Published: Saturday, Mar 7, 2015 Last modified: Friday, May 24, 2024

I was curious to know if there was a real difference between the Network Performance between AWS ap-southeast-1a m3.xlarge ( and a ap-southeast-1a t2.micro (

I configured both instances as 64bit Ubuntu since that has the iperf package readily available and on my wired X220 laptop I ran several iperf tests.

Here are some abbreviated typical results:

wired-to-singtel_FG7003GRV$ for i in; do iperf -c $i; sleep 1; done
[  3]  0.0-10.0 sec   648 MBytes   543 Mbits/sec
[  3]  0.0-10.0 sec   601 MBytes   504 Mbits/sec
[  3]  0.0-10.0 sec  64.8 MBytes  54.2 Mbits/sec
[  3]  0.0-10.1 sec  39.2 MBytes  32.5 Mbits/sec

Full log: is the cheapest Singaporean Digital Ocean droplet (sgp1) and is a Singapore based GPLHost instance. Have a look at for details.

My conclusions

There is no difference in Network Performance between the high of m3.xlarge and low to medium of t2.micro at 500Mbps.

Singtel’s 500Mbps is actually 500Mbps… in one test it even went as high as 542Mbps. 👏

Digital Ocean and GPLHost seem limited to 100Mbps and 50Mbps respectively. Though it’s highly variable. Suck!

Sidenote: Somewhat depressingly, since my landlord doesn’t allow me to run cables to the router, I typically get just 70 Mbits/sec using Wireless AC and my Rpi2 with a USB Wireless-N can only seem to achieve 13 Mbits/sec… i.e. my own internal “final 10m” is underperforming!