Abstract: A system includes a data processing unit (DPU) and DPU resource management circuits. A DPU resource management circuit establishes an interface associated with managing resources of the DPU, where establishing the interface is based on another DPU resource management circuit of the system accessing the DPU. The DPU resource management circuit monitors usage of the resources of the DPU based on auditing data provided by the second DPU resource management circuit. In some cases, the DPU resource management circuit may allocate resources of the DPU to an application. The DPU resource management circuit generates a catalog data structure of the resources of the DPU. The catalog data structure includes entries corresponding to resources of the DPU and functions provided by the resources.

Abstract: Methods and systems are provided for Reducing Write I/O Latency Using Asynchronous Fibre Channel Exchange. A FCP target device may send one or more FC write control information units (IUs) to a FCP initiator device within a first FC exchange to request a transfer of data associated with a FCP write command IU previously sent to the FCP target device by the FCP initiator device within a second FC exchange. The FC write control IUs may be sent without the FCP target device first receiving from the FCP initiator device a sequence initiative of the second FC exchange, and/or may be sent within the first FC exchange concurrently with the FCP initiator device sending one or more FCP data IU sequences within the second FC exchange to the FCP target device. Thus, a full-duplex communication environment may be setup between the FCP initiator device and FCP target device.

Abstract: A FCP initiator sends a FCP write command to a FCP target within a second FC Exchange, and the target sends one or more FC write control IUs to the initiator within a first FC Exchange to request a transfer of data associated with the write command. The first and second FC exchanges are distinct from one another. A payload of each write control IU includes an OX_ID value with which the initiator originated the second Exchange and a RX_ID value assigned by the FCP target for the second exchange. The two Exchanges yield a full-duplex communication environment between the initiator and target that enables the reduction or elimination of latencies incurred in a conventional FCP write I/O operation due to the half-duplex nature of a single FC Exchange. The write control IU may be an enhanced FCP_XFER_RDY IU or a new FC IU previously undefined by the FCP standard.

Abstract: An Ethernet node includes a receiver and transmitter for coupling to an Ethernet link for transceiving Ethernet frames with a remote Ethernet node at a remote end of the Ethernet link. The Ethernet node sends to the remote Ethernet node credit information that enables the remote Ethernet node to know an amount of Ethernet frames the remote Ethernet node is allowed to send to the Ethernet node. The credit information may be specified on a VLAN priority basis and may be included in a header piggy-backed on a host Ethernet frame or may be included in an explicit frame. The credit information may incrementally replenish credits or may specify an absolute credit amount and may be sent in response to processing frames, detection of a frame error, receipt of a request from the remote node or other events.

Abstract: An Ethernet node includes a receiver and transmitter for coupling to an Ethernet link for transceiving Ethernet frames with a remote Ethernet node at a remote end of the Ethernet link. The Ethernet node sends to the remote Ethernet node credit information that enables the remote Ethernet node to know an amount of Ethernet frames the remote Ethernet node is allowed to send to the Ethernet node. The credit information may be specified on a VLAN priority basis and may be included in a header piggy-backed on a host Ethernet frame or may be included in an explicit frame. The credit information may incrementally replenish credits or may specify an absolute credit amount and may be sent in response to processing frames, detection of a frame error, receipt of a request from the remote node or other events.

Abstract: A FCP initiator sends a FCP write command to a FCP target within a second FC Exchange, and the target sends one or more FC write control IUs to the initiator within a first FC Exchange to request a transfer of data associated with the write command. The first and second FC exchanges are distinct from one another. A payload of each write control IU includes an OX_ID value with which the initiator originated the second Exchange and a RX_ID value assigned by the FCP target for the second exchange. The two Exchanges yield a full-duplex communication environment between the initiator and target that enables the reduction or elimination of latencies incurred in a conventional FCP write I/O operation due to the half-duplex nature of a single FC Exchange. The write control IU may be an enhanced FCP_XFER_RDY IU or a new FC IU previously undefined by the FCP standard.

Abstract: An Ethernet node includes a receiver and transmitter for coupling to an Ethernet link for transceiving Ethernet frames with a remote Ethernet node at a remote end of the Ethernet link. The Ethernet node sends to the remote Ethernet node credit information that enables the remote Ethernet node to know an amount of Ethernet frames the remote Ethernet node is allowed to send to the Ethernet node. The credit information may be specified on a VLAN priority basis and may be included in a header piggy-backed on a host Ethernet frame or may be included in an explicit frame. The credit information may incrementally replenish credits or may specify an absolute credit amount and may be sent in response to processing frames, detection of a frame error, receipt of a request from the remote node or other events.

Abstract: Embodiments of the invention relate to performing network communications according to an existing protocol by using frames that are larger than those usually allowed by the existing protocol. Thus, embodiments of the present invention provide for an extension of the existing protocol which allows for the use of larger frames. The larger frames may result from the use of larger payloads and their size may be defined in terms of the payload size. Embodiments provide for use of various negotiation and initialization mechanisms of the existing protocol with additional modifications to allow for the negotiation of the use of larger frames. These modifications may provide for end to end negotiation of a larger frame or payload size. Some embodiments ensure that the negotiations are performed in such a manner that devices that feature the improvements of the present invention can communicate with devices that do not feature these improvements.

Abstract: Embodiments of the invention relate to performing network communications according to an existing protocol by using frames that are larger than those usually allowed by the existing protocol. Thus, embodiments of the present invention provide for an extension of the existing protocol which allows for the use of larger frames. The larger frames may result from the use of larger payloads and their size may be defined in terms of the payload size. Embodiments provide for use of various negotiation and initialization mechanisms of the existing protocol with additional modifications to allow for the negotiation of the use of larger frames. These modifications may provide for end to end negotiation of a larger frame or payload size. Some embodiments ensure that the negotiations are performed in such a manner that devices that feature the improvements of the present invention can communicate with devices that do not feature these improvements.