response to receiving an incoming packet, the packet processor checks workload usage of an isolation group (IG) associated with the incoming packet, the IG being a classification of packets defined for the purpose of processing the incoming packet. The packet processor then determines whether the workload usage of the IG associated with the incoming packet exceeds a predetermined threshold. In response to the workload usage of the IG associated with the incoming packet not exceeding the predetermined threshold, the packet processor forwards the incoming packet to be executed by a packet execution unit, and then it updates the workload usage of the IG associated with the incoming packet based on execution of the incoming packet by the packet execution unit.

Method and system for packet job scheduler in data processing based on workload self-learning

A method for processing data packets in a pipeline and executed by a network processor. The pipeline includes a plurality of logical blocks, each logical block configured to process one stage of the pipeline. Each data packet includes a descriptor and a data. The network processor is coupled to a resource for storing the data. The method reduces latency and enables non-blocking processing of data packets by forwarding a unique identification of a write request from a first logical block to a subsequent second logical block in the pipeline, the write request to modify the data in the resource. The method includes receiving the descriptor for processing at the first logical block, generating the write request and the unique identification for the write request, transmitting the write request to the resource, and transmitting the unique identification towards the second logical block before an acknowledgement is returned by the resource.

Acknowledgement forwarding

A method executed by a controller of a plurality of processing elements to reduce processing time of a data packet in a network element. The processing elements are arranged in a matrix. Each processing element has a point to point connection with each adjacent processing element, known as a hop. Each processing element also includes a separate processing element storage. The data packet includes a data and a descriptor, the data being transmitted to a first processing element for storage before the descriptor is received by the controller, and the data being processed after the descriptor is received. The method includes receiving the descriptor at the controller, determining that the first processing element does not have an available resource for processing the data, determining a second processing element based on a least number of hops to the first processing element, and transmitting the descriptor to the second processing element.

Job homing

A method is performed in a requester device that is coupled with a responder device by a serial interface. The method is one of selecting an approach to handle an error in a request-response communication between the requester and responder devices over the serial interface. The method includes transmitting a request packet, of the request-response communication, over the serial interface, based on a request from an application that is executing on a logical processor. The method includes reporting the error in the request-response communication to an application-layer module. The method includes selecting an error handling approach, of a plurality of error handling approaches, which is to be used to handle the error, with the application-layer module.

Reporting errors on serial interfaces to application layers for flexible error handling approach selection

A method is implemented by a network element to provide scalable hierarchical traffic management (HTM) over a plurality of network layers for a network and eliminate priority leaking caused by quick loopback batch scheduling that analyzes a subset of network layers to shorten processing time and resource requirements when the scalable HTM selects data packets to be forwarded. The method and system function as a low pass filter over the selected data packets to prevent low priority data packets being forwarded where higher priority data packets are available to be forwarded.

Growing community of inventors

Mountain View, CA, United States of America

Michael Feng