Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

Method to calculate back pane speed

Hi:

Can any one tell me how to calculate back pane speed of Cisco 6509 series switch with SUP 720.

Regards

sandipani C

1 REPLY
New Member

Re: Method to calculate back pane speed

This is a tough architecture question but I'll take a stab at answering it for you.

The short answer is that the maximum backplane speed is 720 Giga"BITS" per second. But, this is not a true measure for how much "data" can pass through the switch. Because, in that 720 Gb/s is header overhead, control traffic...a bunch of secret stuff...etc

First: It depends on what type of modules you have installed in the 6500 and if they can take advantage of the faster architectures of the switch. If you have all low end blades that do not support any distributed CEF then your max throughput will be 32 Gb/s. If they all take advantage of the dCEF720 then you will get closure to 720 Gb/s. Then you have to look at the throughput capabilities of each blade. Certain less expensive blades only have a single 1 Gb/s connection to the backplane bus but have 48 RJ45 100Mb/s ports totalling 4.8 Gb/s possible traffic (this is called oversubscribing the blade). Oversubscribed blades are fine for end users generally, but not for servers.

Second: Its more accurate to measure how many packets the switch can handle because packets in ethernet constantly vary in size which causes header overhead percentages to vary greatly. For instance, if you have a lot of small voice packets running through your switch then overall real "data" throughput is less because each of those small packets has a seperate header accoutning for a big percentage of the bandwidth in relation to the data it is carrying. On the other hand, if you have a lot of big fat FTP packets that are usually near maximum packet size, then the ratio of header to data is small and the data throughput is greater. So, when you look at the data/sales pages on the switches and routers it will state how many mega packets per second (Mpps) the switch or router can handle.

A. The xCEF720 architecture refers to the Supervisor Engine 720's innovative architecture that combines proven Cisco Express Forwarding packet-forwarding technology with a 720 Gbps auto-negotiating switch fabric. This architecture enables maximum investment protection for the Cisco Catalyst 6500 Series using existing Cisco Catalyst 6500 Series modules, while enabling higher performance modules with local packet-forwarding capability using the full 720 Gbps switch fabric bandwidth.

The architecture type, or taxonomy, allows multiple generations of Cisco Catalyst 6500 Series modules to be quickly categorized and recognized based on the forwarding and switch fabric characteristics of a given module. Cisco Catalyst 6500 Series module price list product descriptions will be updated to reflect this quick and easy taxonomy, which can be summarized as follows:

* dCEF720?Distributed packet forwarding using an interface module local forwarding card on a given interface card, using a 720-Gbps switch fabric. This forwarding technology provides sustained system throughput up to 400 Mpps.

* dCEF256?Distributed packet forwarding using an interface module local forwarding card on a given interface card, using either a 256-Gbps or the 720-Gbps switch fabric. This forwarding technology provides sustained system throughput up to 210 Mpps. The dCEF256 module connects to the switch fabric using dual 8-Gbps connections.

* CEF256?Centralized packet forwarding using the supervisor engine, using the switch fabric module (SFM)/SFM2 256 Gbps of switch fabric bandwidth or the Supervisor Engine 720 Gbps switch fabric. Note that these modules connect to the switch fabric with single 8-Gbps connections plus another connection to the shared 32-Gbps bus. Many modules can be upgraded to support distributed forwarding with the addition of a distributed forwarding card, yielding higher system throughput.

* Classic?Centralized packet forwarding using the supervisor engine, sharing 32 Gbps of backplane bandwidth.

http://www.cisco.com/en/US/partner/products/hw/switches/ps708/products_qanda_item09186a00801eb824.shtml

489
Views
10
Helpful
1
Replies