Freefuse Cloud Journey on AWS

With careful planning of the solution design and successful POCs , SquareOps came up with a solution to achieve the desired goals

The solution involved:

1. Infrastructure as a code : Using Terraform, infrastructure code for existing resources like VPC, S3, policies, etc., was generated. This code was then extended using a mix of Terraform modules provided by AWS and our own. Our modules were developed to comply with CIS benchmarks and were tested with Prowler scans before being published.

2. Modifying AWS EC2 Autoscaling Group Module: The official Terraform module for AWS EC2 Autoscaling groups was modified to handle the challenges posed by blue-green deployments. This included:
   • Adding a lifecycle policy to the resource autoscaling_group. This policy allowed the creation of a new Autoscaling Group (ASG) before destroying the old one and ignored changes to the name tag.
   • Disabling the use_name_prefix while calling the autoscaling group module. This prevented the Terraform ASG module from creating an autoscaling group with random names.
   • Creating a unique tag id for the autoscaling group’s name, controlled by the Terraform format(“%s-%s”, local.unique, “asg”).
   • Fetching the name of the new autoscaling group created by CodeDeploy using Terraform’s data block data.aws_autoscaling_groups.frontend.names.

3. CodeBuild Pipeline Development: A CodeBuild pipeline was developed to deploy and manage Terraform changes. A script was also used in the pipeline to refresh the Terraform state before planning Terraform changes.

4. Parallel Video Processing: We integrated AWS Simple Queue Service (SQS) into the architecture. The AI video editor’s processing requests were placed in an SQS queue. This allowed multiple videos to be processed simultaneously, significantly reducing the time taken to make edited videos available to users. We also configured the Autoscaling group instances to be provisioned based on the number of pending videos to be processed in the SQS queue. This ensured that the processing capacity of the AI video editor scaled up and down according to demand, optimizing resource usage and cost.

5. Compliance with SOC-2 Controls: tfsec was used in the Terraform pipelines to ensure compliance with SOC-2 controls. AWS account level checks were also configured to comply with SOC-2 controls and remediate any non-compliant resources/services.