Advanced Micro Devices Corporation

Cisco Technology, Inc.

Globalfoundries Inc.

Kishore Karighattam

Robert Alan Williams

Peter Chow

Harish Vasudeva

Prasad P Padiyar

Jian Wu

Vaibhav Ashok Modi

Madhu Babu Kodali

Whu-ming William Young

Chien Meen Hwang

Sangeeta Panigrahy

Madhusudhan Karnam Rao

Sujalendu Das

In one embodiment, a system comprising memory and processor(s), the processor(s) at least adapted to compute representative latency value(s) for a network device, based on a plurality of latency values obtained for a plurality of packets which passed through the network device over a period of time, the plurality of latency values indicative of latency between ingress to and egress from the network device, obtain latency value(s) for packet(s) which passed through the network device after the period of time, the latency value(s) indicative of latency between ingress to and egress from the network device, determine a determination of congestion level(s) for the network device, including analyzing at least one of the latency value(s) in relation to at least one of the representative latency value(s), and generate an indication in accordance with the determination, the indication to be outputted on at least one output device.

Determination and indication of network traffic congestion

According to one aspect, a method includes coupling first and second security units in series between first and second networks. The first security unit obtains packets from the first network, and the second security unit obtains the packets from the first security unit. The first security unit includes first logic arranged to provide security. The second security unit includes second logic arranged to provide security. The method also includes configuring the second security unit in a bypass mode such that the second logic does not provide security, and obtaining a first packet from the first network via the first security unit. The first packet is identified as secure by the first logic. Finally, the method includes providing the first packet from the second security unit to the second network by passing the first packet through the second unit without using the second logic to provide security.

System and method for providing high availability in routers using an L1 bypass

The present invention facilitates network throughput by dynamically generating IPG values, which are employed when recovering from network collisions. Testing a number of IPG values and tracking collisions that occur as a result for each value generate the IPG value. The IPG value that yielded the lowest collisions is then programmed as the IPG value for a network device. This value remains in use by the network device for a stable state time period. After this period, another dynamically generated IPG value is obtained.

Dynamic inter packet gap generation system and method

The present invention facilitates overall system performance by balancing system resource utilization and network throughput. The invention analyzes packets received from host software selects one or more of buffers to coalesce into a single, coalesced buffer. The selection is based upon an initial fragment size selected to facilitate overall system performance. The coalesced buffer and non-coalesced buffers, are passed to a network device for transmission.

Partial coalescing of transmit buffers

The present invention is a four layer architecture that includes reusable components that can be used in varied operating environments and with varied network devices. The architecture includes an operating system dependent layer, an operating system independent layer, a media independent layer, and a media dependent layer. The operating system dependent layer can vary for differing operating system. In contrast, the operating system independent layer is identical or substantially similar for differing operating systems. Similarly, the media independent layer is identical or substantially similar for different network devices (e.g., a family or group of network devices). The media dependent layer can vary according to differing and varied network devices.

Four layer architecture for network device drivers

Aspects for performing localized diagnostics in a station of a home phoneline networking alliance (HPNA) network are described. The aspects include providing a network state and control register comprising a plurality of bits to indicate status of a network state machine of an HPNA media access controller (MAC). A test mode bit is also provided as one of the plurality of bits in the network state and control register. The test mode bit allows overriding of one or more other bits in the network state and control register to direct changes to the network state machine logic state and diagnose performance in response to the changes.

Network state diagnostics for a home phoneline networking alliance media access controller (HPNA MAC)

The present invention provides a method for supporting frame priority in a home phone line network. The method includes: detecting a limited automatic repeat request (LARQ) header in a frame with a priority tag; stripping the LARQ header and a frame check sequence (FCS) in the frame with the priority tag; recalculating the FCS for the stripped frame with the priority tag; and adding the recalculated FCS to the stripped frame with the priority tag. The method strips the LARQ header from a HPNA frame with the priority tag before it is sent to an Ethernet controller. By stripping the LARQ header, the Ethernet controller will correctly recognize the priority tag and send the frame to the appropriate priority queue. In this manner, frame priority is supported.

Mechanism to strip LARQ header and preserve LARQ header in status frame

A method for communicating a Physical Layer (PHY) mean square error (MSE) to an upper layer device driver includes: receiving a frame by the PHY; computing a MSE for the frame by the PHY; sending the MSE and the frame to a Media Access Control (MAC); inserting the MSE into a frame status frame (FSF) associated with the frame by the MAC; and sending the frame and the FSF to the upper layer driver software. With access to the PHY MSE, the upper layer driver software can compute the average mean square error (AMSE) and determine if a change in the payload encoding (PE), or data transmission rate, should be negotiated. In this manner, the data transmission rate can be optimized while maintaining a low error rate.

Method to communicate PHY mean square error to upper layer device driver for rate negotiation

A method for computer MSE to PE tables for rate negotiation has been disclosed. The method obtains probabilistic values for packet sizes in a network and obtains BER values for each FER based on these probabilistic values. An MSE for each PE is then calculated based on the BER values to obtain the upper limit tables. The MSE values in the upper limit tables is then decreased by 2 dB to obtain the lower limit tables. The MSE to PE tables may then be used for rate negotiation as set forth in the HPNA 2.0 specification.

Rate negotiation algorithm

A method for removing unwanted header information from a frame in a network is disclosed. It includes: storing beginning bytes of the frame in a first buffer and remaining bytes in a second buffer, where a size of the first buffer is smaller than the second buffer; determining that the unwanted header information is stored in the first buffer; copying bytes of the frame after the unwanted header information that are stored in the first buffer over the unwanted header information; reporting a number of bytes of the frame stored in the first buffer to be retrieved; and retrieving the reported number of bytes of the frame stored in the first buffer and the bytes of the frame stored in the second buffer. The copying of bytes occurs exclusively in the first buffer. Thus, removing the unwanted header information requires fewer processor cycles and minimizes latency in the packet receive process.

Multiple buffers for removing unwanted header information from received data packets

There is provided a communication network and method for enhancing server availability to client PCS which includes two Ethernet switches. Each one of the two Ethernet switches is connected to a corresponding one of the primary and secondary network interface cards in the file server PC. The two Ethernet switches are interconnected together through an uplink port. As a result, redundancy has been effectively and efficiently provided against the failure of either one of the two switches in order to enable link fail-over across two network segments.

Growing community of inventors

Austin, TX, United States of America

Kishore Karighattam