Solved: 6500 Backplain speed's

ccannon88567 · ‎06-30-2008

Could somebody please clarify;

I am unsure of the meaning of - "6500 has multiple performance options including 32-Gbps, 256-Gbps and 720-Gbps bandwidth options with system throughput of 15 million packets per per second, 30 mpps, 210 mpp or up to 400mpps."

1) Does this mean the backplane switching capacity for switching between local interfaces and line cards?

2)Is the performance option the same as the system throughput?

ie - Is the 32-Gbps option the same as 15 Mpps or the 720 Gbps the same as 400Mpps??

Any info would be much appreciated.

a.alekseev · ‎06-30-2008

http://www.cisco.com/en/US/prod/collateral/switches/ps5718/ps708/prod_white_paper0900aecd80673385.html

Cisco Catalyst 6500 Architecture: Bus Switching Modes

There are three switching modes used by the BUS and fabric ASICs present on CEF256 and CEF720 line cards. These modes determine the header format that is used to transmit the data across the DBus and communicate with other CEF256 and CEF720 line cards. These modes do not apply to line cards that use a DFC. These modes are discussed in more detail below.

Flow-Through Mode

This mode of operation is used by CEF256 modules when there is no crossbar switch fabric present. It enables CEF256 modules to operate as if they were classic line cards. In flow through mode, the entire packet (header plus data) is forwarded by the line card to the supervisor for processing. This mode does not apply to CEF720 modules. When flow through mode is used, performance levels up to 15Mpps can be achieved.

Compact Mode

The compact mode of operation requires a crossbar switch fabric to be present in the system. This can be provided by a switch fabric module, Switch Fabric Module 2 or a Supervisor Engine 720. All modules in the chassis must be fabric enabled (i.e. CEF256, dCEF256, or CEF720) for the switch to operate in Compact Mode. Classic line cards installed in the chassis will negate the ability of the switch to run in Compact Mode. In Compact Mode, only the header is passed over the DBus to the supervisor. The Header is compressed prior to being placed on the DBus which increases the bandwidth available for header transmission. The data portion of the packet is transmitted over the crossbar switch fabric channels. In this mode of operation, the switch can achieve centralized performance of up to 30Mpps independent of packet size.

Truncated Mode

Truncated Mode is used when CEF256 and/or CEF720 line cards are installed in a chassis with a classic line card, but a crossbar switch fabric is present. In this mode, classic line cards will transmit both the header and the data portion of the packet over the DBus. CEF256 and CEF720 line cards will only transmit headers over the DBus and transmit the data portion of the packet over the crossbar switch fabric.

Truncated mode results in centralized forwarding rates up to 15Mpps. In Truncated Mode, since the CEF256 and CEF720 line cards use the crossbar switch fabric to transmit data, overall aggregate bandwidth can actually be higher than the 32-Gbps shared bus capacity.

Performance of DFC enabled line cards are not affected by truncated mode and performance remains the same regardless of the line-card mix in the chassis.

View solution in original post

a.alekseev · ‎06-30-2008

Cisco Catalyst 6500 Series Switching Architectures

Cisco developed the following switching architectures for Cisco Catalyst 6500 Series modules to allow platforms to scale in any deployment:

â€¢ 32-Gbps Bus-Allowing access to a central shared bus

â€¢ 256-Gbps Switch Fabric-Located on the switch fabric module (SFM)

â€¢ 720-Gbps Switch Fabric-Located on the Cisco Catalyst 6500 Series Supervisor Engine 720

Cisco Catalyst 6500 Series Modules

Cisco Catalyst 6500 Series interface modules support the following forwarding technology and switch-fabric combinations.

â€¢ Classic Interface Modules-Use the centralized Cisco Express Forwarding engine located on the supervisor engine's PFC, connect to the 32-Gbps switching bus only, and forward packets at up to 15 mpps

â€¢ CEF256 Interface Modules-Use the centralized Cisco Express Forwarding engine located on the supervisor engine's PFC, connect to both the 256-Gbps fabric located on the supervisor engine with a single 8-Gbps full-duplex fabric connection and the 32-Gbps switching bus, and forward packets at up to 30 mpps

â€¢ dCEF256 Interface Modules-Use the distributed Cisco Express Forwarding engine on the DFC (located on the interface module), connect to a 256-Gbps fabric located on the supervisor engine or a switch fabric module with 16-Gbps full-duplex fabric connections, and forward packets at up to 210 mpps

â€¢ dCEF720 Interface Modules-Use the distributed Cisco Express Forwarding engine on the DFC3 (located on the interface module), connect to the 720-Gbps fabric located on the supervisor engine with dual 20-Gbps full-duplex fabric connections, and forward packets at up to 400 mpps of sustained performance

ccannon88567 · ‎06-30-2008

Thanks for the stat's but I already have these. I am trying to find out if my understanding of the stats presented are correct?

Could somebody please clarify;

****** I am unsure of the meaning of ******

- "6500 has multiple performance options including 32-Gbps, 256-Gbps and 720-Gbps bandwidth options with system throughput of 15 million packets per per second, 30 mpps, 210 mpp or up to 400mpps."

1)***** Does this mean the backplane switching capacity for switching between local interfaces and line cards?

2)***** Is the performance option the same as the system throughput?

ie - Is the 32-Gbps option the same as 15 Mpps or the 720 Gbps the same as 400Mpps??

Any info would be much appreciated

a.alekseev · ‎06-30-2008

http://www.cisco.com/en/US/prod/collateral/switches/ps5718/ps708/prod_white_paper0900aecd80673385.html

Cisco Catalyst 6500 Architecture: Bus Switching Modes

There are three switching modes used by the BUS and fabric ASICs present on CEF256 and CEF720 line cards. These modes determine the header format that is used to transmit the data across the DBus and communicate with other CEF256 and CEF720 line cards. These modes do not apply to line cards that use a DFC. These modes are discussed in more detail below.

Flow-Through Mode

This mode of operation is used by CEF256 modules when there is no crossbar switch fabric present. It enables CEF256 modules to operate as if they were classic line cards. In flow through mode, the entire packet (header plus data) is forwarded by the line card to the supervisor for processing. This mode does not apply to CEF720 modules. When flow through mode is used, performance levels up to 15Mpps can be achieved.

Compact Mode

The compact mode of operation requires a crossbar switch fabric to be present in the system. This can be provided by a switch fabric module, Switch Fabric Module 2 or a Supervisor Engine 720. All modules in the chassis must be fabric enabled (i.e. CEF256, dCEF256, or CEF720) for the switch to operate in Compact Mode. Classic line cards installed in the chassis will negate the ability of the switch to run in Compact Mode. In Compact Mode, only the header is passed over the DBus to the supervisor. The Header is compressed prior to being placed on the DBus which increases the bandwidth available for header transmission. The data portion of the packet is transmitted over the crossbar switch fabric channels. In this mode of operation, the switch can achieve centralized performance of up to 30Mpps independent of packet size.

Truncated Mode

Truncated Mode is used when CEF256 and/or CEF720 line cards are installed in a chassis with a classic line card, but a crossbar switch fabric is present. In this mode, classic line cards will transmit both the header and the data portion of the packet over the DBus. CEF256 and CEF720 line cards will only transmit headers over the DBus and transmit the data portion of the packet over the crossbar switch fabric.

Truncated mode results in centralized forwarding rates up to 15Mpps. In Truncated Mode, since the CEF256 and CEF720 line cards use the crossbar switch fabric to transmit data, overall aggregate bandwidth can actually be higher than the 32-Gbps shared bus capacity.

Performance of DFC enabled line cards are not affected by truncated mode and performance remains the same regardless of the line-card mix in the chassis.