Juniper Networks, Inc.

Cisco Technology, Inc.

John C Carney

Michael E Lipman

Thomas Radogna

Paul J Giacobbe

David Lipschutz

Thomas A Lemaire

Ryan Ross

David Delano Ward

Donald Edward Steiss

Mohammed Ismael Tatar

John H W Bettink

Promode Nedungadi

Doron Oz

Naader Hasani

Marvin Wayne Martinez, Jr

Timothy Patrick Donahue

Mark Warden Hervin

In one embodiment, packet memory and resource memory of a memory are independently managed, with regions of packet memory being freed of packets and temporarily made available to resource memory. In one embodiment, packet memory regions are dynamically made available to resource memory so that in-service system upgrade (ISSU) of a packet switching device can be performed without having to statically allocate (as per prior systems) twice the memory space required by resource memory during normal packet processing operations. One embodiment dynamically collects fragments of packet memory stored in packet memory to form a contiguous region of memory that can be used by resource memory in a memory system that is shared between many clients in a routing complex. One embodiment assigns a contiguous region no longer used by packet memory to resource memory, and from resource memory to packet memory, dynamically without packet loss or pause.

Dynamic temporary use of packet memory as resource memory

A network device may include first logic configured to count data units passing through the network device and to produce a counter value. The network device may include second logic configured to receive the counter value when an indicator is present, and to store the counter value. The network device may include third logic configured to sample the second logic, to receive the counter value, and to operate on the counter value to produce a result.

Data unit counter

In one embodiment, rule-based virtual address translation is performed for accessing data (e.g., reading and/or writing data) typically stored in different manners and/or locations among one or more memories, such as, but not limited to, in packet switching devices. A virtual address is matched against a set of predetermined rules to identify one or more storing description parameters. These storing description parameters determine in which particular memory unit(s) and/or how the data is stored. Thus, different portions of a data structure (e.g., table) can be stored in different memories and/or using different storage techniques. The virtual address is converted to a lookup address based on the identified storing description parameter(s). One or more read or write operations in one or more particular memory units is performed based on the lookup address said converted from the virtual address.

Rule-based virtual address translation for accessing data

A service is applied in a packet switching device to both directions of a flow of packets through the packet switching device, with the application of this Layer-4 to layer-7 service to one direction requiring state information shared from the application of the service to packets traversing in the other direction. The service (e.g. firewall, network address translation) can be applied by different processing complexes which do not share memory; thus, state information is communicated between the processing complexes. When the service is applied by a single processing complex, packets can be directed explicitly to the single processing complex. The inline application of services in a packet switching system typically eliminates the need to change a packet's path through the packet switching system to that through a dedicated application server, and may eliminate the need for a dedicated services card or blade server.

Application of services in a packet switching device

A data processing architecture includes multiple processors connected in series between a load balancer and reorder logic. The load balancer is configured to receive data and distribute the data across the processors. Appropriate ones of the processors are configured to process the data. The reorder logic is configured to receive the data processed by the processors, reorder the data, and output the reordered data.

Scalable packet processing systems and methods

A packet scheduler may include logic configured to receive packet information. The packet scheduler may include logic to receive an operating parameter associated with a downstream device that operates with cell-based traffic. The packet scheduler may include logic perform a packet to cell transformation to produce an output based on the operating parameter. The packet scheduler may include logic to use the output to compensate for the downstream device.

Data unit information transformation

A system receives a set of datagrams and forms frames based on the datagrams, where at least one of the frames includes data associated with multiple ones of the datagrams. The system writes the frames to memory to form superframes in the memory, where each of the superframes includes multiple ones of frames. The system reads the superframes from the memory, recreates the datagrams based on the superframes, and outputs the datagrams.

Scalable central memory switching fabric

A method may include receiving a data unit and identifying a state of a memory storing data units. The method may include selecting a threshold value having a first threshold unit or a second threshold unit based on the state of the memory. The method may include comparing the threshold value to a queue state using the first threshold unit if the memory is in a first state. The method may include comparing the threshold value to the queue state using the second threshold unit if the memory is in a second state.

Dynamic queue management

A controller may include a measurement circuit configured to generate a proxy signal representing delay variations in the controller. The measurement circuit may also generate a measurement value from the proxy signal. A control circuit may be configured to convert the measurement value into a control value. A delay circuit may be adjusted by the control value to alter an amount of delay of a signal.

Dynamic programmable delay selection circuit and method

Hierarchical packet scheduling

A method and a network device for sharing bandwidth among a group of classes of traffic for an interface are provided. Bandwidth may be allocated to at least one traffic class of a first priority for the interface. At least some unused bandwidth of the at least one traffic class may be allocated to at least one other traffic class of a second priority for the interface. In some implementations, weighted constituents may be allocated unused interface bandwidth based on an assigned weight of each of the weighted constituents of the interface.

Growing community of inventors

Bedford, MA, United States of America

John C Carney