MX vs MS Throughput Tests

[dcarney]

Hello,

I have noticed at a few of the properties we manage, that throughput results ran from the MX Security Appliance pull significantly higher throughput results than the MS225 switches do.

Anyone know the reasoning for this? I even tested a property where it is the switch connected directly to the ISP uplink. with the Firewall WAN plugged into said switch. Still, the MX pulled 100mbps of throughput with the switch pulling about 10mbps. 

Is this a limitation of the switches hardware? 

[Adam]

Based on the numbers you provided are you sure there isn't a duplex or negotiation issue somewhere in the mix?  At our core we have some WAN connections connected directly to/through a switch without issues and I get 1G speeds (burstable circuit).  

[ww]

if you mean the test you can start from the dashboard.

its not reliable at the ms for some reason.

mr and mx work fine

[dcarney]

Interesting, thanks for the help

[MerakiDave]

The throughput test under any of the live tools pages is just measuring back-end mtunnel throughput for the control channel between your Meraki device and the Dashboard shard it's connected to back in the data center (cloud).  You can see what the speed test is actually doing (nothing exciting) by running a pcap and looking at all the TCP/7752 traffic.  It's not meant to be an indication of throughput for your clients to/from the Internet, and not necessarily from that actual piece of infrastructure to the Internet.  So for clients you'd use your favorite & regular speed test tools for that. 

There are multiple factors that impact the test results when running a throughput test in the Dashboard, some of which are external/unrelated to your environment.  More on that here: https://documentation.meraki.com/zGeneral_Administration/Tools_and_Troubleshooting/Throughput_test_t...

[dcarney]

Hello, 

Thank you for your help. I do understand this, I am not looking to use it to gauge client speed.I am wondering why the MX Security appliance pulls such higher speed than switches. 

[MerakiDave]

Ah, got it, sorry if I was explaining what you already knew. Definitely not for gauging client speed as you already knew, but also not for even testing true uplink speed from a given piece of infrastructure. Just for others reading/browsing, I'd go so far as to say that the Dashboard throughput test isn't really a "throughput" testing tool the way we'd tend to think of it.

Results are going to vary based on many things including the product, the model and firmware version, naturally the load on the device, which shards they're connected to in which data centers, heck maybe even the time of day and the weather. Just kidding on the last part.

I will make it a point to discuss this with the Dashboard team. I would like the "Throughput" test tool to be renamed to better reflect what it's actually doing... Which I would say is testing the control plane communications to Dashboard.

I didn't write the tool and never saw the code, but I'd imagine there are several "it depends" types of factors. And results can be confusing. For example, I just ran a test on my lab MX65 which got 50Mbps, then an MS120 behind the MX65 which got about 20Mbps, and then an MR42 AP hanging off that switch which got 110Mbps. I have also seen cases where a mesh repeater AP reports higher throughput test results than its corresponding gateway AP.

Anyway, I'm on it, I'll discuss this with the Dashboard PM team at one of the monthly meetings and report back. If it goes several weeks and I haven't checked back, ping me with a reminder.

[dcarney]

Yeah I checked for that. The numbers are an estimate,and I've also replicated this at atleast 3 of my properties. 

I tested from a DMZ switch that is directly connected to the ISP router, linking at 1G FDx, and pulled about 10mbps.

I then tested from the MX Security appliance on the same network, which is using the DMZ switch as an uplink to the ISP, and still pulled significantly higher speeds on the MX. This confuses me as the traffic from the MX goes to the DMZ switch first and then out to the internet. 

[dcarney]

Yeah I checked for that. The numbers are an estimate,and I've also replicated this at atleast 3 of my properties.

I tested from a DMZ switch that is directly connected to the ISP router, linking at 1G FDx, and pulled about 10mbps.

I then tested from the MX Security appliance on the same network, which is using the DMZ switch as an uplink to the ISP, and still pulled significantly higher speeds on the MX. This confuses me as the traffic from the MX goes to the DMZ switch first and then out to the internet.