One day I decided to deploy a t2.micro instance from Amazon’s EC2 for testing purposes.
I did so because I assumed I would get the 1st. year for free which unfortunately was not the case given that my account isn’t new.
Well that didn’t stop me from at least testing the EC2 instance before nuking it.
For those not familiar with the EC2 environment, their t2.micro represents a virtual environment with 1 vCPU and 1GB or RAM as you can see in the table below:
If we are to check out the pricing for the t2.micro instance we’ll see a pricing of $0.013/hour which means $0,312/day and $9.36 per month which is kinda expensive if we consider the volume of resources provided and also cross compare with ArubaCloud’s small VPS which goes at 1 euro per month.
But hey, being Amazon you’d imagine that the price is in the end reflected by the performance. Not?
Not quite because the benchmark I took isn’t impressive at all:
Processor : Intel(R) Xeon(R) CPU E5-2676 v3 @ 2.40GHz
CPU Cores : 1
Frequency : 2394.490 MHz
Memory : 990 MB
Swap : MB
Uptime : 1 min,
OS : Ubuntu 16.04.1 LTS
Arch : x86_64 (64 Bit)
Kernel : 4.4.0-47-generic
Hostname : ip-172-31-45-101
Speedtest (IPv4 only)
Your public IPv4 is 188.8.131.52
Location Provider Speed
CDN Cachefly 37.7MB/s
Atlanta, GA, US Coloat 4.00MB/s
Dallas, TX, US Softlayer 3.14MB/s
Seattle, WA, US Softlayer 2.71MB/s
San Jose, CA, US Softlayer 2.66MB/s
Washington, DC, US Softlayer 5.59MB/s
Tokyo, Japan Linode 5.07MB/s
Singapore Softlayer 1.93MB/s
Rotterdam, Netherlands id3.net 5.59MB/s
Haarlem, Netherlands Leaseweb 8.74MB/s
I/O (1st run) : 67.8 MB/s
I/O (2nd run) : 64.4 MB/s
I/O (3rd run) : 64.4 MB/s
Average I/O : 65.5333 MB/s
Now if you add to the discussion the fact that others have performed different benchmarks and confirmed the issue with Amazon’s EC2 instances, you’ll come to the conclusion that the price they are asking for it isn’t quite worth it given that at the same money you can get true bare metal from Scaleway, Kimsufi or Online.net.