The Truth about WAN router throughput ?

[Michael Marzol]

Hello,

We have a remote office which we just acquired where they state they are running a 300Mbps circuit on a 3800 series router and are actually achieveing these speeds. In my own experience a short time ago, we tried pushing a 50Mbps circuit through a 1751 router and were absolutely choked at a max of about 6Mbps until we upgraded to a 3800 series. So I began doing some research and came across a document (attached) from Cisco showing throughput comparisons, but this just left me with more questions. For example, our 1751 router has a 100Mbps port to which we connected to but the performance white paper states that this model can Fast/CEF Switch up to a max of 6.4Mbps, which was counterintuitive to me being that I was terminating on a 10/100 port set to 100/Full.....but the white paper seemed to have it right, 6Mbps was the best I could get. Now we take the case of the remote office, the same white paper states that a 3824 can handle a max of 179Mbps and yet they're telling me they have no problem pushing speeds in excess of 300Mbps. I suggested that, according to the white paper and my experience with the smaller 1751, at least a 3925 would be needed to achieve 300Mbps. The argument was, "what are you talking about? Our 3800 has a gigabit port, more than enough to handle 300Mbps !" and in their instance, thay seemed to be correct. I know for sure I'm missing a larger picture here and so was hoping someone could put this to rest once and for all.

Regards,

Doug

[Tagir Temirgaliyev]

Numbers are given with 64 byte packet size, IP only, and are only an indication of raw switching performance. 
These are testing numbers, usually with FE to FE, GigE to GigE or POS to POS, no services enabled. As you add ACL's, 
encryption, compression, etc - performance will decline significantly from the given numbers, unless it is a hardware-assisted 
platform, such as the ASR 1000, 7600 or 12000, which process QoS, ACL's, and other features in hardware (or when a hardware 
assist is installed, for instance an AIM-VPN in a 3745 will offload the encryption from the CPU). 
Every situation is different - please simulate the true environment to get applicable performance 
values. 

for web traffic packet size can be 1000 bytes or even 1500 bytes and so speed will be better

and dont forget to rate post

[Joseph W. Doherty]

Disclaimer

The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.

Liability Disclaimer

In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.

Posting

"Your mileage might vary" ;)

As noted by ttemirgaliyev, the router performance sheet you have is for minimum size Ethernet packets, and is without any additional features.  Basically what's listed is "best case" performance for minimal size Ethernet packets.

Router performance is in packets per second, but not only does that vary with the addition of additional "services", but it often varies (much) based on the packet's size.  PPS requirements decrease as packet size increases which usually allows (much) more throughput.

I've attached a newer Cisco whitepaper that lists newer ISR performances under different usage situations (basically more like "real world").  The older ISRs aren't listed, but the whitepaper will show the large throughput differences based on packet size and service usages.  Take special note of the differences between maximum throughput listed in Table 1 and generic WAN deployment recommendations in Figure 1.

Oh and BTW, if you look at your router performance document, also note some of the performance differences for process switched packets versus fast/CEF switched packets.  Just another example how sensitive software based routers can be to the attributes of the traffic.