From Sao Paulo to Savings: GEMA Achieves Performance and Cost Savings with AWS


In today’s dynamic cloud landscape, maximizing the value of your AWS investment is crucial for staying competitive and efficient. That’s where 3HTP Modernization and Optimization Assessment comes in. This service comprehensively analyzes your AWS infrastructure, resources, and services, delving deep into your account to identify optimization opportunities. Our team of experts, including seasoned AWS architects, conducts a live analysis session with you to ensure a personalized and thorough evaluation.

This document summarizes the activities and benefits that GEMA, along with many others, achieved by executing the activities with the help and expertise of our team. Don’t let inefficiencies drain your resources. Let our AWS Resource Optimization Assessment service help you unlock the full potential of your AWS environment, improve efficiency, and reduce costs.

About the customer

GEMA a Chilean web solutions company migrated their AWS services from São Paulo to Virginia for cost & performance gains. 3HTP analysis identified optimization areas, including migrating all the resources to another region, optimizing EC2/RDS instances, and implementing CloudFront for static content. Leveraging economies of scale in Virginia reduced transfer costs & enhanced content delivery for GEMA’s user base.

Customer Challenge

GEMA deployed most of its AWS services in the São Paulo region (sa-east-1). This decision was driven by the perception that proximity to Latin America would result in lower costs and reduced latency compared to regions in the United States.

At the time of the 3HTP optimization report, GEMA’s AWS environment comprised 138 AWS resources, with primary billing expenses in Data Transfer, Amazon S3, Amazon EC2 instances, and Amazon RDS databases. 

3HTP AWS Optimization Project

The following high-level architecture comprised the technical solution implemented with GEMA:

The following are the activities defined in the 3HTP optimization assessment which leads us to execute the project in three phases:

Migration of Compute and Database resources

We start by creating a new landing zone in North Virginia region by leveraging Terraform as an Infrastructure as Code(IaC) tool which helps us quickly deploy the new North Virginia Amazon VPC and all networking resources (including subnets, route tables, NAT gateways, security groups, etc)

A two-wave migration was decided with GEMA’s team to ensure the high-priority instances will be migrated in the first wave. The following activities were performed to execute the migration:

Initial Sync and Replication of Amazon EC2 Instances:

The replication process was executed by utilizing AWS Application Migration Service (AWS MGN), the AWS MGN replication agent was installed and configured on each instance, the process was monitored in the AWS MGN dashboard in the Virginia region by confirming the initial sync of every source server.

Test of replicated Amazon EC2 instances:

AWS MGN helps us perform non-disruptive tests before cutover by quickly deploying test instances in a public subnet in the VPC Staging Area. A series of tests were executed in the testing VPC with the help of the GEMA technical team to ensure that the replicated instances could start and run the applications or processes deployed in every instance.

Cutover of replicated Amazon EC2 instances

When every Amazon EC2 was tested maintenance windows were requested and together with GEMA, the cutover process was applied to every one of the Amazon EC2 instances. This process was executed in two waves. This allows GEMA to migrate transparently to the region with minimal downtime.

Migration of Amazon RDS databases

Amazon RDS snapshot copy functionality across AWS regions was the method chosen to execute the migration process of Amazon RDS instances between regions, this method satisfied GEMA requirements and led us to recreate Amazon RDS in Virginia in a short maintenance window.

Amazon S3 Data replication

Amazon S3 Cross-Region Replication (CRR) was used to copy GEMA’s objects to Virginia. A lifecycle rule was applied in the source bucket to erase all objects when the replication finished.

Resource Optimization

The following activities were performed in this phase:

Amazon EBS optimization

Some Amazon EBS was running in provisioned IOPs (io1) configured using a small number of IOPs comparable to or inferior to general-purpose type volumes. Amazon EBS volumes were replicated via AWS MGN, the volumes were updated to use general-purpose last version (gp3) volume types which reduced costs and improved maximum throughput per volume.

Amazon EC2 Instance Type and Right-sizing Optimization

All instances were reviewed and tuned to their right size based on AWS Cloudwatch compute and memory resource usage, the t2 instances were upgraded/changed to the t3 family to reduce costs and upgrade to the last generation.

Amazon Cloudfront implementation

An Amazon CloudFront distribution was implemented to reduce AWS Data transfer costs associated with consuming the image files stored in AWS buckets

Result and benefits

  • AWS Application Migration Service (MGN) helps plan and execute an Amazon EC2 instance migration between regions with a centralized platform for application testing and cutover activities this workflow helps GEMA achieve minimal downtime.
  • The implementation of a Cloudfront distribution in front of WordPress deployment enhances the performance, cost-effectiveness and reduces the server load by offering streamlined integration between images hosted on Amazon S3 and the origin.
  • Amazon EC2 and Amazon RDS instance right-sizing improve GEMA instances performance by upgrading to last-generation instances.

AWS Optimization project helps GEMA achieve the following savings in their monthly bill:

  • ~ 49% Cost savings on Amazon EC2 Instances and Amazon RDS 
  • ~ 15% Cost savings on data transfer and content delivery 
  • ~ 8% Cost savings on Amazon EBS volumes