Vertical E-commerce
A Tobacco company in China
About The Client
The client is a well-known China state-owned tobacco company, its has an e-commerce platform for retail customers to order cigarettes online.It provides cigarette trading, online payment, internet marketing, industrial parks (points mall) and other functions,whose users has reached 3.3 million.
Project Features
High Concurrency. It has a large number of system registered users and active users with large amounts of daily visits.
Homepage load slowly due to too many elements on the page.
Complex order process.
A variety of large-scale activities. With frequent sales promotion activities and gift distribution based on different rules, there are potential risks of high concurrency, mistakes during gift distribution etc.
Severe consumption of system resource. Due to its own business characteristics, the consumption of the system's hardware resources is severe.
Capacity planning. Capacity planning is required to cope with future business growth in the next few years. Benchmarking test is needed to find out the suitable software and hardware products.
Service Value
The maximum number of concurrent users of the key business scenarios was been improved by 3 to 6 times, with average response time of the same concurrent system shortened by 18 times. System maximum throughput was increased from the original 165QPS (query per second) to 7745QPS.
Homepage loading speed was been improved by 70% and the maximum number of concurrency that the homepage loading can support is increased by 40 times.
Identified potential performance bottlenecks by simulating a variety of order processes and testing the system situation during peak hours.
By simulating panic buying, problems including gifts over distribution and gifts distribution error was found and solved.
The appropriate system architecture was planned based on business growth forecasts for the next few years to provide excellent scalability. With strict benchmarking test, the most suitable software and hardware components was selected.
