OceanBase Cloud helps GCash to handle massive growth via distributed architecture

  • 70%Reduction in data storage space

  • 40%Reduction in database resource costs

  • 80%Fewer data nodes and higher O&M efficiency

Contact Us

Founded in 2017 by the Philippine telecom giant GlobeTelecom, GCash is the number one e-wallet service provider in the Philippines and the country's first “double unicorn”. The company offers users a range of financial services accessible through smartphones. In 2021, GCash was recognized by The Asian Banker (TAB) and International Data Corporation (IDC) for its outstanding inclusive digital financial service program. As of June 2022, the number of registered GCash users reached 66 million, indicating that one out of every two people in the Philippines uses GCash services.

Background

GCash is a leading e-wallet service provider in the Philippines. By the end of 2020, the volume of data had surged, and the used storage space of the databases was approaching the upper limits. GCash had to invest substantial manpower and resources in data sharding and cleanup. Additionally, attempts were made to mitigate the challenges by consistently reorganizing tablespaces. Unfortunately, these workarounds did not prove effective in helping GCash overcome the problem. Their databases could not provide stable support for the business growth. As a result, service availability decreased, and the O&M complexity and costs soared, leaving the DBA team struggling to maintain a large number of MySQL instances. The extreme "internal challenges" that came with the rapid business growth forced GCash to thoroughly upgrade its IT system.

Business Challenges

  • High Operational Costs
    • GCash handled more than one million transactions every day, and the data volume of the legacy databases kept increasing by about 10% every month. Due to the specification limits of MySQL cloud databases, GCash had to upgrade the CPU resources whenever it scaled up the storage in response to the data surges, which led to a waste of resources and high monthly storage costs.
  • Business availability risks
    • The increasing data volume revealed the weaknesses in the database architecture of GCash. Most of these weaknesses existed in key business and service systems, and they could cause considerable impact and jitters in the event of failures and stability issues. The impact of the architecture flaws became more serious along with the continuous business growth, putting GCash at increasing risk of business interruptions.
  • Unsatisfactory user experience
    • In most cases, online traffic surges take place in the transaction and payment system. The system performance determines whether users can smoothly handle their transactions and payments in many high-concurrency scenarios. Given high-concurrency SQL requests in a short period of time, the increase in latency of processing each request by even one millisecond can cause poor user experience. The original MySQL cluster of GCash supported only a limited number of connections due to the architecture design and was extremely sensitive to high-concurrency requests, which could easily lead to system instability. 
       

Solution

  • 1. Advanced data migration and compression technologies
  • OceanBase designed a data migration solution tailored for GCash based on the best practices of the Site Reliability Engineering (SRE) team of Ant Group. The solution involves OceanBase Migration Service (OMS), an automated platform that obtains permissions to access the source database, synchronizes incremental data from the source database to the destination database, and establishes connections between applications and the destination database. The solution also utilizes advanced compression techniques based on the LSM tree, achieving lossless compression of massive amounts of data.
  • 2. Business-oriented architecture transformation
  • OceanBase Cloud supports multitenancy. This feature allows GCash to build resource pools and create approximately 10 OceanBase clusters to accommodate hundreds of legacy MySQL instances in the form of tenants. Tenants that support core and general business systems are hosted in different clusters. This ensures that any jitter occuring in one cluster impacts only 10% of the entire database system.
  • 3. A three-zone high availability architecture
  • If a traditional database experiences server shutdown during a transaction, a forced business switchover is likely to cause the loss of the transaction data. The three-zone high availability architecture of OceanBase Cloud ensures business continuity upon failures in any single zone. Following the upgrade, the number of connections to a single node of OceanBase Cloud is approximately 5 to 8 times that of a MySQL database. The total number of connections supported by a multi-node OceanBase cluster is sufficient to smoothly deal with more drastic traffic fluctuations. 

Benefits

  • Reduction in data storage space by 70% and resource costs by 40%
    • GCash seamlessly migrated hundreds of databases to OceanBase Cloud with zero downtime. On average, each database in OceanBase Cloud takes only 1/10 of the storage space required by a legacy database. The total data storage space was reduced by 70%, saving 40% of database resource costs.
  • Significant improvement in O&M efficiency with 80% fewer data nodes
    • Hundreds of MySQL instances were consolidated into approximately 10 OceanBase clusters. This not only reduces the number of data nodes by 80% but also simplifies the O&M tasks significantly. Moreover, OceanBase Cloud Platform (OCP) supports online DDL operations and intelligent diagnostics based on the OceanBase kernel, which greatly improves the O&M efficiency of Database Administrators (DBAs).
  • Multi-zone financial-grade disaster recovery
    • The three-zone high-availability architecture ensures GCash's business continuity in the event of a single data center failure. The solution reduces the recovery point objective (RPO) to zero and the recovery time objective (RTO) to less than 30 seconds. It provides GCash with capabilities for multi-zone financial-grade disaster recovery. This solution also implements system-wide data encryption, which guarantees data security with less than 5% performance degradation.