DPDK Interfaces Configuration [Документация VAS Experts]
Документация VAS Experts Документация VAS Experts-mobile
in

Settings

  • Show page
  • Old revisions
  • Export to PDF
  • Fold/unfold all
  • Backlinks
  • Show page
  • Old revisions
  • Export to PDF
  • Fold/unfold all
  • Backlinks

    • Differences

    This shows you the differences between two versions of the page.

    Link to this comparison view

    Both sides previous revisionPrevious revision
    Next revision    		Previous revision
    en:dpi:dpi_components:platform:dpi_config [2025/07/22 15:11] – elena.krasnobryzhen:dpi:dpi_components:platform:dpi_config [2025/07/24 11:30] (current) – [dpdk_engine=6: RSS dispatchers per bridge] elena.krasnobryzh
    Line 359: Line 359:
      
     ==== dpdk_engine=6: RSS dispatchers per bridge ==== ==== dpdk_engine=6: RSS dispatchers per bridge ====
    -<note important>This ''dpdk_engine'' is available since version 14.0!</note> +<note important>This ''dpdk_engine'' is available starting from version 14.0!</note> 
    -A separate dispatcher thread is created for each port (device). +This engine is intended for configurations with multiple bridges (dev1:dev2:dev3:...) for 100G+ cards. 
    -Designed for configurations with multiple input and output devices:+
     <code> <code>
     in_dev=41-00.0:02-00.0:c3-00.0:c1-00.0:04-00.0:04-00.1 in_dev=41-00.0:02-00.0:c3-00.0:c1-00.0:04-00.0:04-00.1
    Line 377: Line 377:
     mem_http_parsers=512000 mem_http_parsers=512000
     </code> </code>
    -In this example, six dispatcher threads are created — one per device. Obviously, if there is only one bridge, this engine is equivalent to ''dpdk_engine=1'' — one dispatcher per direction.+This example creates 24 dispatcher threads — 4 dispatchers per bridge.
      
    -<note tip>Total number of required CPU cores equals the number of processing threads plus the number of ports</note>+<note tip>Total number of dispatchers = ''dpdk_rss'' * number of bridges. Starting with version 14.0, the maximum number of dispatchers is 32.</note>
      
    -This engine is designed for configurations with many bridges (dev1:dev2:dev3:...) for 100G+ NICs, replacing the cluster approach. +On-stick devices are supported.\\ 
    -On-stick devices are supported. +SSG configures the cards as follows: 
    -SKAT configures NICs as follows: +  * RX queue count = ''dpdk_rss'' 
    -  * RX queue count = dpdk_rss +  * TX queue count = number of processing threads. Processing threads write directly to their own TX queue on the card.
    -  * TX queue count = number of processing threads. Processing threads write directly to their own TX queue of the NIC.+