Analysys OLAP Algorithm Competition: GBase Achieves Another Great Result

On October 28, the winners of the Analysys OLAP Algorithm Competition were announced at the Analysys A10 Big Data Application Summit 2017! GBase won third place in the commercial group.

Representative of the winning team receiving the award on stage

The competition was divided into commercial and open-source groups, with over 40 top technical teams participating. The 100-day program went through three major stages: internal R&D testing, formal environment testing, and running benchmarks on Analysys's official data environment.

The GBase team chose the mature GBase 8a MPP Cluster, a big data analytical database, for the competition. GBase 8a MPP Cluster uses hybrid row-columnar storage and can handle petabyte-scale structured data. It also leverages advanced data compression algorithms to effectively improve I/O performance. For data processing, it employs a fully parallel MPP + shared-nothing distributed architecture that supports high concurrency, making it highly efficient in real-time computation for ordered funnel analysis.

Backed by a mature big data product and a dedicated team, GBase achieved such outstanding results. As a proven analytical database product, GBase 8a MPP Cluster not only delivers powerful performance in competition scenarios, but also excels in real-world applications. To date, GBase 8a MPP Cluster has successfully supported the core business systems of over 300 customers across various industries, including the People's Bank of China, Agricultural Bank of China, State Grid Corporation of China, China Mobile, and the General Administration of Customs of China, earning wide recognition from users.

Competition Overview:

The OLAP Algorithm Competition was conducted on 4 UCloud cloud hosts, each with 16 cores, 16 GB RAM, and a 300 GB SSD data disk. Eight different scenarios were presented simultaneously with timing started immediately. Each scenario had to be completed within 8 minutes, and the total time for all 8 scenarios was within one hour, assessing the teams' real-time computing accuracy and time efficiency under different scenarios. Each scenario was assigned a different weight based on factors such as the number of events, event complexity, event attributes, and the size of the time window. The total score was the sum of each scenario's score multiplied by its weight. The team with the shortest runtime and highest accuracy across the 8 scenarios won. During the formal data environment benchmark stage, the entire process was video-recorded to ensure the validity and reliability of the competition results.

Participating Teams:

Meituan Dianping, Alibaba Cloud, Xiaomi, Lianjia, Sogou, iFlytek, China Unicom, Yuansu Data, anal, AggreData, GBase , Kyligence, PingCAP, LinkDoc, iflytek, hulu, Truxing, Shanghai Ruimin, Red Cloud Data, Runqian, Fanruan Software, Migu Music, Hangzhou Daishu Technology, Lejian, iFlytek, Shanghai Kuizhi, Qianding Internet, Beijing Jianke Cloud Network, Lexiang Tianxia, Niuban Technology, Beijing University of Posts and Telecommunications, Beijing Jiaotong University, and over 40 other renowned enterprises, research institutions, and individuals

Competition Results:

Commercial Group

1st Place: PingCAP   

2nd Place: Fanruan Software  

3rd Place: GBase

Open-Source Group

1st Place: Wei Wan and Li Benwang of Guangzhou Xiangliangxian Technology

2nd Place: Sun Yerui of Meituan Dianping

3rd Place: Lu Yuekai of Beijing University of Posts and Telecommunications