I live on a campus in the US and I am trying to understand why my upload speed to the online backup service Backblaze on area Y of the campus is 10 times slower than on area X of the campus (20000 vs 2000 kpbs as you can see in the details below).
Area X on campus (Ethernet):
- Backblaze speedtest: Download Speed: 48764 kbps (6095.5 KB/sec transfer rate); Upload Speed: 25035 kbps (3129.4 KB/sec transfer rate)
- speedtest.net speed test: Download Speed: 82460 kbps; Upload Speed: 71640 kbps
- Backblaze actual upload speed using the client: around 20000 kbps
Area Y on campus (Ethernet):
- Backblaze speedtest: Download Speed: 50500 kbps (6312.5 KB/sec transfer rate); Upload Speed: 3249 kbps (406.1 KB/sec transfer rate)
- speedtest.net speed test: Download Speed: 94790 kbps; Upload Speed: 50880 kbps
- Backblaze actual upload speed using the client: around 2000 kbps
I have similar results for wifi but I think that's enough details.
The network administrator suspects that Backblaze is doing IP-based throttling. This surprised me as
- the Ethernet's IP is pretty similar between area X and Y
- the Wifi's IP is pretty similar between area X and Y
- I don't see what Backblaze would target area Y and not Y, since both are on the same campus. Also Backblaze vehemently denies doing any IP-based throttling, and I couldn't find anyone complaining about it.
My feeling is that area Y "as well connected" as X to the outer Internet, but most of the web services work fine, which gives credit to my network administrator's explanation.
Hence my question: is there any way to detect whether an online backup service is doing IP-based throttling, beyond comparing the upload speed with different IPs?
EDIT: Some details:
- all tests where carried out using the same laptop.
- I disabled the throttling in the Backblaze client's options.