Marvell Israel (misl) Ltd.

Marvell International Limited

Marvellworld Trade Ltd.

Xsight Labs Ltd.

Marvell Semiconductor Israel Ltd.

Marvell Israel (m.l.s.l) Ltd.

Marvell Isreal (m.i.s.l.) Ltd.

Marvell Israel Ltd.

Carmi Arad

Gil Levy

David Melman

Tal Mizrahi

Rami Zemach

Martin Leslie White

Ilan Mayer-Wolf

Zvi Shmilovici Leib

Tsahi Daniel

Nafea Bishara

Youval Nachum

Yaniv Kopelman

Dror Bromberg

Aviran Kadosh

Sharon Ulman

Yoram Revah

Lior Valency

Amir Roitshtein

Ehud Sivan

Ronen Tausi

Roi Sherman

Dor Joseph Kampeas

Nir Haim Arad

Jakob Carlström

Ilan Yerushalmi

Denis Krivitski

Guy Koren

Gal Malach

Gideon Navon

Kurt Thomas Boden

Nafea Bshara

Evgeny Shumsky

Ori Rottenstreich

Erez Shaizaf

Yaron Zimerman

Ilan Mayer-wolf

Yosef Katan

A network device includes a receive processor configured to store, in a packet memory, a payload of a packet received from a communication network. The network device also includes a packet processor configured to modify one or more fields of a header of the packet to generate a modified header, perform egress classification of the packet based on the modified header, and store the modified header in the packet memory. The network device further includes a transmit processor configured to transmit the packet in accordance with the egress classification. The transmit processor is configured to, in response to a decision that the packet is to be transmitted from the network device, generate a transmit packet from the payload retrieved from the packet memory and the modified header retrieved from the packet memory and cause the transmit packet to be transmitted to a destination in the communication network.

Egress packet processing using a modified packet header separate from a stored payload

A source switching device in a switching system receives information measured by a target switching device in the switching system. The information is indicative of an amount of data received in a given amount of time by the target switching device via each of two or more first links coupled to the target switching device. The source switching device determines, based at least in part on the information received from the target device, a path, from among multiple paths from the source switching device to the target switching device, for transmission of a packet flow directed to the target switching device. The source switching device transmits, via the determined path for transmission of the packet flow to the target device, one or more packets belonging to the packet flow.

Distributed dynamic load balancing in network systems

An elastic memory system that may include memory banks, clients that are configured to obtain access requests associated with input addresses; first address converters that are configured to convert the input addresses to intermediate addresses within a linear address space; address scramblers that are configured to convert the intermediate addresses to physical addresses while balancing a load between the memory banks; atomic operation units; an interconnect that is configured to receive modified access requests that are associated with the physical addresses, and send the modified access requests downstream, wherein atomic modified access requests are sent to the atomic operation units; wherein the atomic operations units are configured to execute the atomic modified access requests; wherein the memory banks are configured to respond to the atomic modified access requests and to non-atomic modified access requests.

Elastic resource management in a network switch

A network device processes received packets to determine port or ports of the network device via which to transmit the packets. The network device classifies the packets into packet flows and selects, based at least in part on one or more characteristics of data being transmitted in the respective packet flows, a first packet memory having a first memory access bandwidth or a second packet memory having a second memory access bandwidth, and buffers the packets in the selected first or second packet memory which the packets are being processed by the network device. After processing the packets, the network device retrieves the packets from the first packet memory or the second packet memory in which the packets are buffered, and forwards the packets to the determined one or more ports for transmission of the packets.

Hybrid packet memory for buffering packets in network devices

A network device includes a packet processor that: determines at least one egress port via which a received packet is to be transmitted by the network device; modifies one or more fields in a header of the packet to generate a modified header; determines, based at least in part on the modified header, whether the packet a) is to be transmitted or b) is to be discarded; and stores the modified header in a packet memory. In response to the determination that the packet is to be transmitted, a transmit processor of the network device: retrieves a payload of the packet from the packet memory; retrieves the modified header from the packet memory; generates a transmit packet at least by combining the payload of the packet with the modified header; and transmits the transmit packet via the determined at least one egress port of the network device.

In a switching system that comprises a central switching device an at least one port extender device, the central switching device includes at least one port configured to interface with the port extender device, and the port extender device includes a plurality of front ports for interfacing with one or more networks. The central switching device includes a processor that processes packets received from the at least one port extender device, and a plurality of egress queues for storing processed packets that are to be forwarded to the at least one port extender device for transmission via ones of the front ports. The central switching device also includes a flow control processor configured to, responsively to flow control messages received from the at least one port extender device, control transmission of packets to the at least one port extender device to prevent overflow of egress queues of the port extender device.

Traffic management in a network switching system with remote physical ports

A method for responding to a failure of a main die of a switch data-plane device, the method may include applying a secondary packet forwarding process by multiple chiplets, following the failure of the main die and during at least a part of an execution of a synchronous graceful process that follows the failure of the main die; wherein the multiple chiplets are interconnected to each other by a secondary interconnect; wherein the multiple chiplets and are coupled to the main die by a primary interconnect; wherein the applying of the secondary packet forwarding process is less complex than a primary forwarding process applied by the main die while the main die is functional.

Responding to a failure of a main die of a switch data-plane device

A network device processes received packets at least to determine port or ports of the network device via which to transmit the packet. The network device also classifies the packets into packet flows, the packet flows being further categorized into traffic pattern categories characteristic of traffic pattern characteristics of the packet flows. The network device buffers, according to the traffic pattern categories of the packet flows, packets that belong to the packet flows in a first packet memory or in a second packet memory, the first packet memory having a memory access bandwidth different from a memory access bandwidth of the second packet memory. After processing the packets, the network device retrieves the packets from the first packet memory or the second packet memory in which the packets are buffered, and forwards the packets to the determined one or more ports for transmission of the packets.

A network device determines, based on a size of a lookup value, that the lookup value is to be stored across a set of two or more memory banks including a first memory bank and a second memory bank of a database. A first hash function is for determining locations for storing lookup values entirely in the first memory bank, whereas a second hash function is for determining locations for storing lookup values entirely in the second memory bank. A hash operation is performed on the lookup value using the first hash function to determine a memory location for storing the lookup value. A first segment of the lookup value is stored in the first memory bank at the memory location determined using the first hash function, and a second segment of the lookup value is stored in the second memory bank at the memory location determined using the first hash function.

Storing keys with variable sizes in a multi-bank database

At least a payload of a packet that is received by a network device is stored in a packet memory. The packet is processed at least to determine at least one egress port via which the packet is to be transmitted, modify a header of the packet to generate a modified header, and determine, based at least in part on the modified header, whether the packet is to be transmitted or to be discarded by the network device. In response to determining that the packet is to be transmitted, the at least the payload of the packet is retrieved from the packet memory, a transmit packet is generated at least by combining the at least the payload of the packet with the modified header, and the transmit packet is transmitted via the determined at least one egress port of the network device.

A packet processor of a network device determines an amount of free buffer space in a buffer memory currently available for buffering packets, and dynamically determines a value of a threshold for triggering a particular traffic management operation with respect to a packet, to dynamically adjust the value of the threshold based at least in part on a changing amount of free buffer space available for buffering packets in the buffer memory. The packet processor determines, based on a comparison between i) a current fill level of a particular transmit queue in which the packet is to be enqueued and ii) the dynamically adjusted value of the threshold, whether the particular traffic management operation is to be triggered with respect to the packet. When the particular traffic management operation is to be triggered, the packet processor performs the particular traffic management operation with respect to the packet.

Congestion avoidance in a network device

In a network device, a flow classification hardware engine is configured to: store flow state information regarding known flows of packets in a flow information table in association with respective assigned flow identifiers (IDs). The assigned flow IDs are from an ordered set of M flow IDs, where M is a positive integer. In response to detecting new flows of packets, the flow classification hardware engine: i) assigns respective flow IDs, from the ordered set of M flow IDs, to the new flows, and ii) creates respective entries in the flow information table for the new flows. An embedded processor periodically, as part of a background process: i) identifies an oldest assigned flow ID, from the ordered set of M flow IDs, and ii) makes storage space in the flow information table corresponding to the oldest assigned flow ID available for a new flow.

Automatic flow learning in network devices

A local network device of a network switching system determines, based on a first set of received packets that contain markings indicating congestion at one or more other network devices in the network switching system, one or more respective congestion levels of one or more network paths through the network switching system. The local network device selects, based on determined congestion levels, network paths via which a second set of received packets are to be forwarded for load balancing. The local network device alters header information in the second set of received packets, the altered header information to be used by other network devices in the network switching system to make network path selections so that the second set of received packets are subsequently forwarded responsively to the altered header information along the selected network paths within the network switching system.

Methods and apparatus for load balancing in a network switching system

Growing community of inventors

Nofit, Israel

Carmi Arad