Brain Frame is a project initiative under St. Xavier’s College, focused on leveraging Machine Learning (ML) to drive innovation and research.
To support this academic endeavor, the project required a secure, scalable, and cost-effective cloud environment capable of handling ML workloads while supporting multiple users.
The Challenge: The Brain Frame team needed an AWS setup that could:
Deploy Windows servers and MySQL databases for application and data processing. Integrate Amazon SageMaker with S3 storage for ML training and lifecycle management. Support long-term workloads (1 Windows server + 1 RDS) as well as short-term workloads (additional 1 Windows server + 1 RDS for 3 months). Allow 4 concurrent users to access the temporary environment securely. Provide robust monitoring, cost tracking, and security controls for both long- and short-term workloads.
The Solution: Pentagon designed and implemented a cloud-native environment on AWS, tailored to the project’s requirements:
VPC Configuration – Created Brain-Frame-VPC with 2 public subnets for application servers and 4 private subnets for RDS databases. Compute & Database – Deployed 2 Windows EC2 application servers and 2 Amazon RDS (MySQL) instances, with one set configured for long-term use and another for a 3-month period. Security – Configured security groups to whitelist only the client’s static IP for RDP and RDS access. Machine Learning Enablement – Set up Amazon SageMaker Studio in the Mumbai region, integrated with S3 for scalable data storage and ML workflows. Monitoring & Alerts – Implemented CloudWatch for performance monitoring of EC2 and RDS, with SNS notifications for triggered alerts. Cost Management – Configured budget alerts at 80% and 90% thresholds to proactively track monthly AWS consumption.
AWS Services Used:
Amazon VPC Amazon EC2 Amazon RDS (MySQL) Amazon S3 Amazon EBS Amazon SageMaker AWS IAM Amazon SNS Amazon CloudWatch AWS CloudTrail
Benefits
Scalability & Elasticity
Infrastructure scaled easily for both short-term and long-term workloads.
Resiliency & Availability
High uptime ensured through AWS-native services.
Security
Role-based IAM, IP whitelisting, and monitoring protected workloads.
Machine Learning at Scale
SageMaker enabled streamlined ML lifecycle management, data preparation, and accelerated development.
Performance & Cost Efficiency
High-performance workloads with budget alerts kept project costs predictable.
Ease of Maintenance
Automated monitoring and backups reduced operational overhead.
