Background

T3Go is a smart travel ecosystem platform built by Nanjing LingHang Technology Co., Ltd. With the brand vision of "becoming the most trustworthy travel service provider", T3Go is committed to providing users with safe, convenient, and high-quality travel services.

As the company began to grow, its original system was unable to support increased demand for travel services. To address this issue, T3Go has started to iteratively update travel products with Huawei Cloud databases.

展开详情

Challenges

  • Migrating different types of databases is complex

    Data of different T3Go travel services was stored in different databases such as MySQL, MongoDB, and PostgreSQL. T3Go required a database migration solution that can adapt to different database types and requires minimal costs on application rearchitecting.

    Data of different T3Go travel services was stored in different databases such as MySQL, MongoDB, and PostgreSQL. T3Go required a database migration solution that can adapt to different database types and requires minimal costs on application rearchitecting.

  • Migration has to be completed within 30 hours

    T3Go has hundreds of millions of large tables containing various types of data. The amount of relational data alone reaches dozens of terabytes. There are instances where only table data needs to be migrated. T3Go required that the migration be completed within 30 hours and incremental data comparison be performed to verify data consistency.

    T3Go has hundreds of millions of large tables containing various types of data. The amount of relational data alone reaches dozens of terabytes. There are instances where only table data needs to be migrated. T3Go required that the migration be completed within 30 hours and incremental data comparison be performed to verify data consistency.

  • Travel applications run stably without downtime

    The performance of instances that are processing query requests cannot be affected during the database migration. As a result, services are affected. The open-source Canal used by T3Go is frequently restarted. T3Go wants to replace Canal with a reliable service that can report alarms.

    The performance of instances that are processing query requests cannot be affected during the database migration. As a result, services are affected. The open-source Canal used by T3Go is frequently restarted. T3Go wants to replace Canal with a reliable service that can report alarms.

Solutions

  • Improved write performance to complete full synchronization within 30 hours

    Some parameters and settings have been modified to improve write performance of the destination instance on Huawei Cloud. By using flow control, DRS ensures that all data is migrated within 30 hours.

    Some parameters and settings have been modified to improve write performance of the destination instance on Huawei Cloud. By using flow control, DRS ensures that all data is migrated within 30 hours.

  • No impact on source databases

    Based on a thorough survey, DRS sorts out the load of all source instances and provides different flow control schemes for different instances to ensure efficient migration and stable source database performance.

    Based on a thorough survey, DRS sorts out the load of all source instances and provides different flow control schemes for different instances to ensure efficient migration and stable source database performance.

  • Quick adaption to Canal and replacement of Canal

    DRS quickly adapts to Canal and provides high reliability and early alarm reporting.

    DRS quickly adapts to Canal and provides high reliability and early alarm reporting.

Benefits

  • Smooth migration without service interruption

    The parameters, synchronization mode, and architecture of the destination database have been improved, and DRS flow control is used to complete full synchronization within a specified period of time.

    The parameters, synchronization mode, and architecture of the destination database have been improved, and DRS flow control is used to complete full synchronization within a specified period of time.

  • No impact on source services

    DRS flow control ensures that services on the source database are running properly during the migration.

    DRS flow control ensures that services on the source database are running properly during the migration.

  • Improved data subscription stability

    RDS adapts to Canal and combines with DRS high reliability and automatic O&M to greatly improve data subscription stability.

    RDS adapts to Canal and combines with DRS high reliability and automatic O&M to greatly improve data subscription stability.