Abstract: System and method of data routing according to a hierarchical scheduling process. Incoming data traffic is allocated to various queues of a buffer. A scheduling tree has a top level for queues, a bottom level for egress ports, and a plurality of intermediate levels corresponding to different granularities with respect to service categories. Each queue is assigned to a particular node in each intermediate level of the scheduling tree. The scheduling tree traverses through multiple scheduling stages from the bottom to the top level to select a winner node in each level based on a variety of fairness and differentiating variables. As a result, a queue associated with the winner nodes in various levels is selected and data from the queue is read out and sent to the selected egress port for transmission.

Abstract: Embodiments of the present invention relate to a Lookup and Decision Engine (LDE) for generating lookup keys for input tokens and modifying the input tokens based on contents of lookup results. The input tokens are parsed from network packet headers by a Parser, and the tokens are then modified by the LDE. The modified tokens guide how corresponding network packets will be modified or forwarded by other components in a software-defined networking (SDN) system. The design of the LDE is highly flexible and protocol independent. Conditions and rules for generating lookup keys and for modifying tokens are fully programmable such that the LDE can perform a wide variety of reconfigurable network features and protocols in the SDN system.

Abstract: Embodiments of the apparatus for reducing latency in a flexible parser relate to an implementation that optimizes each parser engine within the parser. A packet enters the parser. Each of the parser engines processes the packet if processing is required. Otherwise, the parser engine simply forwards the packet through without processing the packet, thereby reducing latency. Each parser engine includes a memory. The memory stores bypass data and status information that indicates whether parsing for this packet is completed and, thus, no further processing is required by subsequent parser engines. Each parser engine also includes a counter, which is incremented whenever a packet enters the parser engine and is decremented whenever a packet exists the parser engine. A packet bypasses the parser engine based on the counter of the parser engine and the status information of that packet.

Abstract: Embodiments of the present invention relate to multiple parallel lookups using a pool of shared memories by proper configuration of interconnection networks. The number of shared memories reserved for each lookup is reconfigurable based on the memory capacity needed by that lookup. The shared memories are grouped into homogeneous tiles. Each lookup is allocated a set of tiles based on the memory capacity needed by that lookup. The tiles allocated for each lookup do not overlap with other lookups such that all lookups can be performed in parallel without collision. Each lookup is reconfigurable to be either hash-based or direct-access. The interconnection networks are programed based on how the tiles are allocated for each lookup.

Abstract: Embodiments of the present invention relate to a Lookup and Decision Engine (LDE) for generating lookup keys for input tokens and modifying the input tokens based on contents of lookup results. The input tokens are parsed from network packet headers by a Parser, and the tokens are then modified by the LDE. The modified tokens guide how corresponding network packets will be modified or forwarded by other components in a software-defined networking (SDN) system. The design of the LDE is highly flexible and protocol independent. Conditions and rules for generating lookup keys and for modifying tokens are fully programmable such that the LDE can perform a wide variety of reconfigurable network features and protocols in the SDN system.

Abstract: The invention describes a network lookup engine for generating parallel network lookup requests for input packets, where each packet header is parsed and represented by a programmable parser in a format, namely a token, which is understandable by the engine. Each token can require multiple lookups in parallel in order to speed up the packet processing time. The sizes of lookup keys varies depending on the content of the input token and the protocols programmed for the engine. The engine generates a super key per token, representing all parallel lookup keys wherein the content of each key can be extracted from the super key through an associated profile identification. The network lookup engine is protocol-independent which means the conditions and rules for generating super keys are full programmable so that the engine can be reprogrammed to perform a wide variety of network features and protocols in a software-defined networking (SDN) system.

Abstract: Embodiments of the apparatus for modifying packet headers relate to a rewrite engine that represents each protocol header of packets in a generic format specific to that protocol to enable programmable modifications of packets, resulting in hardware and software flexibility in modifying packet headers. Software programs generic formats in a hardware table for various protocols. The rewrite engine is able to detect missing fields from a protocol header and is able to expand the protocol header to a maximum size such that the protocol header contains all possible fields of that protocol. Each of the fields has the same offset irrespective of which variation of the protocol the protocol header corresponds to. In a bit vector, all newly added fields are marked invalid (represented by 0), and all existing fields are marked valid (represented by 1). Software modification commands allow data to be replaced, removed and inserted.

Abstract: Embodiments of the present invention relate to multiple parallel lookups using a pool of shared memories by proper configuration of interconnection networks. The number of shared memories reserved for each lookup is reconfigurable based on the memory capacity needed by that lookup. The shared memories are grouped into homogeneous tiles. Each lookup is allocated a set of tiles based on the memory capacity needed by that lookup. The tiles allocated for each lookup do not overlap with other lookups such that all lookups can be performed in parallel without collision. Each lookup is reconfigurable to be either hash-based or direct-access. The interconnection networks are programmed based on how the tiles are allocated for each lookup.

Abstract: Embodiments of the present invention relate to multiple parallel lookups using a pool of shared memories by proper configuration of interconnection networks. The number of shared memories reserved for each lookup is reconfigurable based on the memory capacity needed by that lookup. The shared memories are grouped into homogeneous tiles. Each lookup is allocated a set of tiles based on the memory capacity needed by that lookup. The tiles allocated for each lookup do not overlap with other lookups such that all lookups can be performed in parallel without collision. Each lookup is reconfigurable to be either hash-based or direct-access. The interconnection networks are programmed based on how the tiles are allocated for each lookup.

Abstract: A multicast rule is represented in a hierarchical linked list with N tiers. Each tier or level in the hierarchical linked list corresponds to a network layer of a network stack that requires replication. Redundant groups in each tier are eliminated such that the groups in each tier are stored exactly once in a replication table. A multicast replication engine traverses the hierarchical linked list and replicates a packet according to each node in the hierarchical linked list.

Abstract: A multicast destination table contains a list of links. The list of links includes the main link that is currently in use and alternate links to reach the same destination. The links in the list of links are ordered based on a metric. Each of the links is stored as an entry in the multicast destination table. A multicast replication engine traverses the list of links until an enabled link in the list of links is reached, and replicates a packet according to data associated with the enabled link in the list of links.

Abstract: A software-defined network (SDN) system, device and method comprise one or more input ports, a programmable parser, a plurality of programmable lookup and decision engines (LDEs), programmable lookup memories, programmable counters, a programmable rewrite block and one or more output ports. The programmability of the parser, LDEs, lookup memories, counters and rewrite block enable a user to customize each microchip within the system to particular packet environments, data analysis needs, packet processing functions, and other functions as desired. Further, the same microchip is able to be reprogrammed for other purposes and/or optimizations dynamically.

Abstract: Embodiments of the present invention relate to fast and conditional data modification and generation in a software-defined network (SDN) processing engine. Modification of multiple inputs and generation of multiple outputs can be performed in parallel. A size of each input or output data can be large, such as in hundreds of bytes. The processing engine includes a control path and a data path. The control path generates instructions for modifying inputs and generating new outputs. The data path executes all instructions produced by the control path. The processing engine is typically programmable such that conditions and rules for data modification and generation can be reconfigured depending on network features and protocols supported by the processing engine. The SDN processing engine allows for processing multiple large-size data flows and is efficient in manipulating such data. The SDN processing engine achieves full throughput with multiple back-to-back input and output data flows.

Abstract: A forwarding pipeline of a forwarding engine includes a mirror bit mask vector with one bit per supported independent mirror session. Each bit in the mirror bit mask vector can be set at any point in the forwarding pipeline when the forwarding engine determines that conditions for a corresponding mirror session are met. At the end of the forwarding pipeline, if any of the bits in the mirror bit mask vector is set, then a packet, the mirror bit mask vector and a pointer to the start of a mirror destination linked list are forwarded to the multicast replication engine. The mirror destination linked list typically defines a rule for mirroring. The multicast replication engine mirrors the packet according to the mirror destination linked list and the mirror bit mask vector.

Abstract: Embodiments of the present invention relate to a centralized table aging module that efficiently and flexibly utilizes an embedded memory resource, and that enables and facilitates separate network controllers. The centralized table aging module performs aging of tables in parallel using the embedded memory resource. The table aging module performs an age marking process and an age refreshing process. The memory resource includes age mark memory and age mask memory. Age marking is applied to the age mark memory. The age mask memory provides per-entry control granularity regarding the aging of table entries.

Abstract: This disclosure describes techniques and apparatuses enabling low-to-high speed cut-through communication without creating an overrun condition. By so doing, the techniques and/or apparatuses enable communication interfaces to communicate at higher speed, such as by avoiding store-to-forward latency.

Abstract: A software-defined network (SDN) system, device and method comprise one or more input ports, a programmable parser, a plurality of programmable lookup and decision engines (LDEs), programmable lookup memories, programmable counters, a programmable rewrite block and one or more output ports. The programmability of the parser, LDEs, lookup memories, counters and rewrite block enable a user to customize each microchip within the system to particular packet environments, data analysis needs, packet processing functions, and other functions as desired. Further, the same microchip is able to be reprogrammed for other purposes and/or optimizations dynamically.

Abstract: A multicast rule is represented in a hierarchical linked list with N tiers. Each tier or level in the hierarchical linked list corresponds to a network layer of a network stack that requires replication. Redundant groups in each tier are eliminated such that the groups in each tier are stored exactly once in a replication table. A multicast replication engine traverses the hierarchical linked list and replicates a packet according to each node in the hierarchical linked list.

Abstract: Embodiments of the present invention relate to a centralized table aging module that efficiently and flexibly utilizes an embedded memory resource, and that enables and facilitates separate network controllers. The centralized table aging module performs aging of tables in parallel using the embedded memory resource. The table aging module performs an age marking process and an age refreshing process. The memory resource includes age mark memory and age mask memory. Age marking is applied to the age mark memory. The age mask memory provides per-entry control granularity regarding the aging of table entries.

Abstract: A forwarding pipeline of a forwarding engine includes a mirror bit mask vector with one bit per supported independent mirror session. Each bit in the mirror bit mask vector can be set at any point in the forwarding pipeline when the forwarding engine determines that conditions for a corresponding mirror session are met. At the end of the forwarding pipeline, if any of the bits in the mirror bit mask vector is set, then a packet, the mirror bit mask vector and a pointer to the start of a mirror destination linked list are forwarded to the multicast replication engine. The mirror destination linked list typically defines a rule for mirroring. The multicast replication engine mirrors the packet according to the mirror destination linked list and the mirror bit mask vector.