Revolutionizing Real-Time Travel Data: How DragonPass Leveraged RisingWave Streaming Database

Revolutionizing Real-Time Travel Data: How DragonPass Leveraged RisingWave Streaming Database

Company Background

Since its inception in 2005, DragonPass has been providing diversified travel services, including VIP lounges, to its global members. As of the end of 2019, DragonPass boasts a worldwide presence, spanning over 140 countries and regions, 600+ cities, 700+ airports and high-speed train stations, with over 30 million members and partnerships with over 400 renowned enterprises across various industries, including banking, card organizations, insurance, airports, hotels, internet, and mobile services. In the realm of travel, DragonPass has established itself as a global leader in service network coverage.

In addition to serving individual travelers, DragonPass collaborates with over 200 banks, credit card issuers, and telecom operators. Through their global offices, DragonPass helps partners engage with their customers and enhance their travel experiences.

DragonPass offers a one-stop white-label travel solution, encompassing airport lounges, dining, retail, and car services. Their user-friendly platform is accessible via the web and mobile devices, making it the ultimate tool for customer interaction in the travel industry.

Situation Before Using RisingWave

Before DragonPass embraced RisingWave, they had minimal real-time data processing capabilities. They predominantly operated on a T+1 processing model, where data was collected within a day (T), then processed and summarized in batch jobs running at the end of the same day or early the next day (T+1). The specific data processing workflow included:

  1. Data initially stored in distributed file systems such as HDFS.
  2. Subsequent data processing and computations performed using data processing engines like Apache Spark and other commercial analytical engines. These engines executed various data extraction, transformations, cleaning, and loading to prepare data for further analysis and reporting.
  3. Processed data transferred to the relational database system (RDS for MySQL) for persistent storage and querying.
  4. Finally, DragonPass utilized tools like Superset to query and visualize data stored in RDS. These visualization tools allowed users to create dashboards and reports for data analysis and monitoring key performance metrics.

Current T+1 data pipeline of DragonPass

While the T+1 technology stack could meet data processing requirements in certain scenarios, it couldn't provide real-time data analysis and immediate insights due to its batch processing nature. This is precisely why DragonPass decided to seek a more modern and real-time data processing solution with RisingWave.

System Selection

DragonPass encountered significant challenges in selecting a solution for real-time data processing and analysis. One of the primary hurdles was their reluctance to adopt the stream processing framework Apache Flink, primarily due to two key reasons: high costs and a steep learning curve.

1. High Cost Concerns:

Introducing a new technology stack often carries financial implications, and Apache Flink was no exception. DragonPass recognized that adopting Apache Flink could potentially result in substantial costs, including infrastructure upgrades and ongoing maintenance expenses. These financial considerations were crucial for a company aiming to optimize operations and allocate resources efficiently.

2. Steep Learning Curve:

The second challenge revolved around the complexity associated with Apache Flink. While it is a powerful tool for stream processing, it is known for having a significant learning curve. For DragonPass, a company focused on delivering world-class travel experiences, investing substantial time and effort into mastering a complex technology was not an ideal path forward. The fast pace of the travel industry demanded solutions that could be swiftly adopted with minimal disruption to existing operations.

Why They Chose RisingWave

In response to these challenges, DragonPass set out to explore a different approach. They recognized the need for a more accessible and cost-effective solution that would enable them to leverage the advantages of real-time data processing and analysis while avoiding the hurdles posed by high costs and a formidable learning curve.

This strategic pivot led DragonPass to explore alternative options, ultimately introducing them to RisingWave, an innovative streaming database that illuminated a promising path forward. Notably, RisingWave's compatibility with popular database management tools, including DBeaver, further enhanced its appeal. By choosing RisingWave, DragonPass found a solution perfectly aligned with their objectives, empowering them to overcome these challenges and embark on a journey toward real-time data analytics with newfound confidence and agility.

DragonPass's initiation into the realm of real-time data processing began with their introduction to RisingWave. This pioneering streaming database solution not only provided a fresh perspective on their data challenges but also solidified their partnership. RisingWave swiftly crafted an internal dashboard tailored to DragonPass's precise specifications, all accomplished within a matter of days, utilizing Superset as the foundation for this remarkable achievement.

Implementation Process

DragonPass's risk monitoring requirements could be abstracted as real-time metric aggregation within a time window to monitor recent instances of misuse. If implemented using Apache Flink, an external system (such as MySQL or Redis) would need to be separately deployed for storing the computed results from Apache Flink, which as a computation engine, lacks storage capabilities exposed to external systems. This increased complexity and maintenance costs.

Potential data pipeline with Apache Flink to rick metrics monitoring

RisingWave, on the other hand, is a streaming database that provides its storage capabilities and can expose stored data as tables or materialized views for online querying. Moreover, RisingWave is PostgreSQL-compatible, allowing the utilization of existing PostgreSQL ecosystem capabilities to meet requirements effectively.

Data pipeline with RisingWave adopted by DragonPass

Following their initial encounter with RisingWave and mutual understanding, DragonPass recognized RisingWave as an excellent tool to address their existing challenges. To achieve this goal, RisingWave engineers deployed RisingWave on DragonPass's internal machines alongside Superset. During the support process, RisingWave engineers implemented a Superset-RisingWave plugin to facilitate data visualization and integration. Additionally, by using DBeaver, a popular database management tool, they further streamlined integration, ensuring efficient and user-friendly control of RisingWave's capabilities.

Within this dynamic ecosystem, DragonPass leveraged RisingWave's robust capabilities to construct materialized views (MVs) that acted as repositories for their real-time data. These MVs facilitated not only data storage but also swift access and retrieval, ensuring that DragonPass had timely insights necessary for informed decision-making. Superset, in conjunction with these MVs, played a pivotal role in crafting a series of dashboards tailored to DragonPass's precise requirements. These dashboards provided an interactive and visually engaging interface for data analysis, empowering DragonPass's teams to explore and interpret their data with ease.

One of the standout features of RisingWave fully capitalized on by DragonPass was its window aggregation capability. This powerful feature enabled DragonPass to extract valuable insights in real-time, all without the need for extensive custom coding. It allowed them to aggregate and analyze data efficiently within defined time windows, offering a granular view of membership card usage trends. With RisingWave's window aggregation feature at their disposal, DragonPass had the agility and precision required to monitor and respond to abnormal usage patterns swiftly, contributing to improved decision-making and enhanced customer experiences.

Results

The results were nothing short of impressive. In just a matter of days, DragonPass had a fully functional real-time data monitoring system up and running. The agility and ease of implementation significantly reduced development time and costs compared to traditional approaches. RisingWave proved to be the catalyst that enabled DragonPass to embark on a new era of real-time travel data analytics, revolutionizing their services and enhancing the travel experience for millions of customers worldwide.

Testimonial

"At DragonPass, we understand the value of real-time data in the travel industry. Thanks to RisingWave, we were able to transition from T+1 processing to real-time data analytics within days, without the need for extensive coding or high costs. The RisingWave and Superset integration has been a game-changer for us, and we're excited about the endless possibilities it brings to our business."

DragonPass's journey with RisingWave showcases the transformative power of accessible and efficient real-time data processing solutions. By choosing RisingWave, DragonPass not only addressed their immediate data challenges but also set the stage for a future of innovation and growth in the travel industry. This success story is a testament to the evolution of travel data analytics, made possible by the next generation of streaming database technology.

The Modern Backbone for Your
Event-Driven Infrastructure
GitHubXLinkedInSlackYouTube
Sign up for our to stay updated.