Elite Cloud Assists SCUD in Migrating to AWS: A Customer Case Study

Overview

The objective of this project was to migrate SCUD’s quantitative finance trading system from Google Cloud Platform (GCP) to Amazon Web Services (AWS). SCUD is a leading technology company focused on the Web3 quantitative trading domain, dedicated to driving innovation and integration of blockchain technology and digital assets. The team comprises experienced blockchain experts, fintech professionals, and technology luminaries. Their core business is Web3 quantitative trading, combining blockchain technology and quantitative trading strategies to provide investors with an efficient, secure, and transparent digital asset trading experience. They develop smart contracts and decentralized finance tools, leveraging the decentralized nature of blockchain and the intelligence of quantitative trading to offer customers exceptional trading strategies.

Project Goals and Pain Points

• GCP Managed Service Migration: A professional technology service provider was needed to assist in matching corresponding AWS services, including service adaptation, migration feasibility research, and validation. Due to product differences, system architectures, and deployment methods, the migration and rebuilding workload was substantial.
• Customer Self-Built Data Warehouse and Aurora PostgreSQL Continuous Synchronization: The customer’s self-built GreenPlum 7 data warehouse is PostgreSQL-compatible and needed to synchronize with a PostgreSQL database. This involved synchronizing numerous data tables, including tables with millions of rows and dozens of smaller tables. A Redshift alternative was recommended, but the customer currently still uses their self-built GreenPlum data warehouse.

Expected Project Outcomes

• Successful migration of Cloud Store files, Elasticsearch, and PostgreSQL databases, with data integrity verified.
• Functional testing of the Jasmine Mint application services.
• EKS cluster elastic scaling time ≤ 10 minutes.
• Real-time synchronization to the core Aurora PostgreSQL database via DMS.
• Email alerts for AWS cloud resources and application-related monitoring metrics.
• Cost control with clear billing.

Proposed Architecture Diagram

To meet the customer’s requirements, the overall business system deployment architecture is illustrated in the following diagram:

• Architecture Explanation:

  • Landing Zone Planning: Assisted the customer in reasonable network layer division, with the main business concentrated in the private network and proper routing and security group policies.
  • Multi-Availability Zone Deployment: Improved business security and reduced single points of failure.
  • Utilization of AWS Services: Including IAM, CloudTrail, etc., to assist in managing permissions and operation records.

• Monitoring and Alert Notification Design:

  • Configured CloudWatch monitoring and alerts for events such as EC2 status changes, memory, and CPU utilization.
  • Provided custom alerts for core database PostgreSQL load and DMS failures.

• Core Business EKS Cluster Design:

  • Worker Nodes deployed across multiple availability zones, using Karpenter for Kubernetes auto-scaling.
  • Jasmine Mint core business deployed within the EKS cluster, with services exposed via ALB Controller Ingress.
  • Single EKS container cluster, with 2-20 nodes deployed in Private Subnets.
  • Node group security groups only allow access from the EKS cluster and ELB.

Challenges and Solutions

• Database Migration and Synchronization:

  • Importing GCP Elasticsearch snapshots to AWS OpenSearch.
    Planned to use snapshot capture, export to S3 for migration, requiring advance evaluation of snapshot time under high Elasticsearch load.
  • Synchronizing multiple tables from the customer’s self-built data warehouse to the core PostgreSQL database.
    Utilized DMS best practices, splitting multiple instances to synchronize different tables, improving fault tolerance and performance.

Through the professional services provided by the Elite Cloud team, SCUD successfully migrated its quantitative finance trading system from GCP to AWS, achieving the expected project outcomes and providing customers with a more efficient, secure, and reliable digital asset trading experience.