Servers inherently have significantly greater bandwidth and connection speeds compared to home connections.
However, if you plan to serve many people in certain regions or even across the world, you may want to check speeds from one place to another.
Just as in the real world, actual physical locations may not matter as much as the transport infrastructure between major providers. Certain types of providers have better peering to certain regions, that have more direct connections.
Bad peering will cause more hops to occur between providers, reducing speed and reliability. This must be factored into your choice of provider.
Ideal Connection Speed Test
To test your connection's maximum speed under the most ideal conditions, use Ookla's Speedtest from the command line.
bench.sh from Freevps.us to test network connection speeds to various major hosting providers around the world. This will help you determine which areas have the best peering, and where to put your US reverse proxy.
wget freevps.us/downloads/bench.sh -O - -o /dev/null|bash
On a Hetzner (DC 11) server:
Output of `bench.sh` by FreeVPS.us System Info ----------- Processor : Intel(R) Xeon(R) CPU E3-1245 V2 @ 3.40GHz Frequency : 3798.304 MHz Memory : 15786 MB Arch : x86_64 (64 Bit) Kernel : 3.10.0-327.4.4.el7.x86_64 Hostname : CentOS-72-64-minimal Speedtest (IPv4 only) --------------------- Your public IPv4 is 184.108.40.206 Location Provider Speed CDN Cachefly 79.3MB/s Atlanta, GA, US Coloat 16.5MB/s Dallas, TX, US Softlayer 16.0MB/s Seattle, WA, US Softlayer 13.5MB/s San Jose, CA, US Softlayer 12.1MB/s Washington, DC, US Softlayer 5.73MB/s Tokyo, Japan Linode 8.08MB/s Singapore Softlayer 6.00MB/s Rotterdam, Netherlands id3.net 61.3MB/s Haarlem, Netherlands Leaseweb 86.0MB/s Disk Speed ---------- I/O (1st run) : 131 MB/s I/O (2nd run) : 127 MB/s I/O (3rd run) : 128 MB/s Average I/O : 128.667 MB/s