So someone had this question. It’s a good one. I didn’t know the answer so I found out….
Articles here and here would indicate that network traffic should be both intermittent and light. So….lets check it out.
The Test: I built a Windows 8 VM in my lab. Updated it, updated default Windows Store Applications, created a new user account that I had never logged in as and then setup a netsh trace session from an elevated command prompt:
mkdir c:\trace cd c:\trace netsh trace LAN capture=yes report=yes tracefile=c:\trace\file.etl
netsh trace LAN capture=yes report=yes tracefile=c:\trace\file.etl
And then I logged out and logged into the new ID. The plan was to run it for 20-30 minutes but I went downstairs to talk to the wife and you know how that goes so an hour or so later I remembered my test! Went upstairs and logged out, logged in as my administrator user and opened an elevated command prompt and did:
cd c:\trace netsh trace stop
netsh trace stop
It looked like this:
Which isn’t horribly exciting, but the etl file and file.cab file are
So I put them on my debugging machine via SkyDrive and went to town.
Very little traffic, most of it in spurts, every 5 minutes as expected.
You can see it in the graphic above. I’ll actually need to do a longer trace to get a better feel for statistics, but I wouldn’t sweat a Windows 8 RT or Pro device on your network.
How can i know the impact only in trace?
is not safe...
The question that got me looking into this in particular was how many proxy connections does a single Windows RT device create. From looking at a default configuration of live tiles updating, it appears to be one session from the WNS.
The debate will shift then to how each application impacts performance of the network I suppose.
Nice post! The proof is in the puddin'.