Overview
Elite Cloud assisted FANWE in completing the migration of their entire platform system from Alibaba Cloud to AWS. Since 2021, the client had deployed their platform on Alibaba Cloud in Hong Kong. Considering the growth of their overseas business needs, FANWE aimed to improve global coverage by deploying their applications globally on AWS, and increase efficiency through leveraging AWS’s big data platform solutions.
Elite Cloud’s services included helping the client understand AWS services, designing the cloud architecture, assisting with project testing, service migration, and deployment optimization. The migration plan began in January 2023 and was expected to take approximately three months to complete. Prior to the migration, it was necessary to thoroughly understand the client’s resource usage on Alibaba Cloud, their business deployment architecture, as well as map out the business modules and understand data usage. After the migration, Elite Cloud will continue to ensure stable operation and incremental growth of the business.
This project involved migrating existing Alibaba Cloud resources and deploying new services, including:
- Current hosted services on Alibaba Cloud: RDS PostgreSQL, ACK, Redis, Elasticsearch, OSS, Kafka;
-
- PostgreSQL versions 12 and 13, with a total data size of approximately 10TB across multiple business databases, some used for data analysis with high IO loads. Some databases needed version upgrades before migration.
- Most Redis instances were small configurations compatible with version 5.0.
- Elasticsearch 7.10 used for tag search and log viewing, needing adaptation to Amazon OpenSearch.
-
Alibaba Cloud’s CDN service and object storage, including CDN for dynamic network acceleration and partial webpage downloads, as well as OSS for data storage and archiving. These needed to be migrated and adapted to CloudFront and S3, with access control and storage tiering design.
-
Migration of all Relation services to AWS, including containerized services currently on Alibaba Cloud’s ACK. This involved migrating the containerized platform from the vendor to AWS EKS, requiring a detailed understanding of Relation’s resource usage, deployment architecture, and data usage on Alibaba Cloud.
Project Goals & Pain Points
- Alibaba Cloud’s network could not meet business needs, with requirements spanning Southeast Asia to the Middle East, Europe, and the Americas. AWS’s global infrastructure with over 400 edge locations connected via a global backbone network aligned with the client’s requirements.
- Migrating hosted services from Alibaba Cloud involved complex technical services due to product differences, system architectures, and deployment methods between the two vendors, resulting in significant workload.
- The client needed to find an AWS partner with extensive experience and strong technical capabilities to rapidly complete the Proof of Concept (PoC) testing and deployment migration for the Relation platform. The partner needed to provide robust technical and staffing support to quickly complete the architecture design, migration, and deployment of Relation on AWS.
Reasons for Choosing AWS
- Security and Compliance: Relation operates in the financial industry and must adhere to strict regulations while maintaining a reliable disaster recovery plan.
- Industry Solutions: AWS has a strong presence in Southeast Asia with multiple Availability Zones. In this region, AWS has many customers similar to Relation, offering extensive experience supporting Relation’s deployment and business expansion.
- Cost Management Flexibility: The AWS platform provides flexible resource purchasing options, allowing dynamic adjustment of resource usage to help users save costs.
Expected Project Outcomes
The client will consider the project successful if the following main criteria are met:
- Databases migrated normally, with data integrity verification passed;
- All application service functionalities tested normally;
- EKS cluster elastic scaling time ≤ 30 minutes;
- Data warehouse application views displayed normally;
- Average application latency from various office locations met requirements;
- AWS cloud resource and application-related monitoring metrics alerted to enterprise WeChat;
- Self-hosted Prometheus + Grafana server monitored normally.
Recommended Architecture Diagram
To meet the client’s needs, the overall business system was deployed with the following architecture:
Detailed design plans for network design, security design, EKS cluster design, log collection, monitoring/alerting, and Greenplum data warehouse design can be referenced in the original text.
Successful Migration Delivery Criteria
- Solution meets requirements;
- Migration documentation is complete, and application and data migration is finished;
- Post-migration functionality meets requirements;
- Stable operation for 15 business days.