Migrating and Optimizing infor's Pricing and Billing Solution

Key Takeaways

1. infor's Complete Billing System (CBS) is a batch processing application primarily based on the Microsoft stack, used by commercial banks for treasury management services.
2. infor decided to move CBS to the cloud on AWS, primarily to take advantage of the cloud's strategic benefits like scalability, security, and cost optimization.
3. The migration journey involved making decisions around identity and authentication, data storage, application deployment, and licensing.
4. The key challenge was to reduce the batch processing time by 33% without making any code or platform changes.
5. The solution architecture team, comprising AWS Solutions Architects and infor's cloud architect, followed a structured approach to optimize the performance of the CBS application on AWS.
6. The optimization process involved several rounds of testing and adjustments to the infrastructure, focusing on instance types, storage configurations, and database licensing.
7. The final outcome was a 48% reduction in batch processing time, a 74% reduction in SQL Server licensing costs, and a 24% reduction in the total cost of ownership (TCO).
8. The success of the exercise was attributed to the team's ability to fail fast, fail often, and fail cheap, enabling them to arrive at the optimal solution quickly and cost-effectively.

Migration and Optimization Journey

Assessing the Options

1. Identity and Authentication: Evaluated the use of Managed Active Directory (AD) service vs. self-managed AD on Amazon EC2.
2. Data Storage: Explored options like Amazon RDS for SQL Server, self-managed SQL Server on Amazon EC2, and RDS Custom for SQL Server.
3. Application Deployment: Considered both platform-managed (AWS Elastic Beanstalk) and self-managed (Amazon EC2) options for running the .NET applications.
4. Licensing: Evaluated the use of license-included instances, Bring Your Own License (BYOL) with License Mobility, and BYOL on dedicated hosts.

Optimizing the Performance

1. Establishing the Baseline: Collected performance data from the on-premises environment, as infor did not have access to this information.
2. Iterative Testing and Adjustments:
   • Tested different instance types and storage configurations for the application and database servers.
   • Focused on metrics like CPU utilization, memory utilization, EBS throughput, and IO operations per second (IOPS).
   • Made incremental changes to the infrastructure based on the findings, such as increasing EBS throughput, leveraging NVMe-based instant storage for tempDB, and splitting the database storage into separate volumes.
3. Achieving the Target: After four rounds of testing and adjustments, the team was able to achieve a 48% reduction in batch processing time, meeting the original goal of a 33% reduction.

Cost Optimization

1. SQL Server Licensing: Evaluated the impact of SQL Server licensing costs and recommended the use of SQL Server Standard edition, which resulted in a 74% reduction in licensing costs.
2. Total Cost of Ownership (TCO): The optimized infrastructure and licensing decisions led to a 24% reduction in the overall TCO.

Key Factors for Success

1. Structured Approach: The team followed a well-defined process of setting goals, identifying constraints, testing, and iterating on the solution.
2. Collaborative Effort: The joint team of AWS Solutions Architects and infor's cloud architect worked together to address the challenges.
3. Fail Fast, Fail Cheap: The team's ability to quickly test, fail, and iterate on the solution enabled them to find the optimal configuration efficiently.