Media Coverage: GBase Sets Its Sights on the 'Crown Jewel' of Big Data Infrastructure Software

The global big data era is well underway, from the Obama administration’s 'Big Data Research and Development Initiative' in 2012 to China’s State Council issuing the Action Plan for Promoting Big Data Development in 2015; from Google’s post-Hadoop 'troika' of Caffeine, Pregel, and Dremel shaping the global trajectory of big data technology, to the industry consensus that a hybrid architecture combining MPP and Hadoop is the optimal solution for big data analytics; from China’s big data market surpassing tens of billions of yuan last year to sustained high growth of 40% over the next three years, big data has opened new paths for national governance, transformed everyday life, and triggered massive technological upheaval in IT. Looking back at 2015, the importance of big data infrastructure is self-evident, and the database—as the core of foundational software—provides the technical backbone for big data analytics, continuing to generate long-term value in the big data space.

Over the past three decades, the database landscape has been dominated by an oligopoly of Oracle, IBM, and Microsoft. However, in the face of demands brought by big data, traditional databases encounter performance bottlenecks when handling queries, statistics, and analysis on data above 100 TB, especially at petabyte scale. The global database structure is undergoing a revolutionary shift: the once-monolithic OldSQL world is giving way to a landscape where OldSQL, NewSQL, and NoSQL jointly support diverse application types. The Chinese database vendor GBase (General Data Technology) had anticipated the arrival of big data. As early as 2008, it initiated the design and development of the analytical database GBase 8a, having seen customer needs for processing massive, multi-source, and heterogeneous data. Dr. Wu Xin, who had spent 12 years as a senior database expert at Oracle France, returned to China and took on the role of CTO at GBase, shouldering the responsibility as chief architect of GBase 8a. With an early start and high ambition, and cumulative R&D investment exceeding RMB 200 million, GBase 8a emerged as a purpose-built database product, seizing market opportunities during the big data wave with mature commercial-grade capabilities. 2015 was a year of rapid growth for GBase 8a. Looking back at the past year with Dr. Wu Xin, we can appreciate the extraordinary journey of a Chinese next-generation database that keeps pace with global technology and market standards.

Relentless Product Refinement and Technological Breakthroughs

GBase 8a released its first standalone version on April 1, 2010, and has undergone eight years of continuous evolution, making new technological breakthroughs and optimizations. Wu Xin summarized that in 2015 three major versions were released. Compared with earlier versions, v8.5.11.4 enhanced reliability; v8.5.11.5 doubled performance; and v8.6 transformed from an MPP architecture to a federated architecture, enabling online expansion with 'append-write', doubling batch 'write' performance, and increasing the maximum number of nodes in a single cluster from 192 to 300. These achievements are backed by over 1,000 person-months of R&D effort, 13 new feature categories, 650,000 new lines of code, and more than 400 documents spanning 4,800 pages. In Wu Xin's view, these numbers reflect the solid groundwork the GBase team laid to earn user recognition for GBase 8a.

Chinese databases are forging ahead into core applications of critical industries, where database quality is paramount. Wu Xin further explained that, building on tens of thousands of test cases from traditional manual testing, GBase built a leading automated database testing platform that conducts comprehensive testing of GBase 8a through automation. The platform can automatically generate vast numbers of test cases that simulate all possible user inputs and scenarios, achieving full coverage without blind spots. Based on this rigorous, efficient, and robust testing system, GBase 8a can rapidly iterate, with release cycles as fast as three days, enabling truly rapid response to frontline customer needs.

Hybrid Architecture Jointly Supports Big Data Applications

GBase 8a clusters are based on an MPP technical architecture. When discussing the relationship between MPP and Hadoop in processing big data, Wu Xin cleared up a common misunderstanding from the early days of the big data era: Hadoop is not the only technical solution for big data analytics. Having felt the industry shift firsthand, he noted that GBase had begun proposing the concept of 'industry big data' at least three years earlier—big data is not just a product of the internet; its applications in industry are no fewer than those in internet. The mainstream approach now is for MPP to store high-value-density data for analysis and mining, while Hadoop handles low-value-density data for base storage, cleansing, and simple classification queries. International vendors have already launched platform-type products of this converged architecture, such as EMC’s Pivotal and HP’s HAVEn. GBase 8a MPP clusters have been deployed in hybrid configurations with Hadoop in numerous high-end industries, including finance and telecom, forming a comprehensive solution that supports big data applications. GBase has also launched its own converged big data unified processing platform, GBase UP (Unified Platform). Wu Xin asserted that GBase had achieved technological and market parity with mainstream international vendors in database research.

Outstanding Market Performance Supported by User Trust

Speaking of market performance, Wu Xin shared that GBase 8a has been successfully deployed in sectors including telecom, banking, insurance, electric power, public security, security, customs, taxation, social insurance, finance, health and family planning, food and drug administration, statistics, oceanography, and defense. In telecom, it covers 16 provinces across China’s three major carriers. GBase 8a holds the leading market share among next-generation MPP databases. As a software product, its quality matures through continuous deployment—the rapid growth of GBase 8a is inseparable from user trust and support, with credit going equally to GBase and its users. Wu Xin gave an enthusiastic example: Agricultural Bank of China (ABC) replaced the foreign product Sybase with GBase 8a, breaking through a scalability bottleneck. Data processing capability improved 100-fold, and the data scale jumped from terabytes to petabytes. By introducing big data technology, ABC built its first ever enterprise-wide data warehouse at the headquarters level, expanding its business insight from branch-level to the entire bank. It became the first of China’s four major state-owned banks to complete the replacement of a foreign product in a core financial system with a domestic solution, successfully deploying a Chinese financial big data platform. Meanwhile, GBase 8a advanced through this project with several qualitative leaps; once ongoing projects are completed, the total cluster scale will exceed 400 nodes. Since going live, the system processes more than 6,000 complex jobs daily, handles an incremental data volume of over 2 TB per day, and its largest table has surpassed 100 billion rows with 1 billion new rows daily. Wu Xin proudly commented that these figures alone demonstrate GBase 8a has become a truly world-class database to be proud of.

In the telecom industry, which imposes extremely strict database requirements, GBase 8a outperformed many competitors and became the only domestic MPP database to pass China Mobile’s 'Centralized Data Warehouse' POC test; the only other products to pass were foreign solutions Greenplum and Vertica. Wu Xin briefly described the scale: 120 compute nodes, 30 load nodes, 800 TB of raw data, the largest table containing 3.6 trillion rows, and the smallest 800 million rows. Even at this level of difficulty, GBase 8a performed excellently and ranked among the top three. In 2015, GBase 8a beat out foreign competitors in 31 POC tests across four major sectors—securities regulation, human resources and social security, customs, and telecom—earning an excellent reputation and marking the end of the era where Chinese databases were considered unable to support core business systems, giving users more choice.

In 2015 alone, GBase 8a supported big data processing platforms for key industry users including China Unicom, China Telecom, the China Banking Regulatory Commission, the Ministry of Public Security, the State Administration of Taxation, and the General Administration of Customs, with projects cumulatively deploying 1,019 nodes and processing more than 5 PB of data. Wu Xin stressed that these benchmarks prove GBase 8a is on par technologically with foreign products, and even locally ahead in supporting data scale and system concurrency.

Aiming for the Crown Jewel of Big Data Infrastructure Software

Data is like energy, and data security is critical. The databases that store and compute big data are the 'crown jewel' of big data infrastructure software—and GBase is determined to capture it. This year, GBase became the first Chinese database vendor to list on the New Third Board and enter the capital market. The company will continue to invest in product R&D, product testing, application migration, and adaptation optimization, focusing its efforts on the database business and striving through relentless effort to seize that crown jewel. Discussing competition with international databases, Wu Xin said that in the big data analytics field, GBase 8a has already won out over long-established database giants Oracle, IBM, and Microsoft, and has repeatedly come out on top in PKs with emerging forces like HP and EMC, rekindling hope for Chinese databases in the industry and making the GBase team proud.

Product, Service, and Partnerships: Poised for Acceleration

Looking ahead to 2016, Wu Xin said the focus will remain on product, service, and partnership models. He will lead the R&D team to tackle the new application form of GBase 8a virtual clusters. Virtual clusters enable hierarchical management and support cross-data-center deployment, with multi-data centers managed in layers, and each data center creating independent sub-clusters according to its needs and scale. Facing challenges from key technologies such as metadata management for ultra-large clusters, communication models, and cross-heterogeneous-cluster access, the GBase team has prepared for intensive, closed-door R&D, striving to achieve the goal of a 1,000-node ultra-large cluster as early as possible. The reason GBase 8a has frequently delivered good news in high-end sectors like finance, telecom, and government is not only its world-class product, but also the quality support services behind it. Wu Xin mentioned that GBase will further expand its technical support team, strengthen its well-established technical system (three horizontals, three verticals, and one loop), continue to respond quickly to user needs, and reinforce value-based marketing through solutions. GBase 8a’s growth requires a sound ecosystem and the support of partners. GBase is expanding into new business models, having formed deep partnerships with companies like DHC Software, H3C, and Founder, some approaching a scale of ten million.

Pooling strengths in talent, technology, and industrial development, the GBase team will continue to uphold a spirit of craftsmanship, meticulously polishing GBase 8a—this crown jewel of big data infrastructure software—to underpin the implementation of the national big data strategy, offer users a premium alternative, and drive forward the development of the big data industry in China.