主要观点总结
阿里云PolarDB在TPC-C基准测试中刷新世界纪录，展现性能与效率的双重突破。其创新的「三层解耦」架构设计大幅提升效率，实现极致的弹性伸缩能力。该云原生数据库保证数据准确性、一致性和完整性，已支持全球超过10,000家企业的数字化转型。
关键观点总结
关键观点1: 阿里云PolarDB在TPC-C基准测试中创下新世界纪录
PolarDB以每分钟20.55亿笔交易超越原纪录，单位交易成本降低近40%。
关键观点2: PolarDB的架构设计实现高效弹性伸缩
采用「三层解耦」设计，计算、内存和存储的独立扩展大幅提升效率。
关键观点3: PolarDB保证数据准确性、一致性和完整性
在8小时压力测试中完成2.2万亿次数据操作，保证100%数据正确性，tpmC波动率低至0.16%。
关键观点4: PolarDB支持全球众多企业的数字化转型
包括TNG eWallet、TVB、enish, Inc.和DOKU等在内的超过10,000家企业客户使用PolarDB。
正文
阿里云PolarDB凭借创新的「三层解耦」架构刷新TPC-C基准测试世界纪录
阿里巴巴集团的数字技术与智能骨干业务阿里云，其云原生关系型数据库PolarDB在TPC-C基准测试中创下新世界纪录。TPC-C基准测试由事务处理性能委员会（TPC）设计，用于衡量联机事务处理（OLTP）系统的性能。
数据来源：https://www.tpc.org/tpcc/results/tpcc_results5.asp
在TPC-C测试中，PolarDB以每分钟20.55亿笔交易（tpmC）超越原纪录2.5倍，同时将单位交易成本（price/tpmC）降低至人民币0.8元（美元0.11），相比原纪录降低近40%。此外，在8小时的压力测试中，PolarDB完成了2.2万亿次数据操作，保证了100%的数据正确性，同时tpmC波动率仅为0.16%，比TPC-C基准所要求的2%波动率低一个数量级。
PolarDB在TPC-C基准测试中所达到的20.55亿tpmC，其处理能力相当于2020年天猫双11购物节期间交易峰值的59倍。
阿里云数据库产品事业部总裁李飞飞博士表示：「这一成就是我们团队在数据库技术领域不懈追求创新与卓越的最佳证明。PolarDB在TPC-C基准测试中的出色表现，展现了其应对极限工作负载的能力，更体现了我们致力于提供高性价比、可扩展解决方案的承诺。凭借创新的架构设计，云原生数据库的管理将变得如同搭建积木般简单。未来，我们将继续助力客户高效管理和运用数据，在数字时代蓬勃发展。」
TPC-C基准测试旨在评估数据库在极端条件下的性能表现。它不仅要求数据库在极大压力下保持数据准确性，还需在软硬件故障情况下保证数据一致性，并在压力条件下确保数据的可用性、一致性和完整性。TPC-C为业界公认最权威的OLTP基准之一。
PolarDB性能刷新记录，得益于其创新架构的强力支撑。PolarDB的Limitness架构采用「三层解耦」设计，通过实现计算、内存和存储的独立扩展，大幅提升效率，从而达到极致的弹性伸缩能力。PolarDB还透过软硬件结合及先进的数据库单机事务优化技术，如事务处理优化、索引结构优化和I/O路径优化等，进一步强化其卓越性能。
PolarDB是一款云原生关系型数据库，专为需要快速性能、高并发和自动扩展的关键业务数据库应用而设计。PolarDB已支持超过10,000家来自全球各行各业的企业客户的数字化转型，包括马来西亚最大的金融科技平台TNG eWallet；总部位于香港、全球最大商业华语电视制作公司之一的电视广播有限公司（TVB）；日本移动游戏开发公司enish, Inc.；以及印尼领先的支付技术公司DOKU等。
阿里云提供广泛的自研云原生数据库产品，包括关系型数据库PolarDB、数据仓库AnalyticDB、多模数据库Lindorm，以及数据管理服务（Data Management Service，DMS）等。阿里云连续五年获评为Gartner®云数据库管理系统（DBMS）魔力象限领导者。
Alibaba Cloud, the digital technology and intelligence backbone of Alibaba Group, has set a new world record with its cloud-native relational database PolarDB in the TPC-C benchmark. Developed by the Transaction Processing Performance Council (TPC), the TPC-C benchmark evaluates the performance of Online Transaction Processing (OLTP) systems.
Source: tpc.org
According to the TPC-C results, PolarDB achieved 2.055 billion transactions per minute (tpmC), 2.5 times higher than the previous record holder, while reducing the cost per transaction (price/tpmC) by almost 40% to CNY 0.8 (USD 0.11). Additionally, during an 8-hour stress test, PolarDB successfully processed 2.2 trillion data operations with 100% data accuracy, and maintained a tpmC fluctuation rate of just 0.16%, an order of magnitude lower than the TPC-C benchmark requirement of 2%.
PolarDB’s achievement of 2.055 billion tpmC in the TPC-C benchmark is 59 times the peak transaction volume recorded during the Tmall 11.11 Shopping Festival in 2020.
Dr Feifei Li, President of Database Products Business at Alibaba Cloud Intelligence, said: “This remarkable achievement is a testament to our team’s relentless pursuit of innovation and excellence in database technology. PolarDB's performance in the TPC-C benchmark showcases its capability to handle the most demanding workloads while emphasizing our commitment to providing cost-effective and scalable solutions. With innovative architecture, managing cloud-native databases will be as simple as ‘building blocks’. We will continue to support our customers to effectively manage and utilize data to thrive in the digital era.”
The TPC-C benchmark is designed to assess a database's performance under extreme conditions. It challenges a database to maintain data accuracy under significant stress, guarantee data consistency during software and hardware failure, and ensure data availability, consistency, and integrity under stress conditions. TPC-C is recognized as one of the most authoritative OLTP benchmarks in the industry.
PolarDB’s record-breaking performance is driven by its innovative architecture. Its Limitless Architecture features a "three-layer decoupling" design, which enhances efficiency by enabling the independent scaling of computing, memory, and storage for optimal elasticity. This exceptional performance is further boosted through software-hardware integration and advanced database kernel optimization technologies, including transaction processing improvement, index structure refinements, and I/O path enhancements.
PolarDB is a cloud-native relational database that is designed for business-critical database applications that require fast performance, high concurrency, and automatic scaling. This solution has supported the digital journey of over 10,000 enterprise customers across various sectors worldwide, including TNG eWallet, the largest fintech platform in Malaysia; Television Broadcasts Limited (TVB), a major player in Chinese TV program production based in Hong Kong; enish, Inc., a mobile game development company in Japan; and DOKU, one of Indonesia's leading payment technology companies.
Alibaba Cloud offers a broad range of self-developed, cloud-native database products, including the relational database PolarDB, data warehouse AnalyticDB, multimodal database Lindorm, and Data Management Service (DMS), among others. For the fifth year in a row, Alibaba Cloud has been named a Leader in Gartner® Magic Quadrant™ for Cloud Database Management Systems (DBMS).
