1

My office has a 150Mbps synchronous fiber, burstable to 250Mbps.

Based on a document I found, my Cisco 1941 is capable of 153Mbps MAX:

http://www.cisco.com/web/partners/downloads/765/tools/quickreference/routerperformance.pdf

I was seeing these speeds on average 50Mbps, with bursts every now and again for 100-300Mbps; starting 2 weeks ago, I was barely able to get 16Mbps. I have since rebooted all my switches and the Cisco 1941, but I'm still not able to get more than 60Mbps.

How do I determine if my Cisco 1941 is performing properly? How do I verify I'm getting maximum throughput on my router? Why is my Cisco 1941 not able to perform higher than 60Mbps consistently?

UPDATE 2016-06-21

I'm testing performance using an ISP provided performance tool. Connecting directly to the ISP provided router/gateway, I see 80Mbps UP/140Mbps DOWN. Behind my own router, I see less than 40Mbps UP and less than 50Mbps DOWN.

I have flow monitors watching incoming and outgoing traffic as well, which confirm what I see behind my router.

My ISR acts as a firewall, NAT/port-forwarding and SSL VPN server.

  • What is the router doing? Be as precise as you can. Do you know which switching paths you are using? Are you doing NAT? IPSec? Firewalling? – David Schwartz Jun 22 '16 at 0:55
  • How are you testing speeds? Is this to some kind of internet-based speed test? – Tedwin Jun 22 '16 at 0:56
  • What have you done to see what is happening during perceived periods of low throughput? What services are you running? What is the CPU utilization? Are packets being punted to process switching, or are they being CEF switched? Is someone hogging the bandwidth? And so on. Cisco provides many tools in IOS to check the performance. Actually, 60 Mbps throughput for a 1921 is really not that bad, depending on the services running. – Ron Maupin Jun 22 '16 at 2:13
2

Found the problem. It was buried in the policy-map for class class-default. There was a police for capping traffic at 60000000 bytes put in by a previous IT group that's no longer with the company.

My assumption is they were trying to conserve some bandwidth for some reason, but instead of segregating traffic with VLANs and QoS, they went this route. Changed the cap to 150000000 until I can build a proper solution.

Thanks for your assistance everyone.

| improve this answer | |
1

Did you read the document you quote? It quite clearly states: Knowing the performance for a specific router platform is not a good indication of how well a specific feature will perform. If a feature is supported in the CEF path, for instance, and we know the feature-free CEF throughput in a specific configuration, then we only know the platform's "never-to-exceed" performance but we do not know the actual performance of any given feature, which will always be less

You say you are running firewall, NAT and VPN so you shouldn't expect the max pps quoted.

The problem is that unless you know what the traffic profile was before it went slow, and what the traffic profile is now, then you won't be able to tell. if for example you are accessing servers with a higher RTT with TCP or the ISP has increased your latency for other reasons then nobody can answer this question accurately. It is interesting that you were bursting to 300Mbps but now you get 140Mbps without the router in place. It is to be expected that the router will add some latency and thus affect maximum throughput.

If you want to know what impact the router is having, then you should first capture network traffic both sides and measure the latency of the device and check out whether it interacts with TCP or other protocol in any particular way.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.