How to Start Instance Selection (and optimal sizing)

In this article, we will show you how to start an instance (and choose an instance size) through the AWS Marketplace. 

EC2 Region Selection

EC2 is being hosted in multiple locations globally called, “Regions”.

  • US East (Virginia)
  • US West (Oregon)
  • US West (Northern California)
  • EU West (Ireland)
  • Asia Pacific (Singapore)
  • Asia Pacific (Sydney)
  • Asia Pacific (Tokyo)
  • South America (Sao Paulo)

You must select the region prior to initiating an EC2 instance.

For example you would like to create a site with Tokyo viewers as your target, and you are based in USA, you may choose the “Tokyo” Region (closest region) for faster transfer rates.


Users who live in Japan, the Asia Pacific (Tokyo) region is recommended.


EC2 Instance Types

There are multiple EC2 instances, and they are referred to as “family” that is broadly categorized by its performance.

Small CPU resources: T1* 

The t1 instance has a free trial period and you may use this for a blog site, some small management applications, bastion host, and sites with low traffic.

General purpose: M1*, M3

For medium-sized databases, data processing, encryption, caching, SAP enterprise applications (of Microsoft SharePoint), and other small scale apps.

Computing optimization: C1, CC2

For high traffic web application, ad delivery, batch processing, video encryption, analysis of variance, high-energy physics, genome analysis, and computational fluid dynamics.

Optimize memory: M2, CR1 

Deploy a large-scale enterprise application analysis, high-performance database, distributed memory cache, memory analysis, and genome assembly, SAP, of Microsoft SharePoint, etc.

Storage Optimization: HI1*, HS1

A common usage case of HI1: transaction database scale-out and NoSQL databases, including Mongo B and Cassandra.

A typical use case of HS1: data warehouse, Hadoop, and cluster file system.

GPU instance: G2, CG1*
A typical usage case of G2: Game streaming, graphics workload of the server side of the 3D application streaming, and others.
The usual usage case of CG1: computational chemistry, rendering, financial modeling and engineering design.

The vCPU, ECU and memory ratios are determined by the instance family: “micro”, “small”, “medium” and “large”.

Though T1, M1, C1, CC2, M2, CR1, CG1, and HI1 are previous generation instances, they are still supported by AWS. However, for best performance, choose the latest generation of instances. 

Below you may check the vCPU, instance family, ECU, memory, and so on:

OSbit vCPU ECU Memory (GiB) Instance
T1 t1.micro 32-bit
One Variable 0.615 Only EBS The very low
M1 m1.small 32-bit
One One 1.7 1 x 160 Low
M1 m1.medium 32-bit
One Two 3.75 1 x 410 Medium
M1 m1.large 64-bit Two Four 7.5 2 x 420 Yes Medium
M1 m1.xlarge 64-bit Four Eight Fifteen 4 x 420 Yes High
M2 m2.xlarge 64-bit Two 6.5 17.1 1 x 420 Medium
M2 m2.2xlarge 64-bit Four 13 34.2 1 x 850 Yes Medium
M2 m2.4xlarge 64-bit Eight 26 68.4 2 x 840 Yes High
M3 m3.xlarge 64-bit Four 13 Fifteen Only EBS Yes Medium
M3 m3.2xlarge 64-bit Eight 26 Thirty Only EBS Yes High
C1 c1.medium 32-bit
Two Five 1.7 1 x 350 Medium
C1 c1.xlarge 64-bit Eight Twenty Seven 4 x 420 Yes High
CC2 cc2.8xlarge 64-bit 32 88 60.5 4 x 840 10 Gigabit
CR1 cr1.8xlarge 64-bit 32 88 244 2 x 120 SSD 10 Gigabit
HI1 hi1.4xlarge 64-bit Sixteen 35 60.5 2 x 1,024 SSD 10 Gigabit
HS1 hs1.8xlarge 64-bit Sixteen 35 117 24 x 2,048 10 Gigabit
G2 g2.2xlarge 64-bit Eight 26 Fifteen 1 x 60 SSD Yes High
CG1 cg1.4xlarge 64-bit Sixteen 33.5 22.5 2 x 840 10 Gigabit




Which instance should I choose?

For your first blog, you may choose the free trial, so we can start with a t1.micro instance.

A general corporate site should be alright with t1.micro because corporate sites usually have a monthly PV of 100,000 PV~30 (about 10,000 as a guide).

On a blog media with a monthly PV of 500,000 PV~100 (about 10,000 PV as a guide), you can choose the m1.small or m1.medium instance.

It is possible to optimize gradually while checking in the cloud, etc. Watch the CPU usage for the first time from a small instance (especially when the optimal instance is unknown to the site).

Furthermore, it’s also possible to scale the instance up such as during a high web traffic (so the images will load quickly). And when the web traffic returns back to normal, you can scale your instance down.

The AWS usage is not a monthly fee, but is determined by the amount of time spent.


If you create a new account in AWS, the t1.micro + AMIMOTO entitles you to a free trial (up to 750 hours per month for the first year).

Usage is calculated at the 30th of the month and let’s say that you only used up 720 hours this month, then you can plan to run multiple instances for the remaining hours.


How to Start

At the “Amimoto Self-Hosting Plan”, you can choose the region where you’d like to launch the instance. For example, “Tokyo” region:

Screen Shot 2014-10-22 at 13.46.30


Once you click “Launch AMI”, you will be redirected to AWS Marketplace (login required).

Detailed settings can be made at the “1-Click Launch” tab. For more info on how to use, check this guide and continue.


Please note that the data transfer fee and EBS fee are paid for separately. In addition, the cost varies on the location.

At the AWS Marketplace right hand tab, you may see the estimated cost like this:

For instance t1.micro Asia Pacific at (Tokyo) Region

For t1.micro Asia Pacific (Tokyo) Region.

For instance m1.small US East in (Virginia) Region

For m1.small instance US East (Virginia) region.