I did what you said:
> You can easily verify this by sending a file from one PC to another and then checking on the VZW usage reports when they update.
MiFi 5510L Ver 1.1 (L1.34.23 M1.34.20). The only real difference is I force Data Usage to update before and after testing by turning the Jetpack off, waiting a minute, turning it back on, and after it comes back up, read Data Usage. Because of the lousy detail of Data Usage, files transferred between computers should be large enough to change tenths of a gigabyte by a couple digits. On the most recent test, I just grabbed the first five OEM provided Windows 7 driver files I have in the source drive, with the mouse and dropped them into the target drive. 104 MB in files, 135 seconds with 0.08 GB data usage.
The first time I encountered this was approximately 2 years ago. I confirmed it several times with my 4510 and then with my 5510 along with clean installs of Ubuntu and Windows with minimal drivers. I have also done it at the beginning of my cycle as the very first thing to eliminate the possibility of delayed usage showing up.
Alright I did some testing to verify what I have always claimed to be true.
Enable file sharing on two PCs connected with a MiFi 4510L (Router FW 02.01, AP FW Version 2.10-18.104.22.168, Modem FW version 2.28). I left the internet connection enabled and also enabled MAC filtering to ensure I didn't get any other random machines to jump on and tamper with my results. I also installed Glasswire since I was working with two Win7 laptops. Glasswire can show data usage and break it down by application which is very handy for a situation like this. Both machines were allowed to finish auto updates from a separate line before conducing the test to increase the effectiveness of this scenario.
I transferred a folder of installation files (the same stuff you get from the full software solution for Win7 64Bit on support.hp.com) for my HP CM1312 printer from the C:\temp drive of one machine to another. This folder was rated at 427 MB according to the properties in Windows Explorer. I monitored the data transfer with Glasswire to confirm that data was moving and benchmarked its progress to compare against. This test was conducted 2x.
2:00 - Power up Jetpack and VZW data usage says 0.17GB from verizonwireless.com
- Allowed the Jetpack to sit for a while and confirm no other shenanigans were going on
2:45 - File transfer begins
- Glasswire confirms movement, lists the local data as a "System" application
- Glasswire shows data transfer speeds of 1 MB/s (Note this is not Mbps)
- Windows estimates about 8 minutes to complete the transfer
2:55 - Glasswire shows data transfer is complete
- Glasswire shows the System application has sent a total of 437.6 MB (I take it there is a little overhead)
2:57 - Reboot the Jetpack
2:58 - VZW.com posted one data session of .017 GB back from 1:32 PM which accounts for what was originally displayed
3:00 - Allow the devices to sit and observed any other traffic
3:21 - File transfer round 2 begins
- Glasswire starts recording and all stats look similar to before
3:29 - File transfer is complete
- Glasswire shows a total of 431 MB have been sent to the new PC (records updated because of the reboot to the Jetpack earlier and the IP Addresses actually reversed due to DHCP)
- Jetpack powered down
7:50 - Next morning - VZW.com reports .018 GB usage for this Jetpack
- Data details show 5 total sessions each with negligible data
|02/26/2015||01: 32 PM||DATA ONLY SHARE||0.0179230393 GB|
|02/26/2015||01: 52 PM||DATA ONLY SHARE||0.0001893267 GB|
|02/26/2015||02: 16 PM||DATA ONLY SHARE||0.0000059269 GB|
|02/26/2015||02: 20 PM||DATA ONLY SHARE||0.000033224 GB|
|02/26/2015||02: 24 PM||DATA ONLY SHARE||0.0001508956 GB|
So there you have it sports fans. I should have racked up around 1GB of data from my one test in a matter of 16-20 minutes of actual usage time. VZW records do not show anything suggesting that my local data was getting accounted against me. The small data sessions that were recorded were accounted for by Glasswire as random IE downloads and policy updates.
I do have access to more Jetpack models and would be happy to conduct the same experiment with them if we feel this is a hardware model or firmware issue. I encourage and invite anyone else to conduct the same experiment on their own and compare results against me. As long as you are using Glasswire or another data monitor of some kind to verify the activity is local or not it should be a valid test.