Here's what I'm trying to achieve and the steps I've taken so far:
I have a large One Drive roll out about to happen @ 500 branch offices and each user data upload set will be unique. These offices have a slew of upload pipe's from 3 to 50 Mb/s up. My thought was that I could set a shaping rule for One Drive to Priority to "Low" and that would let all other traffic (which I've read is marked "Normal") have more bandwidth available.
This doesn't work as I believe it should.
Details about my config -
200/7 cable connection
MX64/MS-225-48/MR18 test lab
Down/Up speeds are set and accurate
The dashboard see's this traffic as "Sharepoint" not sure why.. outdated definitions?
Policy is set on MX device
I know its correct due to hard bandwidth limits working perfectly
The test I perform is
-start uploading a file in a browser window
-start large file sync with one drive
Monitor Task Manager for each programs usage
Doesn't matter which one I start first often the One Drive upload will consume more bandwidth than the browser upload
Am I missing something here?
I would say get a packet capture of the One Drive app uploading and adding the IP(s) that is it communicating with. I wonder if the app is not getting categorized the same way as the browser upload.
Also what is the menu path you went through to get "Sharepoint"? I am able to specifically select OneDrive.
I can select one drive also but that's now how the device see's the traffic. It's identified as Sharepoint for reasons beyond me. The rule I have works if I select specific bandwidth limits so I know that it's seeing the traffic correctly.
Ok. The way the "High, Normal, and Low" traffic shaping settings work is by giving a fraction of the flows to a specific application.
High: 4/7 flows
Normal: 2/7 flows
Low: 1/7 flows
I'm not sure how much traffic is on your network, so is it possible the proportion of OneDrive traffic was considerably higher than other traffic?
Also I wonder if the OneDrive app initiates multiple flows of traffic when syncing. So in theory it 4 simultaneous traffic flows vs a single traffic flow that the browser might do. So when traffic shaping rules see that traffic it ends up giving the OneDrive app four times as much network flow.
We are in the same boat here just flipped on down/up. I was expecting the same result but see similar findings as you. I've opened a ticket with Meraki about possible solutions.
>The dashboard see's this traffic as "Sharepoint" not sure why.. outdated definitions?
@ciph3r OneDrive for Business is stored in your Sharepoint site (as sub-sites, just hidden from your view). So that is the reason for that one.
>I was under the impression that I would see the Ubunutu ISO pull 6Mbps, and the YouTube pull 1Mbps
@NolanHerring that QoS is for traffic leaving the site (or being uploaded). So seeing the downloaded traffic being split like this would be expected.
@ciph3r a better test might be to run something like www.speedtest.net in your browser during the OneDrive for Business upload, and monitor the upload performance. The upload performance should only be affected in a small way.
@PhilipDAth - You're right about the one drive for business.
If you look at the traffic for my client (first post) the second entry down is a speed test site that allows you to pick file sizes to test with. I picked a larger file which gave me time to see results.
Here is a cap of the behavior in action. This is Chrome pushing a file upload while OD is syncing a new folder.
That's right, in the Meraki documentation for that it specifies uplink.
And yeah, you would need to specify that bandwidth limit for the download.