Will the real router performance numbers please stand up?

[patrick.hurley]

We have this document related to router performance on Cisco's Website.

http://www.cisco.com/warp/public/765/tools/quickreference/routerperformance.pdf

We have this report from Miercom for an 1841.

http://www.miercom.com/dl.html?fid=20040904&type=report&PHPSESSID=6cd0835d99e8f7695c21a5868d311951

The difference between what is reported in the first .pdf as performance and the second .pdf as performance for the 1841 router is a decrease of 95% of the rated router performance as stated in the first document.

How are we to provide accurate information to customers regarding the capabilities of Cisco products without having the actual performance ratings of each platform?

[paolo bevilacqua]

I do not see much of a contradiction but different testing parameters.

Mier tested 2 Mbps full duplex with 3DES, Qos and other services enabled, Cisco do not provide this figure directly.

Then Mier tested with no features and got an aggregate of 130 Mbps with big packets.

As in Cisco tradition they tested with the smaller packet size, no features, and hit the ceiling at 75 Kpps, or 38 MBps.

[mihanlin]

Hello Patrick,

Your questions: "How are we to provide accurate information to customers regarding the capabilities of Cisco products without having the actual performance ratings of each platform?" can be answered by the Cisco PDF you sourced, which states:

"Every situation is different - please simulate the true environment to get applicable performance values."

The fact is that the Cisco document is a guide of raw routing performance, usually from Ethernet to Ethernet interfaces. As soon as you add encryption, ACLs, QoS, PBR etc then this will naturally affect performance.

If you are looking for information on devices to recommend to your customers, please engage Cisco sales who will be able to help further to determine specific conditions for your customers.

Hope this is of assistance.

Kind Regards,

Michael.

Cisco TAC.

[patrick.hurley]

I guess what I am saying that the figures are only raw performance figures. If you added the worst case for each platform, it would be easier to choose the right platform.

The other gentleman that answered doesn't have a realistic response. I can't personally purchase each platform to model performance as a pre-sales engineer nor as the document states, simulate performance. I have to go by what information I receive from Cisco.

If you added the scenario for each platform with all features running and the performance that can be expected then its easier for me to provide guidance to my customers.

I hope this is of assistance to Cisco.