Challenge
IDeaS has been one of the earliest pioneers of multi-tenant SaaS offerings. Its market-leading core SaaS platform has been leveraging private and hybrid cloud for over 20 years.
In recent years, IDeaS business exploded, posing new challenges. These growth-based challenges required a new approach to datacenter and application development strategy to support:
- Onboarding new customers necessitated additional capital expenditures in the form of hardware purchases and investments in datacenter management capacity
- A faster, more agile and elastic production release cycle was needed that would remove orchestration constraints across more than 30 different components
- IDeaS’ relentless focus on customer experience required a solution to potential performance degradation during releases and code merges
- Enhanced financial reporting granularity would allow for an increase in financial management velocity through more accurate datacenter cost tracking and differentiation
SAS, IDeaS parent company, already having made significant investment in AWS, allowed IDeaS leadership to have further confidence that migrating to AWS would provide the elasticity, scalability, cost efficiency and financial transparency they were looking for.
Solution
As an AWS Advanced Consulting Partner and a technology partner to IDeaS for more than three years, DataArt was invited to participate in IDeaS’ migration initiative from the beginning. Leveraging the existing AWS governance provided by SAS, IDeaS were able to focus exclusively on migrating the workloads.
Laying a Foundation
The first steps were designed to take advantage of the existing managed AWS environment. This included using Amazon Cognito to support managed authentication for web and mobile clients. Other management features included application permissions and feature flagging based on purchased subscription profiles. These were implemented with Amazon RDS and DynamoDB.
Migration strategy: Containerization “Conveyor Belt”
IDeaS and DataArt then continued with assessment, design, planning, and cost estimation for the workload migrations. The joint team was able to leverage the existing code written in Java to streamline the migration process. DataArt smartly migrated existing Java components and then standardized workload deployment and operations through containerization.
After iterative discussions and proof-of-concept implementations, the team settled on an architecture using AWS ECR to store container images, and Amazon ECS on AWS Fargate to deploy and run the containers.
Once the approach was proven out with the first successful migrations, the team then repeated the pattern for rest of the components, creating a “conveyor belt”. This “conveyor belt” approach greatly streamlined the process and accelerated migration timelines, and ultimately IDeaS’ time to market.
As the platform was containerized, engineering teams gained the ability to release to production without resource contention nor team dependencies. Infrastructure management, redeployment and scaling became entirely automated.
Resulting in the following release process:
Comprehensive approach
To enable successful end-to-end migration, the DataArt and IDeaS team took a comprehensive approach leveraging both AWS and 3rd party services, including:
- Remote access from IDeaS networks to AWS networks provided by AWS VPN.
- Customer-authentication leveraging Cognito.
- Customer access through AWS Application Load Balancer or AWS API Gateway
- Modernized existing relational databases for direct migration to AWS RDS
- Migration of on-prem NoSQL solution based on MongoDB to MongoDB Atlas in AWS
- RabbitMQ migration to CloudAMPQ with peered VPC
- Infrastructure build our with Terraform.
- Continuous integration improvement through BitBucket Git repository and BitBucket Pipeline.
- Support multi-availability zone capable VPCs supporting public and private subnets through AWS Fargate services.
Outcomes
- IDeaS and DataArt’s approach enabled incredibly fast workload migration to ensure efficiency, repeatability, and reliability, drastically minimizing code changes.
- By leveraging cloud-native containerization capabilities offered by AWS, IDeaS were able to benefit from automatic scalability, elasticity, performance, and high availability native to AWS cloud.
- Increased capacity scaling and scaling, as well as reduced IT organizational strain during demand spike by leveraging AWS managed services.
- Improved team productivity and release velocity (up to several production releases in a single day), resulting in increased business agility.
- Account specific billing and greater financial reporting granularity through this new native AWS approach.
AWS services used include: AWS Fargate, Amazon ECS, Amazon ECR, Amazon RDS, Amazon ElastiCache, Amazon API Gateway, AWS Lambda.
