Broadcom Corporation

Sun Microsystems, Inc.

Cisco Technology, Inc.

Avago Technologies International Sales Pte. Limited

Oracle America, Inc.

Howard M Frazier

Wael William Diab

Scott R Powell

Yongbum Kim

Shimon Muller

Patricia Ann Thaler

Burhan Masood

Thomas James Edsall

Howard A Baumer

Ariel Hendel

Ali Ghiasi

Andreas V Bechtolsheim

Michael David Johas Teener

Michael Johas Teener

Glen Kramer

Brad Matthews

Lowell David Lamb

Robert J Stone

Leo A Hejza

Leo Hejza

A system and method for backchannel closed loop feedback for channel equalization over Ethernet. Closed loop backchannel feedback for real-time transmitter adaptive equalization is provided for a pair of non-ideal duplex communication channels. The real-time transmitter adaptive equalization enables use of low cost relaxed specification transmitter modules at high data rates.

System and method for backchannel closed loop feedback for channel equalization over Ethernet

Aspects of a method and system for a distinct physical pattern on an active channel to indicate a data rate transition for energy efficient Ethernet. In this regard, one or more distinct physical patterns may be transmitted on one or more active channels of a network link during an inter-packet gap to control a data rate on the link. The unique physical pattern may be transmitted instead of or in addition to one or more IDLE symbols. The distinct physical pattern may communicate a data rate to be utilized on the link and/or indicate when a data rate transition should occur on the link. The distinct pattern may be transmitted and/or the data rate transition may occur during a specified inter-packet gap or during a specified packet boundary. The distinct physical pattern may comprise one or more control characters and/or an ordered set of voltage levels, symbols, and/or characters.

Method and system for a distinct physical pattern on an active channel to indicate a data rate transition for energy efficient ethernet

Systems, devices, and methods of implementing 50 Gb/s Ethernet using serializer/deserializer lanes are disclosed. One such device includes circuitry operable to provide a media access control (MAC) interface. The MAC interface is associated with a port having a 50 Gb/s link rate. The device also includes circuitry operable to generate Ethernet frames from data received at the MAC interface and circuitry operable to distribute the Ethernet frames across a group of serial/deserializer (SERDES) lanes associated with the port, the group having size N. The device also includes circuitry operable to transmit the distributed Ethernet frames on each of the SERDES lanes at a 50/N Gb/s rate.

50 Gb/s Ethernet using serializer/deserializer lanes

A device for pre-emption in passive optical networks may include a first media access control (MAC) module configured to receive a first type of data traffic and transmit the first type of data traffic to a MAC merge module. The device may further include a second media access control (MAC) module configured to receive a second type of data traffic and transmit the second type of data traffic to the MAC merge module. The device may further include the MAC merge module configured to receive the first and second types of data traffic from the first and second MAC modules, respectively, and provide the first and second types of data traffic for transmission over a port. The MAC merge module may be configured to pre-empt the transmission of the first type of data traffic over the port in favor of the second type of data traffic.

Pre-emption in passive optical networks

In an Ethernet network comprising link partners that are coupled via an Ethernet link, an energy efficient network (EEN) communication control policy may be established above a MAC layer. The EEN control policy may determine a power level mode of operation for the link partners. Power level mode control data may be communicated from hardware, software and/or firmware of upper layer protocols to PHY devices. The control data may specify when, where and/or how long to implement the power level mode. The control data may be communicated via a 'magic' packet comprising start and/or end of frame delimiters and bits that indicate the power level mode. The PHY devices may detect, parse and/or interpret the 'magic' packet and may enter and/or exit a lower power mode of operation. The PHY devices may respond to by generating a magic packet and sending the packet to upper layers.

Method and system for packet based signaling between A Mac and A PHY to manage energy efficient network devices and/or protocols

While transmitting a first Ethernet frame from the first buffer onto an Ethernet link, a first Ethernet device may stop transmitting the first frame prior to completing transmission of the frame. The first Ethernet device may then transmit a second frame from a second buffer onto the Ethernet link. The first Ethernet device may resume transmission of the first frame from the first buffer onto the Ethernet link. A second Ethernet device may receive, via the Ethernet link, a first portion of a first Ethernet frame and store the first portion of the first Ethernet frame in a first buffer. The second Ethernet device may then receive, via the Ethernet link, a second Ethernet frame and store the second Ethernet frame in a second buffer. The second Ethernet device may then receive, via the Ethernet link, a second portion of the first Ethernet frame and append it to the contents of the first buffer.

Packet preemption for low latency

An Ethernet link may comprise silent and active channels and may support energy efficient Ethernet communication. Training parameters from the one or more active channels may be utilized for determining and/or adjusting training parameters for silent channels prior to activation. Training parameters for silent channels may be determined based on copying training parameters from active channels. Determination of training parameters for silent channels may be based on a weighted average of the active channel training parameters. A delta between active channel training parameters from a prior time and subsequent time may be utilized to determine a correction factor for adjusting training parameters for a silent channel from a prior time. Silent channels may be adjusted based on active channel training parameters and then subsequently may be trained. Training parameters may be adjusted for one or more of an echo canceler, a near-end crosstalk canceler and a far-end canceler.

Method and system for training an ethernet channel based on an active channel to support energy efficient ethernet networks

Aspects of a method and system for compensated time stamping for time-sensitive network communications are provided. In this regard, one or more timestamps generated in an OSI layer above the physical layer may be adjusted based on parameters associated with an amount of time in which data traverses a PHY of the network device. Communications of the network device may be managed based on the adjusted one or more timestamps. The parameters may comprise one or more of: average ingress PHY traversal time, average egress PHY traversal time, variance of ingress PHY traversal time, and variance of egress PHY traversal time. One or more network links coupled to the network device may be characterized based on the one or more adjusted timestamps. The parameters may be stored in one or more registers within a PHY of the network device.

Method and system for compensated time stamping for time-sensitive network communications

Various aspects of a method and system for low-latency networking are provided. Latency requirements of traffic to be communicated along a network path comprising one or more Ethernet links may be determined. A maximum size of Ethernet frames utilized for communicating the traffic may be determined based on the latency requirements. The maximum size of the Ethernet frames may be determined based on a data rate of one or more Ethernet links along the network path. A single device may utilize different maximum packet sizes for different ports/links on which it communicates. One or more messages indicating the determined maximum size may be communicated among devices along the network path to coordinate maximum packet sizes.

Method and system for low-latency networking

Systems, devices, and methods of implementing 50 Gb/s Ethernet using serializer/deserializer lanes are disclosed. One such device includes circuitry operable to provide a media access control (MAC) interface. The MAC interface is associated with a port having a 50 Gb/s link rate. The device also includes circuitry operable to generate Ethernet frames from data received at the MAC interface and circuitry operable to distribute the Ethernet frames across a group of serializer/deserializer (SERDES) lanes associated with the port, the group having size N. The device also includes circuitry operable to transmit the distributed Ethernet frames on each of the SERDES lanes at a 50/N Gb/s rate.

50 Gb/s ethernet using serializer/deserializer lanes

A system and method for carrier deferral for full duplex energy efficient Ethernet (EEE) physical layer devices (PHYs). A carrier deferral signal can be asserted to a media access control layer to indicate to the media access control (MAC) layer that transmission of data is to be deferred due to a power savings initiative in the physical layer device. In one example, the carrier deferral signal is used when a PHY is awakened by the MAC when the MAC has something to transmit. In another example, the carrier deferral signal is used when a PHY is switching link rates.

System and method for carrier deferral for full duplex energy efficient ethernet PHYs

An Ethernet link may comprise silent and active channels and may support energy efficient Ethernet communication. Training parameters from the one or more active channels may be utilized for determining and/or adjusting training parameters for silent channels prior to activation. Training parameters for silent channels may be determined based on copying training parameters from active channels. Determination of training parameters for silent channels may be based on a weighted average of the active channel training parameters. A delta between active channel training parameters from a prior time and subsequent time may be utilized to determine a correction factor for adjusting training parameters for a silent channel from a prior time. Silent channels may be adjusted based on active channel training parameters and then subsequently may be trained. Training parameters may be adjusted for one or more of an echo canceller, a near-end crosstalk canceller and a far-end canceller.

Aspects of a system for keyboard, sound and mouse (KSM) over LAN A/V bridging and A/V bridging extensions for graphics thin client applications may include an I/O controller within an LAN subsystem, which enables generation of response data based on input signals from a keyboard device and/or mouse device that is coupled to a computing device. The generated response data may be encapsulated within an encapsulating PDU. The encapsulating PDU may comprise an Ethernet frame and/or and IP packet. The encapsulating PDU may be transmitted via a network based on a traffic class designation. The traffic class designation may enable the transmission of the generated response data via the network based on specified latency target values. A time stamp value may be generated for the encapsulating PDU. The encapsulating PDU may include at least one data type identifier, which indicates that the encapsulating PDU contains the generated response data.

Method and system for keyboard, sound and mouse (KSM) over LAN A/V bridging and A/V bridging extensions for graphics thin client applications

An Ethernet network may comprise multi-rate link partners that may be coupled via an Ethernet link. The Ethernet link may comprise a plurality of channels wherein one or more may be silent and/or idle channels set to a low(er) power that may enable energy efficiency. Silent channels may be utilized for transmitting signals comprising one or more distinct physical patterns for controlling data rate transitions. For example, symbols may be utilized for generating the distinct physical patterns. The distinct physical patterns may communicate a new data rate, when to transition the data rate, which channel(s) may bear the data rate transition and/or a method for implementing the data rate transition. The method may comprise activating or suspending a data traffic channel, modifying a signal constellation, modifying PAM levels and/or adjusting inter-frame gap. The new data rate may be determined based on past traffic and/or expected traffic on the network link.

Growing community of inventors

Newark, CA, United States of America

Howard M Frazier