Arista Networks, Inc.

Joseph Olakangil

Anuraag Mittal

Nitin Ravindra Karkhanis

Purushothaman Nandakumaran

Manjula Gopalakrishnan

Lakshmikantha Chowdary Pothula

Sandeep Gawai

Swaroop George

Lakshmikanth Chowdary Pothula

Software tables in the control plane of a network device are downloaded into the data plane hardware after a reboot of the control plane. Valid entries are identified and downloaded while reducing the downloading invalid or otherwise unused entries to improve download times and reduce disruption of traffic resulting from the download. In some instances, blocks of entries containing a high ratio of valid to invalid entries are identified and downloaded to the hardware. In some instances, entries in the software tables that differ from corresponding entries in the hardware (diff entries) are identified and downloaded. In some instances, chunks containing diff entries and non-diff entries are identified and downloaded.

Reduced outage time with improved download to hardware tables based on valid entries

In general, embodiments relate to a method, for managing a network device, that includes accessing, by a feature agent of the network device, an allocation data structure, wherein the allocation data structure specifies a first portion of memory and a second portion of memory, identifying, using the allocation data structure, the first portion of the memory to be used during an upgrade, wherein the second portion of memory is used for storing a network device table, wherein the network device table is used by a packet transmission component while the upgrade is being performed, and upon completion of the upgrade, updating the allocation data structure to specify that the packet transmission component use a second network device table and stop using the network device table, wherein the second network device table is initially populated during the upgrade.

Reducing a network device upgrade outage using a reduced hardware resource scaling

Systems and methods are provided herein for implementing multi-table OpenFlow flows that have combinations of packet edits. This may be accomplished by a network device receiving a first flow entry with a first set of actions to be installed into a flow table. The network device may determine that the first set of actions includes edits to a plurality of fields of a matched data packet. In response, the network device may change the first set of actions of the first flow entry to edit a first field of the data packet and create a second flow entry with a second set of actions to edit a second field of the data packet. The network device may install the first and second flow entries into one or more flow tables of the network device.

Systems and methods for implementing multi-table OpenFlow flows that have combinations of packet edits

Techniques for implementing multi-table OpenFlow using a parallel hardware table lookup architecture are provided. In certain embodiments, these techniques include receiving, at a network device from a software-defined networking (SDN) controller, flow entries for installation into flow tables of the network device, where the flow entries are structured in a manner that assumes the flow tables can be looked-up serially by a packet processor of the network device, but where the flow tables are implemented using hardware lookup tables (e.g., TCAMs) that can only be looked-up in parallel by the packet processor. The techniques further include converting, by the network device, the received flow entries into a format that enables the packet processor to process ingress network traffic correctly using the flow entries, despite the packet processor's parallel lookup architecture, and installing the converted flow entries into the flow tables/hardware lookup tables.

Implementing multi-table OpenFlow using a parallel hardware table lookup architecture

Systems and methods are provided for programming a network device. A method includes receiving a wild card entry at the network device, the network device including a ternary content addressable memory (TCAM) table and an exact match (EM) table. The method determines whether the wild card entry is compatible with the EM table. In response to determining that the wild card entry is compatible with the EM table, the method determines the available space in the EM table, the usage of the TCAM table, and at least one flow characteristic of the wild card entry. The method evaluates the determined available space in the EM table, usage of the TCAM table, and the at least one flow characteristic against a set of stored rules that select the EM table or the TCAM table. The method programs the wild card entry in the EM table or TCAM table based upon the selection.

Growing community of inventors

San Jose, CA, United States of America

Joseph Olakangil