In recent years, Perfect Diary, a cosmetics brand founded in 2017 by Yatsen Global, has launched a series of best-selling products, such as "Little Black Diamond" lipsticks and "Animal" eyeshadow palettes. With their sleek design and impressive cost-effectiveness, Perfect Diary's makeup products have gained popularity among young female consumers and seen rapid sales growth. In many Chinese Internet communities, Perfect Diary is hailed as "The Pride of Domestic Brands".
In August 2019, Perfect Diary began to work closely with Alibaba Cloud to adapt to its explosive business growth. Based on Alibaba Cloud's mature database services, the cosmetics brand quickly replicated best practices for e-commerce databases and adopted a hybrid transaction/analytical processing (HTAP) solution that consists of ApsaraDB PolarDB, Data Transmission Service (DTS), and AnalyticDB. In April 2020, during the 3rd Anniversary Campaign, Perfect Diary saw a 50-fold increase in its system throughput compared to six months ago.
In 2019, Yatsen Global began to aggressively push its new retail business forward. The company has taken advantage of its online sales experience and data and integrated online and offline channels to build beauty stores that provide a seamless and hassle-free shopping experience. How can Perfect Diary, which was borne in the e-commerce market, grab a piece of the highly competitive offline market? Cloud migration has assumed an important role in helping Perfect Diary drive its new retail business.
Unlike traditional offline stores, Perfect Diary's offline stores allow shoppers to use their mobile phones to go through the entire shopping process, including try-on, order placement, and pick-up. With this method, Perfect Diary can turn its offline sales data into digital information to facilitate analytics and decision-making. In addition, it can direct offline shoppers to its online sales channels that work better for subsequent activities.
Alibaba Cloud's mature Platform as a Service (PaaS) database services, which work in a simple and efficient manner, have addressed the demands of Perfect Diary's operations staff for online data processing and analytics. This allows the brand's technical team to focus on its business and supports its rapid business growth more effectively.
We are seeing endless e-commerce shopping festivals and live-streaming promotions throughout the year. Hot items often sell out in a few seconds after the "three, two, one" countdown. These exciting moments seem to be carnivals, however, they also present big challenges to the high concurrency capabilities of databases. The system must not only support immediate ordering with quick response, but also prevent overselling. In case of delay in delivery due to overselling, shoppers will be likely to lose patience and cancel their orders as they wait for the goods to arrive. If this is the case, shoppers will also have the impression that the merchant is unable to support fast delivery. This scenario can have negative effects on repurchase decisions and fan retention.
With the rapidly increasing number of offline stores and the huge traffic generated by live streaming, Perfect Diary's business system must handle explosive business growth. However, this system also encountered some database related problems that are commonly seen in e-commerce companies. These problems include overselling due to inconsistent inventory data, difficulty in increasing database throughput, time-consuming processes of configuration upgrade or downgrade, and poor statistics and analytics performance of SQL. To address these problems, Perfect Diary and Alibaba Cloud began to work closely together. They wanted to quickly replicate best practices for e-commerce databases that are based on Alibaba Cloud's mature database services.
After a detailed discussion, Alibaba Cloud developed a two-phase plan for Perfect Diary's architecture planning and technical evolution. The first phase lasted for only a month and focused on optimization. During this phase, Alibaba Cloud quickly developed a stress testing framework with the Alibaba Cloud Performance Testing tool. Then, Alibaba Cloud stress-tested the complete process to identify the resource usage and bottlenecks, optimize each instance, and analyze the resource usage and system bottlenecks of each database.
Perfect Diary set a new sales record during Double 11 in 2019. The sales in the first 28 minutes exceeded the sales in 24 hours during Double 11 in 2018. With the new record, Perfect Diary became the first domestic brand to rank first in the "Double 11 Top-selling Makeup Brands" list.
During the second phase, on the last day of stress testing for Perfect Diary's 3rd Anniversary Campaign in April 2020, the speed of the order processing system reached up to 10,000 transactions per second (TPS). Meanwhile, the corresponding write speed of Apsara PolarDB reached 100,000 TPS. This indicated that the system throughput had increased by 50 folds in six months.
On the day of the campaign on April 14, millions of users rushed into the system immediately upon the beginning. Hundreds of thousands of "Puppy" eyeshadow palettes sold out within tens of seconds. The number of orders placed per second broke the record, with a peak value that was several times higher than the historic peak. The business traffic at the peak was 50 times higher than half a year ago, and the system ran smoothly.
As Perfect Diary's user base and gross merchandise value (GMV) continue to grow rapidly, the system must meet increasingly higher reliability requirements. Alibaba Cloud's database solutions with active geo-redundancy will provide best practices for this purpose. In the first quarter of 2020, with the help of Alibaba Cloud Database Backup Service (DBS), Perfect Diary quickly implemented geo-redundancy to meet the requirements of Grade III Protection of Information Security. In addition, Perfect Diary has become one of the earliest Alibaba Cloud Database Autonomy Service (DAS) customers since the public preview phase ended. The service provides the cosmetics brand with a variety of powerful and convenient database features, such as automatic detection, automatic repair, automatic optimization, automatic operation and maintenance (O&M), and automatic security. Perfect Diary is currently reconstructing distributed databases. After the reconstruction, the e-commerce system will support linear performance expansion.
Perfect Diary's database architecture
Perfect Diary began to use Alibaba Cloud ApsaraDB for Redis clusters in August 2018, and began to intensively use Apsara PolarDB for MySQL in August 2019 to adapt to its explosive business growth.
At present, Perfect Diary's database system is almost completely built on Alibaba Cloud's products. As a brand that is well-known for releasing popular products by crossover merchandising and co-branding, Perfect Diary has addressed many common challenges in the industry by using Alibaba Cloud's PaaS database services. With the cloud migration of its entire database system, the brand has set a benchmark for all "unicorn" enterprises to use digital infrastructure to drive rapid business growth.
In terms of market share, Alibaba Cloud is one of the top three providers in the global cloud database market and the top provider in China. Alibaba Cloud has been rapidly advancing its service technologies and has been widely recognized. It is the only Chinese enterprise that has been placed in Gartner Magic Quadrant for Operational Database Management Systems for two consecutive years. To date, more than 400,000 database instances have been migrated to Alibaba Cloud, including the databases of leading enterprises in fields such as government services, retail, finance, telecommunications, manufacturing, and logistics.
Tech Insights - Two-phase Commit Protocol for Distributed Transactions
Alibaba Cloud Community - November 9, 2021
Alibaba Clouder - November 16, 2020
Iain Ferguson - May 9, 2022
ApsaraDB - February 9, 2021
Alibaba Clouder - April 21, 2020
Aliware - November 10, 2020
Migrating to fully managed cloud databases brings a host of benefits including scalability, reliability, and cost efficiency.Learn More
Secure and easy solutions for moving you workloads to the cloudLearn More
Migrate your legacy Oracle databases to Alibaba Cloud to save on long-term costs and take advantage of improved scalability, reliability, robust security, high performance, and cloud-native features.Learn More
An easy transformation for heterogeneous database.Learn More
More Posts by ApsaraDB