Abstract: A method for reducing latency in a networking application includes receiving data at a physical coding sublayer (PCS) from a media access control (MAC) sublayer, where the MAC sublayer utilizes a first clock domain operating at a first clock frequency. The method further includes performing one or more functions in the PCS on the data in the first clock domain. The method also includes transmitting the data on one or more physical medium attachment (PMA) lanes, where the one or more PMA lanes utilize a second clock domain operating at a second clock frequency, where the first clock frequency and the second clock frequency have a fixed ratio. The data is transmitted from the first clock domain to the second clock domain without buffering the data. The method also includes performing one or more functions in the PCS on the data in the second clock domain.

Abstract: Apparatuses, methods, program products, and systems for emulating end-host forwarding behavior is disclosed. A data module receives a data packet at an ingress port of a network switch. The data packet includes a source address and a destination address. The network switch includes ports designated as uplink ports and different ports designated as server ports. An update module creates an entry in the forwarding database based on the data packet. The entry for the data packet includes the source address and either an undefined server port and an undefined uplink port in response to the ingress port being an uplink port where an undefined port includes an identifier for a port that does not exist, or the ingress port and one of the one or more uplink ports in response to the ingress port being a server port. A transmission module forwards the packet to a destination address.

Abstract: Embodiments presented herein describe techniques for parsing an Internet Protocol version 6 frame and skipping extension headers of the frame. A configurable skip list is provided that specifies extension headers for a networking device to skip when parsing the frame. The networking device examines “next header” fields of each extension header to determine a next extension header in the chain. If the next extension header matches an extension header in the skip list, the networking device iterates to the next header in the chain until the end of the chain (or an extension header that does not contain a match in the list) is reached.

Abstract: Embodiments presented herein describe techniques for parsing an Internet Protocol version 6 frame and skipping extension headers of the frame. A configurable skip list is provided that specifies extension headers for a networking device to skip when parsing the frame. The networking device examines “next header” fields of each extension header to determine a next extension header in the chain. If the next extension header matches an extension header in the skip list, the networking device iterates to the next header in the chain until the end of the chain (or an extension header that does not contain a match in the list) is reached.

Abstract: Method and apparatus for reducing latency in a networking application comprises receiving data at a physical coding sublayer (PCS) from a media access control (MAC) sublayer, wherein the MAC sublayer utilizes a first clock domain operating at a first clock frequency. The method further comprises performing one or more functions in the PCS on the data in the first clock domain. The method also includes transmitting the data on one or more physical medium attachment (PMA) lanes, wherein the one or more PMA lanes utilize a second clock domain operating at a second clock frequency, wherein the first clock frequency and the second clock frequency have a fixed ratio. The method also comprises performing one or more functions in the PCS on the data in the second clock domain.

Abstract: Method and apparatus for reducing latency in a networking application comprises receiving data at a physical coding sublayer (PCS) from a media access control (MAC) sublayer, wherein the MAC sublayer utilizes a first clock domain operating at a first clock frequency. The method further comprises performing one or more functions in the PCS on the data in the first clock domain. The method also includes transmitting the data on one or more physical medium attachment (PMA) lanes, wherein the one or more PMA lanes utilize a second clock domain operating at a second clock frequency, wherein the first clock frequency and the second clock frequency have a fixed ratio. The method also comprises performing one or more functions in the PCS on the data in the second clock domain.

Abstract: Embodiments presented herein provide a TCAM-based access control list that supports disjunction operations in rules. According to one embodiment, a numeric range table is tied to the access control list. Each entry in the numeric range table includes an encode field that provides for scanning TCP flags in a TCP header of an incoming Ethernet frame. Further, each entry provides a first mask and a second mask used to test for desired set and unset TCP flags in a given frame. Each entry also provides an operation field that performs a disjunction operation that compares the first mask, the second mask, and set TCP flags in a given frame.

Abstract: Embodiments described herein provide techniques for atomically updating a ternary content addressable memory (TCAM)-based access control list (ACL). According to one embodiment, a current version bit of the ACL is determined. The current version bit indicates that a rule in the ACL is active is the version flag in the rule matches the current version bit. Through these techniques, a first set of rules can be modified to create a second set of rules (e.g., by insertions, deletions, and replacements, etc.).

Abstract: Embodiments described herein provide techniques for atomically updating a ternary content addressable memory (TCAM)-based access control list (ACL). According to one embodiment, a current version bit of the ACL is determined. The current version bit indicates that a rule in the ACL is active is the version flag in the rule matches the current version bit. Through these techniques, a first set of rules can be modified to create a second set of rules (e.g., by insertions, deletions, and replacements, etc.).

Abstract: Link aggregation is a practice that uses multiple Ethernet links between two end points in order to obtain higher bandwidth and resiliency than possible with a single link. A flow distribution technique is provided to distribute traffic between the two end points equally across all links in the group and achieve greater efficiency. The flow distribution technique generates and sub-divides a hash value based on received packet flow. The divided portions of the hash value are used in a hierarchical fashion to select a link to use for this packet.

Abstract: According to one embodiment, a method for handling congestion in a network includes determining that there is congestion on a first device in a network, setting a congestion indicator in a header of a packet to indicate an amount of congestion at the first device, sending the packet to all devices that send traffic to the first device, receiving the packet having the multi-bit indicator in a header thereof at a device that sends traffic to the first device, and reducing a congestion window by a factor of between about 5% and about 50% based on a severity of the congestion indicated by the multi-bit indicator, wherein the congestion window is reduced by a greater factor when the congestion is indicated as being more severe. Other systems and methods for handling congestion in a network are described according to more embodiments.

Abstract: A technique for operating a high performance computing cluster includes monitoring workloads of multiple processors. The high performance computing cluster includes multiple nodes that each include two or more of the multiple processors. Workload information for the multiple processors is periodically updated in respective local job tables maintained in each of the multiple nodes. Based on the workload information in the respective local job tables, one or more threads are periodically moved to a different one of the multiple processors.

Abstract: A technique for operating a high performance computing (HPC) cluster includes monitoring communication between threads assigned to multiple processors included in the HPC cluster. The HPC cluster includes multiple nodes that each include two or more of the multiple processors. One or more of the threads are moved to a different one of the multiple processors based on the communication between the threads.

Abstract: A packet parser has a set of marker elements each comprising a one bit latch and connected to store flag values from the results of the application of parser rules. Some marker elements are connected to provide the stored marker values as input to the parser rule logic to be taken into account in the processing of subsequent parser rules and some are connected to control external hardware. Some markers are reset at the end of each packet. A special toggle marker element toggles its value when its address is selected and other marker elements are connected to store, when its own address is selected, the value of the toggle element. Other markers toggle their own value when selected.

Abstract: A packet parser has a set of marker elements each comprising a one bit latch and connected to store flag values from the results of the application of parser rules. Some marker elements are connected to provide the stored marker values as input to the parser rule logic to be taken into account in the processing of subsequent parser rules and some are connected to control external hardware. Some markers are reset at the end of each packet. A special toggle marker element toggles its value when its address is selected and other marker elements are connected to store, when its own address is selected, the value of the toggle element. Other markers toggle their own value when selected.

Abstract: Provided is a method for operating a host network controller for a network processor where the host network controller has at least one register. A restricted set of state data may be stored in at least one register. The restricted set of state data may be indicative of fewer possible states than those defined in accordance with a standardized communication protocol. The host network controller may implement a finite state machine based on the restricted set of state data stored in the at least one register such that the finite state machine may have fewer possible states than those defined in accordance with the standardized communication protocol.

Abstract: A network interface controller for providing a connection for a device to a network. The network interface controller may include a sideband port controller. The sideband port controller may provide a sideband connection between the network and a sideband endpoint circuit that is operative to communicate information with the network via the sideband. The sideband port controller may include a transmit data route having an input for receiving packets from the sideband endpoint circuit and an output for passing packets received from the sideband endpoint to the network. A packet parser is connected to the transmit data route. The packet parser is operative to read data from packets received from the sideband endpoint and is further operative to analyze the data.

Abstract: Aspects of the present disclosure are directed towards a network interface controller that could provide a connection for a device to a network. The network interface controller can include a sideband port controller. The sideband port controller can provide a sideband connection between the network and a sideband endpoint circuit that can be operative to communicate with the network via a sideband. The sideband port controller can include an event notification unit operative to compile information into an event notification packet. The sideband port controller can further include a packet parser. In embodiments, the packet parser could be operative to analyses a packet to provide an indication that the packet contains the event notification packet. In embodiments, the sideband port controller could be operative to forward the information in the event notification packet to the sideband endpoint circuit, responsive to that indication.

Abstract: A network interface controller (NC) that can provide a connection for a device to a network. The NC can include a sideband port controller. The sideband port controller can provide a sideband connection between the network and a sideband endpoint circuit that can communicate information with the network via the sideband. The sideband port controller can include a receive data route that has an input for receiving packets of data from the network and an output for passing the packets of data received from the network to the sideband endpoint circuit. The receive data route may include a buffer to receive the packets of data from the network and to pass the packets of data received from the network to the sideband endpoint.

Abstract: A network interface controller includes a media access controller connected to receive an in-band packet and further connected to receive a sideband packet. The network interface controller includes a host adapter that includes a receive route connected to receive the in-band packet and the sideband packet from the media access controller, and further connected to transmit the in-band packet to a host. The network interface controller includes a sideband port controller comprising a sideband receive buffer. The host adapter further includes a first receive buffer to store the in-band packet and to store the sideband packet. The host adapter further includes an arbiter connected to allow, at a time, the in-band packet to advance from the first receive buffer along the receive route towards the host and further connected to allow, at a different time, the sideband packet to advance to the sideband receive buffer of the sideband port controller.