Cisco Technology, Inc.

David Walker

Kenneth Michael Rose

Kevin C Wong

Venkateshwar Rao Pullela

Matthew Todd Lawson

Daniel R Ullum

Kaichuan He

Yu Kwong Ng

Vijay A Nebhrajani

Deepak Srinivas Mayya

Ravikiran Kaidala Lakshman

Tanjore K Suresh

Thomas James Edsall

Kalyan Ghosh

Sagar Borikar

Ranganathan Rajagopalan

Mick R Jacobs

Amit Surendra Singh

Chickayya Naik

Shrikant Jay Vaidya

Michael A Benning

Lalit Kumar

Gaetano Borgione

Srividhya Nagarajan

Kar Wai Kam

Reese Faucette

Ravi Shankar Palagummi

This disclosure describes techniques for a network interface controller (NIC) to communicate over a network according to a specified transmission communication protocol. The NIC receives, from a host processor, an indication of a data unit and an indication of a header template. The NIC generates, independent of the specified transmission communication protocol, metadata for transmitting a segment of the data unit over a network. The NIC generates, at last partly using the metadata, header fields that comply with the specified transmission communication protocol. The NIC processes the header template and the header fields to generate a packet header, and the NIC transmits, over the network, a packet including the packet header and the segment of the data unit.

Configurable segmentation offload

Presented herein are methodologies for increasing effective throughput on a network. A method includes receiving a command request via a communication bus, the command request including a command ID, determining, based on the command ID, whether data in the command request is to be joined with data from other command requests having the same command ID, when it is determined, based on the command ID, that the data in the command request is to be joined with other data from other command requests having the same command ID, writing the data to a selected buffer in which the other data is already stored, and causing the data and the other data in the buffer to be sent as a payload of a single packet across a communications fabric.

Increasing throughput of non-volatile memory express over fabric (NVMEoF) via peripheral component interconnect express (PCIe) interface

In one embodiment, a direct memory access (DMA) controller within a host device obtains a packet to be processed by the host device, where the host device comprises a host processor, a network interface controller (NIC), and a co-processor of the NIC, and where the co-processor is configured to perform one or more specific packet processing operations. The DMA controller may then detect a DMA descriptor of the packet, and can determine, according to the DMA descriptor, how the packet is to be moved for processing within the host device. As such, the DMA controller may then move the packet, based on the determining, to one of either a host main memory, a NIC memory, or a co-processor memory of the host device.

Offloading data movement for packet processing in a network interface controller

Traffic between a first device and a second device across a plurality of communication lanes of a communication link is stalled. A first device side of the communication lane is powered-down on at least one of the plurality of communication lanes to deactivate the at least one communication lane. A power-down request is sent from the first device to the second device to power-down a second device-side of the at least one communication lane. Traffic between the first device and the second device is restarted over the active communication lanes.

Dynamic communication link scaling

Various techniques can be used to handle frames within multi-stage switching fabric. For example, in one method, a frame and an associated frame header are received at a switching fabric stage. The associated frame header includes a first field and a second field. The method selects one or more fabric points of exit within the switching fabric stage, based on the second field. The first field is used to select one or more other fabric points of exit within another switching fabric stage, and thus two different fields within the associated frame header specify fabric points of exit. The method then sends the frame to the selected fabric points of exit within the switching fabric stage.

Frame handling within multi-stage switching fabrics

The method can be implemented on a processor executing software instructions stored in memory. In one embodiment of the invention, the method includes receiving an Ethernet frame, wherein the Ethernet frame comprises a Transmission Control Protocol (TCP) header, wherein the TCP header comprises a TCP header length value. When the Ethernet frame is received, the TCP header length value is compared to a predetermined value.

Apparatus and method for filtering IP packets

An apparatus and method for unconditionally loading a value into first memory of an first integrated circuit, which operates in one of several different modes depending on value stored in the first memory. In one embodiment, apparatus comprises a printed circuit board. The first integrated circuit (IC) is mounted on the printed circuit board, wherein the first IC comprises a first memory device, and wherein the first IC is configured to operate in a first mode when a first value is stored in the first memory device, and wherein the first IC is configured to operate in a second mode when a second value is stored in the first memory device. The printed circuit board also includes a second IC mounted thereon. The second IC comprises a second memory device that stores the first value. A third IC mounted on the printed circuit board is configured to provide a copy of the first value stored in the second memory device to the first IC for storage in the first memory device, wherein the third IC is configured to provide the copy of the first value to the first IC without condition.

Method using virtual ASIC pins to reconfigure hardware

A self-adjusting load balancing among multiple fabric ports. A plurality of first values is received in response to receiving a first frame, wherein each of the first values is related to a quantity of data stored in a respective one of a plurality of buffers of fabric ports. First identifiers are also received, each of which corresponds to a respective one of a first subset of the plurality of buffers. A subset of the first identifiers is selected based on one or more of the first values. Thereafter one of the first identifiers contained in the selected subset is selected. Ultimately the first frame is transmitted to the buffer that corresponds to the selected one of the first identifiers.

Self-adjusting load balancing among multiple fabric ports

In one embodiment, a programmable priority encoder is configured to receive inputs, including an ordered list of a plurality of input request values each representing either a request or a non-request, and a starting position within the ordered list of the plurality of input request values. The programmable priority encoder is configured to generate an identification of a result position of a first input indicating said request in order from a position identified from the starting position within the ordered list. In one embodiment, the programmable priority encoder includes a hierarchal structure of logic blocks including a plurality of columns of logic blocks; wherein a first-stage column of the plurality of columns of logic blocks is configured to operate on at most N input values; and wherein the ordered list of the plurality of input request values consists of N input request values.

Programmable priority encoder

A method for receiving packet data at a communication channel and transmitting the packet data over serial links of the communication channel. The packet data is sliced into n-bit data portions which are concatenated with a header prior to transmitting an n-bit portion across one of the serial links of the communication channel. The header includes a CRC to provide improved error detection.

Enhanced error detection in multilink serdes channels

Disclosed is a method and apparatus for checking link layer protocol frames such as Ethernet frames. The method can be implemented on a processor executing software instructions stored in memory. In one embodiment of the invention, the method includes receiving an Ethernet frame, and counting data bytes of the Ethernet frame to generate a total number of counted bytes. The total number of counted bytes can be used to calculate a data length of a datagram of the Ethernet frame. Once calculated, the datagram data length can be compared to a predetermined value. If the datagram length does not fall within an acceptable range of the predetermined value, the Ethernet frame may be dropped so that the Ethernet frame does not reach its final destination.

Apparatus and method for detecting tiny fragment attacks

A method, system, and apparatus to transmit replicated multicast packets over a plurality of physical network links that are combined into one logical channel or link so that the replicated multicast packets are distributed over more than one network link is disclosed. It is further disclosed that distribution over the network links is accomplished, in part, through analyzing the multicast packet for information other than ethernet addresses. Such information can include a tag header including destination interface information.

Apparatus and method for utilizing aggregate network links for multicast switching

A method is disclosed for remotely monitoring network traffic through a generic network. A first data packet, which indicates a first destination network element, is received. A second data packet, which contains at least a part of the first data packet, is generated. The second data packet indicates a second destination network element that is configured to monitor network traffic. The second destination network element differs from the first destination network element.

Method and apparatus for remotely monitoring network traffic through a generic network

An integrated circuit monitors the most active traffic flow rates on a communications network by using a leaky bucket model having a variable fill rate. As a switch receives packets, the packet identifications are sampled. A sampled packet identification is compared to record identifications in a table of identifications. If the sampled and record identifications match, an activity value for the packet identification is increased by an amount inversely proportional to an activity value associated with the record identification. If the sampled and record identifications do not match, the activity value is decreased. Record identifications are removed from the table when the activity value decreases to a specified level. New sampled identifications are added to the table if empty records exist.

Growing community of inventors

San Jose, CA, United States of America

David Walker