Case Studies
22 July, 2024
August 10, 2021
Migration from Azure to AWS with Infrastructure as Code – Startell
About the Customer
Startell is a premier technology platform that connects marketers with influential creators to help automate influencer marketing and custom content development. Their comprehensive global database allows brands to target influencers with a broad range of interests, ensuring the perfect partnerships for every campaign.
As a digital marketing marketplace with developed & integrated AI technology, Startell creates cost-effective optimization on any marketing campaign.Their user-friendly platform gives clients the tools to create, track and analyze their projects and audience engagement on an hourly basis, all from one secure place.
Project Description
Startell was originally deployed on Azure and became interested in migrating their resources to AWS to both leverage AWS special services (such as AWS Recognition for the application) and scale their operations.
TeraSky DevOps team implemented the project as “lift-and-shift” with the following stages:
Customer’s Challenges
Before this project, Startell worked on Azure in a completely manual environment without any automated process.
This made the process slow and inefficient because the load was not required at all times. Developers had to create and configure a new instance manually to scale out.
Manual creation of a new environment was time-intensive and caused differences within the original environment.
All services used a public IP address, which caused significant security concerns. The access list only prevented people from accessing their services.
All updates had to be deployed manually on all instances.
Startell wanted to use the Infrastructure as a Code approach to create consistent and efficient additional environments (Dev, Training, Stage, Prod).
In addition, Startell wanted to deploy all of its services with the ability to allow easy resource scaling according to existing demands and to achieve cost savings.
Web Application should be deployed on beanstalk to remove all operational overhead and introduce ease deployment for light weight services.
Proposed Solution
TeraSky experts performed an assessment process including detailed discovery sessions with the Startell team to better understand their requirements and to define an optimal solution. To address the defined objectives, TeraSky’s Solution Architect recommended using the following AWS services:
Project Achievements
Once the project was completed, Startell was able to perform quick and frequent updates to their web application code in a streamlined manner, without leaving their IDE. Elastic Beanstalk service automatically handled the deployment details of capacity provisioning, load balancing, auto-scaling, and application health monitoring.
Further, redesigning the environments to use Auto Scaling Groups led to cost savings while at the same time allowing greater flexibility and improving the process significantly.
RDS service provided faster response time, which significantly improved overall service performance.
With our expert assistance, Startell reduced the cost of ownership thanks to effective resource management planning, effectively reducing the time consuming tasks to mere minutes through automation. As a result, they are prepared to meet their future challenges.
Performance
Startell was impressed by the sudden improvement in performance using the new DB service and ec2 instances.
Enhanced security: The new environment achieved enhanced security via OpenVPN, separated subnets between all services, and specifically tuned Security Groups.
Repeatability: The entire environment can be easily redeployed in another region or to recreated in a separate environment for testing and production the moment it is required. At this stage a single environment only is required.
Scalability: All backend servers were deployed inside of an autoscaling group to prevent reaching the computational limit during peak hours and to achieve cost savings by reducing a manually deployed instance to offload the compute time needed during operational hours.
22 July, 2024