OceanBase A Cost Effective, Ultra Fast, Highly Scalable Distributed Relational Database - Why Distributed Databases - What is OceanBase - Key Competitiveness - OceanBase Cloud - Solutions - Case Studies TABLE OF CONTENTS �� Why Distributed Databases �� �� What is OceanBase �� �� Key Competitiveness �� Data Consistency, High Availability, and Flexible Scalability �� HTAP Capability �� Unified Architecture Highly Compatible Advanced Compression Technology Native Multitenancy �� OceanBase Cloud �� �� �� �� Solutions �� �� Scalable OLTP �� Database Consolidation �� Real-time Analytics Multi-cloud Architecture �� �� Key Features Available Zones �� �� Case Studies �� �� �� GCash �� DANA �� Ctrip Industrial and Commercial Bank of China Cainiao Ziel Home �� �� �� �� Why Distributed Databases Data Management Has Become Increasingly Complex due to Explosive Data Growth The development of cutting-edge technologies, such as cloud computing, Internet, artificial intelligence, Internet of Things, and blockchain, has spawned novel software and hardware products in various industries, leading to explosive data growth. Meanwhile, the mining and application of the value of massive data are becoming the main forces that drive innovation, and the complexity of data computing has increased significantly. Number of Terminals Connected to Internet Worldwide Unit: Billion �� ���� �� ����E �� �� ��� ��� Global CAGR: ��.�% ����E ����E ����E ����E *Data source: Gartner Challenges to Conventional DBMS in the Big Data Era Performance bottleneck Facing the business-driven growth in the data scale, conventional databases are incompetent in maintaining their performance, while the performance of distributed databases can be scaled indefinitely. Insufficient analytical capabilities Conventional database solutions require separate analytical systems, which leads to data redundancy and latency. Distributed databases support hybrid transaction and analytical processing (HTAP), which significantly improves analytical efficiency and reduces data redundancy. Total Cost of Ownership (TCO) is high Monolithic database systems are difficult to scale out and have to be designed per the maximum business traffic. Reliability comes at a high cost through over provisioning. However, a native distributed architecture supports flexible scaling and achieves high availability at a low cost. �� Why Distributed Databases Distributed Database Adoption around the World A research by International Data Corporation (IDC) shows that about ��.�% of the enterprise customers interviewed have deployed distributed databases. Among them, more than ��% express positive attitude toward distributed databases, in which about ��% have seen significant improvement in the database performance and are either satisfied or very satisfied with the distributed database systems. The figure below describes the reasons why the interviewed enterprises prefer distributed databases to conventional mololithic databases. Cloud and distributed architecture is the future and require planned strategic investment and require strategic investment in advance. ��% To improve the stability and availability of the database system. ��% To reduce hardware cost, thus bringing down the total cost of database system. To flexibly scale out/in hardware resources for better price-performance ratio. ��% Conventional databases cannot meet the high concurrency requirement in the age of digital mobility and smart devices. Conventional databases cannot efficiently manage massive and rapid growing data. ��% ��% In Other Vendors to Consider for Operational DBMSs published in ����, Gartner pointed out that distributed architecture was the future of database technology. A distributed architecture significantly improves the database capacity to store and manage massive data, well handles highly concurrent user requests, and provides automatic scaling in response to business fluctuations at much lower costs in comparison to a conventional database. �� What is OceanBase OceanBase is a highly cost effective, high performance, scalable native distributed relational database, ���% independently developed by Ant Group. It has been supporting Alibaba and key e-commerce events such as Double �� over the past years, and has been widely deployed in core systems among customers in various industries. OceanBase supports deployment model includes on-premise, cloud SaaS and multi-cloud, hybrid model, and can handle Hybrid Transactional/Analytical Processing (HTAP) with one database engine. OceanBase Database has been listed in "The Translytical Data Platforms Landscape, Q� ����", a report released by Forrester, which demonstrates OceanBase's analytical capabilities on top of strong OLTP. OceanBase is highly compatible with both Oracle DB and MySQL DB, users can easily migrate from conventional databases to OceanBase. World #� �� Years The only distributed database that tops Supporting Double Evelen Shopping Festival TPC-H record. time. TPC-C benchmark test and breaks the ��� + for �� years, with ��� million QPS at peak Serving over ��� customers and partners from various industries across the globe. �� What is OceanBase Why OceanBase? The enterprise-ready distributed SQL database has been supporting the Double ��, which is the world's biggest online shopping event with �� million QPS at peak time and over ��,��� nodes. It is designed to deliver ultra-fast performance, elastic scalability, and cost-effectiveness for your transactional and operational analytics workloads. Cost Effectiveness High compression technology saves ��%-��% storage space without compromising performance. Multi-tenancy achieves higher resource utilization. Ultra-fast Performance Supporting the Double �� with �� million QPS at peak time. Ranking No.� in TPC-C benchmark test at ��� million tmpC. Scalable OLTP An OceanBase cluster can be scaled out to �,��� nodes transparently, handling petabytes of data and a trillion rows of records. Oracle/MySQL Compatible OceanBase is highly compatible with Oracle/MySQL, which ensures that zero or few modification is needed for migration. �� Key Competitiveness As a native distributed database, OceanBase Database adopts a unified architecture that supports both distributed and monolithic mode. It has achieved significant breakthroughs in many aspects, such as automatic scaling, high availability, disaster recovery, storage engine, distributed transaction processing, HTAP, compatibility with mainstream databases, and multitenancy. OceanBase database has been operating in complex and demanding sectors such as financial core systems, as well as massive user traffic during ��.�� eCommerce festival. Native Distributed Architecture Ensures Data Consistency, High Availability, and Flexible Scalability Instead of using SAN storage devices that require a dedicated network, OceanBase Database can be deployed in a shared-nothing cluster based on commodity servers and data center network. An OceanBase cluster automatically manages the allocation and dynamic balancing of computing and storage resources and supports automatic scale-out or scale-up with linearly improved read and write performance. SQL execution and data storage are supported on all service nodes, and partition replicas are autonomously managed by each node. The cluster runs only one type of database service processes, without depending on external services, which ensures easy Operation & Maintenance (O&M) management. OceanBase Database provides unified database services to applications, supports ACID and global indexes. This means it performs as a standalone system in application development. Strong Data Consistency OceanBase Database provides a variety of built-in mechanisms for strong consistency verification. These mechanisms guarantee data reliability by automatically identifying errors, such as inter-replica data inconsistencies, network errors, silent data corruption, and table-index inconsistencies. System High Availability OceanBase Database has strong disaster recovery capabilities. Exceptions in the same IDC or interruptions across IDCs do not cause data loss or service downtime. Transparent Scalability An OceanBase cluster can be scaled out to include thousands of nodes, with automatic load balancing mechanisms. The process of scaling is transparent to applications. �� Key Competitiveness The Unified Architecture Adapts to Business Growing Stages OceanBase Database adopts a unified architecture. When deployed in monilithic mode, it delivers nearly the same performance as conventional monolithic databases. When deployed with three replicas, it provides higher availability than a conventional database that delivers the same performance in the primary/standby mode. OceanBase Database supports linear vertical scaling by increasing the node specification and linear horizontal scaling by deploying more nodes in a zone. Such a unified architecture allows OceanBase Database to grow with business growth and be adaptable to the needs of different customers, including small personal sites, core banking systems, and giant e-commerce platforms. Business Scale Small Business States Singapore United Low-spec Server OceanBase OB Indonesia Medium Business High-spec Server OB Core Business Large Business High-spec Server High-spec Server OB OB High-spec Server OB OB OB OB OB OB OB OB OB OB OB OB OB �� Key Competitiveness HTAP Capability: A Unified Engine for Both OLTP and OLAP Workloads HTAP is a highly expected database capability. With its distributed architecture, OceanBase Database has strong transaction processing capability while handling analytical tasks such as data analysis and batch processing. OLAP and OLTP are supported by the same set of engine, which reduces the costs significantly. Thanks to the native distributed technology, OceanBase Database provides high-performance OLTP capabilities. It integrates both transaction processing and real-time data analytics capabilities in one system, which means that the same set of data can be used for different tasks. By doing so, it guarantees data consistency, minimizes data redundancy, and significantly reduces the total O&M costs for enterprises. Separated Systems: High Workload HTAP Engine: Hybrid Workload in One System Step �: OLTP Requests Async Transfer Step �: OLAP Requests OceanBase Cluster OLTP Requests + OLTP Requests To meet HTAP requirements, OceanBase Database also provides various resource isolation methods, such as physical isolation of multiple zones and isolation of database connections based on CPU resource groups. OceanBase Database also automatically identifies and isolates slow queries to prevent them from affecting the overall transaction response time. Online Transaction System Data Analytics System Computing Resources Data Storage �� Key Competitiveness Highly Compatible with Mainstream Databases OceanBase Database is compatible with mainstream open source and commercial databases. SQL syntax, commands, objects, operation habits, stored procedures, C APIs, and precompilers are supported in OceanBase Database. Applications can be migrated to an OceanBase database with a few or even no changes, saving considerable labor and time costs. In addition, users can create different database tenants in the same OceanBase cluster. Open-source Database Different database tenants in one OceanBase cluster Complete compatibility with data types, SQL syntax, and operation habits Commercial Database Support for PL/SQL, OCI, Pro*C, and other advanced features Database � Tenant OceanBase Cluster Database � Tenant Advanced Compression Technology Significantly Reduces Storage Costs Data compression is key to the reduction of massive data storage. OceanBase Database implements a distributed storage engine featuring a high data compression ratio. Thanks to the adaptive compression technologies, the LSM-tree-based storage engine balances the system performance and compression ratio in a creative manner, which is impossible in a conventional database that stores data in fixed-size chunks. Moreover, data and logs are separately stored in the system to further reduce the storage costs. The LSM-tree-based storage engine greatly reduces the storage costs by using encoding and compression technologies. For the same amount of data, the storage space required in OceanBase Database is much lower than that in MySQL and Oracle, approximately �/�-�/� OceanBase Database significantly improves the stability and security of a business system and reduces storage costs by ��%-��% �� Key Competitiveness Native Multitenancy Ensures Resource Isolation and Flexible Management OceanBase Database adopts a native multitenant architecture, which supports multiple tenants in one cluster. Each tenant can be considered as a database service. Data and resources of one tenant are isolated from other tenants, and tenant resources are uniformly allocated within the cluster. Users can create different database tenants and configure the number, type, and storage location of data replicas as well as computing resources for each tenant. Suitable for Microservice Architecture As business system of an enterprise becomes more complex, a monolithic service architecture will cause overwhelming workload in engineering and management. If a microservice architecture is used, the enterprise can add and adjust services only by adding new microservice nodes. However, as each microservice needs a separate database, more services require more databases, leading to challenges for database reliability and O&M management. With the multitenant architecture of OceanBase Database, the administrator needs to maintain only a small number of clusters, with guaranteed isolation of tenant data and resources and improved database stability. Suitable for Multitenant SaaS Services Generally, software-as-a-service (SaaS) cloud vendors provide multitenant services. If the spaces of multiple business tenants are logically isolated by tenant name in a single database, operations of different tenants can affect each other. If each business tenant uses a separate database, the large number of scattered databases can be a real headache for O&M engineers. The scalability is limited, let alone the high costs. The native multitenant architecture of OceanBase Database provides a better balance between resource isolation and costs, and tenants of any specification can be scaled in and out separately. Application � Application � Application � Oracle MySQL Oracle Resource Isolation, Data Isolation, and Elastic Scalability OceanBase Cluster CPU CPU CPU Memory Memory Memory Resource Pool � Resource Pool � Resource Pool � Storage Storage Storage Each application tenant has a dedicated resource pool. �� OceanBase Cloud OceanBase Cloud provides fully managed database services with elastic scalability, ultra-fast performance, and high compatibility on global cloud infrastructure. OceanBase Cloud provides database solutions on the cloud, covering support, scaling, monitoring, diagnosing, development, migration, backup and restore. Cloud Service Tools Public Cloud Service Cluster version Monolithic version OMA Multicloud Service OCP ODC OMS History version Third-party tools OceanBase Database Kernel Database Paxos multi-replica Online DDL Data compression Cloud Infra Beijing Data compression Concurrent processing Distributed transaction High availability ··· Other Cloud Vendors AWS Alibaba Cloud Shenzhen Transparent scaling HTAP Hangzhou Shanghai Indonesia Singapore United States Cluster version O&M Console Cluster Management Tenant Management Performance Monitoring Warning Management Capacity Management Backup Recovery SQL Diagnose SQL Audit Security Management History Events �� OceanBase Cloud Key Features Multi-cloud Deployment Support deployment on multiple cloud vendors, including AWS, Alibaba Cloud, Tencent Cloud. You can easily manage databases across cloud platforms via a unified console. Multi-tenant Flexibility The native distributed multi-tenant architecture isolates data and roles among tenants. It allows independent tenant configuration for replica number, location, and compute resources for improved hardware and storage resource utilization. Real-time Operational Analytics OceanBase uses one copy of data for both transaction and operational analytics workloads, which ensures real-time analytics and strong data consistency. ��.��% SLA For On-Premises and Hybrid Cloud deployment, the revolutionary "Three Regions, Five IDCs" architecture supports a region-level disaster recovery with zero data loss and application recovery time less than �� seconds. Available Zones �� available regions �� available zones ����+ nodes Available regions Available soon Available zones cover major countries and regions across the globe, meeting the requirements of global enterprises. �� Solutions Scalable OLTP Auto Scaling | High Performance | Cost Effective OceanBase Database, adopting a next-generation native distributed architecture, provides users with strong performance, high compatibility, high scalability, and low cost through technologies such as multi-tenant, partitioned transparent scaling, advanced coding compression, helping customers achieve higher efficiency at lower costs. The Challenge With data grows significantly, traditional monolithic databases experience performance bottleneck, which requires scaling out or scaling up. However, the conventional monolithic database cannot be horizontally expanded and has to be split into multiple physical instances. This makes the number of instances as many as tens or even thousands, posing a serious challenge to operation and stability. Heterogeneous storage systems According to various business needs in different scenarios, an enterprise often needs to deploy different databases, such as Oracle, MySQL, DB�, and PostgreSQL. These different systems are complex to operate and maintain, which brings large investments in resource and human power. High resource costs The conventional standalone database cannot be horizontally expanded and has to be split into multiple physical instances. This makes the number of instances as many as tens or even thousands, posing a serious challenge to operation and stability. As the number of instances grows, the cost of software licenses rises accordingly. Resource fragmentation The utilization of resources, such as CPU and memory, varies in different scenarios. Sometimes resources are not sufficient, but sometimes a large amount of resources are idle. Resources cannot be "pooled" and cannot be managed in a centralized way, leading to a serious waste of resources. The Solution With OceanBase Migration Service (OMS), data can be quickly migrated to OceanBase Database without any downtime, with minimal or even zero modification. Resources are isolated on a per tenant basis and data is maintained independently between business modules. The resources of a tenant can be increased or decreased as needed, which is transparent to applications. OceanBase's large cluster mode integrates multiple instances into one cluster for unified management, flexible management and scheduling, significantly improving resource utilization. With the powerful leader distribution and read/write routing strategies for clusters, OceanBase has proven experience in managing high concurrency systems, which has been adopted in Ant Group for years. �� OceanBase Database comes with a powerful intelligent control console that automatically detects and analyzes typical issues, which significantly improves O&M efficiency. Tenant � Tenant � Tenant � Tenant � AZ � AZ � OB Node � OB Node � OB Node � Leader Follower Follower Follower Leader Follower Leader Follower Follower Follower Follower Leader MySQL/Oracle �TB AZ � MySQL/Oracle �TB Benefits High compression ratio Using LSM-Tree based self-developed storage engine in combination with data encoding and compression algorithms, OceanBase Database achieves a high compression ratio of more than ��% (compared with traditional databases), significantly reducing storage costs. High performance OceanBase Database uses the shared-nothing architecture and distributes leader replicas across nodes to implement multi-node data write, which makes it easy to handle ultra-high concurrency requirements. In addition, it uses the industry-leading table group technology to optimize the processing of a distributed transaction, achieving an execution efficiency as high as that of a single-node transaction. High compatibility OceanBase Database supports Oracle- and MySQL-based tenants in the same cluster. It is compatible with most features of Oracle and MySQL, such as SQL syntaxes, stored procedures, functions, C APIs, and precompilers, minimizing business migration and modification costs. Multi-tenant OceanBase Database realizes kernel-level virtualization and isolates resources and data between tenants. Different tenants can share the same cluster, which significantly improves resource utilization and reduces costs. High scalability An ultra-high-throughput OceanBase cluster can be built with common PC servers. The cluster can be flexibly scaled as needed without sharding, so that cost growth is predictably smooth. �� Solutions Real-time Analytics T+� HTAP | Multi-model Support OceanBase Database uses one set of database engine to process both online transaction processing (OLTP) and online analytical processing (OLAP) tasks. This avoids the time-consuming ETL process in real-time data warehouses. The Challenge Poor real-time performance Offline data warehouses often only provide T+� data capability. Moreover, the separate construction of real-time data warehouse is relatively complicated for small and medium-sized enterprises, and requires additional human power costs to maintain the data synchronization link. Poor resource isolation If resource isolation in an HTAP architecture cannot be guaranteed, once the database that undertakes key transactions is affected by analytical and batch-type business, it often directly leads to a drop in online transactions and serious damage to business. Limited capability Due to the synchronization latency, it is impossible to guarantee real-time data consistency in traditional primary/standy cluster in case of large transactions or DDL operations. SQL optimizer of conventional databases cannot support complex analysis of multiple tables across shards. The Solution OceanBase Database's HTAP mode makes it possible for direct business analytics query without waiting for T+� data, enabling real-time marketing analytics directly in the online library. MySQL 25-49 Partitioned Table Leader MySQL 50-74 MySQL 75-99 Real-time Data Warehouse Queries OLAP Core Transactions OLTP MySQL 00-24 OceanBase Cluster Follower Follower �� Benefits Strong Data Update Capability Relational database-based data update capability with very low latency between replicas. Multi-model Support Support HBase and Table model APIs, and supports semi-structured data formats such as JSON. HTAP Capability Provide a reliable TP/AP resource isolation solution with the resource group isolation technology. Distributed Parallel Computing Engine The Powerful SQL optimizer supports parallel computing and analysis of massive data. Flexible Scalability A single cluster can have a maximum of �,��� nodes, with data capacity reaching petabyte level. �� Solutions Database Consolidation Resource Pooling | Manageability | Cost Optimization With the distributed multi-tenant architecture, OceanBase Database achieves service-oriented multi-database consolidation, and realizes resource pooling while ensuring resource isolation and system availability, which significantly improves resource utilization and management efficiency and helps SaaS enterprises reduce their overall costs. The Challenge Instance fragmentation Multiple different business applications within an enterprise, or the resource isolation needs of SaaS enterprises for different customers, may lead to the need to deploy a large number of database instances, resulting in serious fragmentation. When key business or customer demands surge, it is difficult to scale out flexibly and to guarantee performance and availability. Resource waste Due to the fragmented deployment of database resources, an enterprise often reserve an extra portion of resources for an instance for request growth in a short period of time in order to ensure certain performance elasticity. Such resource reservation actually causes huge resource waste from the overall business perspective and raises the cost of enterprise resources. Complex management A large number of database instances increases management complexity. It is difficult for the DBA team to put hundreds or even thousands of instances under fine-tuned management. In case of a failure, jitter, or other events, it may take a long time for the DBA team to recover services. As a result, the management costs may rise. The Solution The multi-tenant solution helps improve resource utilization by integrating database instances of different businesses. With the multi-replica design based on the Paxos protocol, high availability can be ensured for each resource unit. The multi-tenant solution not only applies to large and medium-sized enterprises where resources of different business lines need to be put into a resource pool, but also applies to SaaS enterprises where instances of different specifications need to be provided for different customers, which ensures resource isolation at lower costs. �� Small-sized merchant A Medium-sized merchant B Daily business growth Small tenant � � cores, � GB memory Medium tenant � � cores, �� GB memory Large tenant � �� cores, ��� GB memory Large-sized merchant D Scale-out for major promotions Small tenant � � cores, � GB memory Auto Scaling Large tenant � �� cores, ��� GB memory Medium tenant � �� cores, �� GB memory Medium-sized merchant C Evolved into a large-sized merchant In-place Scale-out OBServer � Medium tenant � �� cores, �� GB memory Auto Scaling Large tenant � �� cores, ��� GB memory Medium tenant � �� cores, �� GB memory Small tenant � � cores, � GB memory OBServer �( Auto Scaling) Small tenant � � cores, � GB memory OBServer �( Auto Scaling) Benefits Resource pooling: OceanBase cluster natively implements resource isolation and virtualization in the multi-tenant solution. Hundreds of database instances can be deployed in one cluster, with data and resources isolated for each instance and computing resources scaled out in seconds. Flexible scaling: With the multi-node distributed architecture, you can immediately upgrade a single server, and flexibly scale to multiple servers with auto load balancing. OceanBase Database automatically completes the data migration and multi-dimensional data verification without human intervention. Unified Management: AInstead of managing several hundreds of scattered instances, DBAs now manage clusters in OceanBase. Data loading, alarming, and tuning are performed on a per cluster basis. Minor failures can be automatically recovered, significantly improving maintenance efficiency and emergency response speed. Cost Effectiveness: Resource pooling improves resource utilization. And the advanced compression algorithm reduces storage costs by �/�, compared with traditional databases. The total cost of ownership (TCO) can be reduced by about ��% to ��% for large and medium-sized enterprises. �� Solutions Multi-cloud Architecture Multi Cloud | Hybrid Cloud | Unified Management OceanBase can provide customers with a multi-cloud, cross-cloud, or hybrid cloud deployment architecture, which achieves cloud vendor-level disaster recovery. The Challenge ‒ Single point failure: In the case of a black swan event, such as crash, network failure, or power outage, single-cloud deployment may encounter service interruption and even data loss or capital loss, which is unbearable for many enterprises. ‒ Asynchronous data replication: In a traditional database, cross-cloud data replication is asynchronous, which requires manual operation and maintenance (O&M). In the case of an IDC-level disaster, data loss or capital loss may occur. ‒ Cloud infra depandency: Some cloud-native databases depend heavily on cloud vendor-specific hardware and infrastructure. Otherwise, these databases cannot provide services or can only provide limited services. The Solution OceanBase provides public cloud services that are compatible with the cloud infrastructure of multiple cloud vendors. These services are managed in an all-in-one console. OceanBase provides consistent product features, interfaces, accessibility, and user experience for different cloud deployment modes. OceanBase connects on-premises systems with public clouds over secure network facilities. If your service encounters heavy load, failure, or disaster, the service can be quickly switched to other cloud services. For users in areas prone to natural disasters or users seeking cloud neutrality, an OceanBase cluster can be deployed across different clouds by connecting the virtual private clouds (VPCs) of different cloud vendors. If a single cloud becomes unavailable, the service can be quickly recovered, with zero data loss. �� Independent, prviate cloud, and public cloud deployment Physical Resource Self-built IDC A Multi-cloud BareMetal Prviate Cloud A Public Cloud Existing applications OceanBase DBMesh New generation core Existing applications IaaS New generation applications、 applications Paas OceanBase DBMesh Online business IaaS Paas OceanBase DBMesh Benefits ‒ Multiple Deployment Modes: OceanBase supports multi-cloud and hybrid cloud deployment. An OceanBase cluster can be deployed across different clouds, which achieves cloud vendor-level disaster recovery. ‒ Unified Management Console: OceanBase provides a unified multi-cloud O&M console for users to manage all instances, which significantly improves management efficiency and reduces risks. ‒ DBMesh: OceanBase Database Proxy (ODP) supports transparent access to OceanBase, which shortens the access link and improves the database performance and availability. ODP supports applications coded in different programming languages via SQL proctocol. �� Case Studies GCash GCash is one of the leading mobile payment providers and the largest e-wallet in the Philippines. By ����, it had more than �� million users. On average, one out of every three Filipinos is a GCash user. Challenges There was a bottleneck in the scalability of the MySQL database. Engineers were required to work overnight, on average � days a week, to perform database sharding and data cleaning, which resulted in frequent business interruptions lasting � to � hours each. The large number of MySQL instances increased operation and maintenance (O&M) costs. Solution OceanBase Migration Service (OMS) is adopted to migrate the whole business to OceanBase Database with zero downtime. The storage architecture is upgraded to optimize the overall database architecture, providing strong support for GCash's long-term development. Zone � Zone � Zone � Billing Tenant � Tenant � Tenant � Marketing Tenant � Tenant � Tenant � Transaction Tenant � Tenant � Tenant � Payment Tenant � Tenant � Tenant � Accounting Tenant � Tenant � Tenant � Financial Network Tenant � Tenant � Tenant � Real-time Data Tenant � Tenant � Tenant � Core Clusters Bypass Clusters �� Benefits OceanBase helps GCash reduce costs, improve efficiency, and implement financial-grade disaster recovery during periods of rapid business growth. High compression ratio: After the business is migrated to OceanBase Database, the required storage space is reduced by ��% due to the high compression ratio of the database. High availability: OceanBase Cloud's three-zone high availability architecture ensures business continuity in the event of any single-zone failure. High resource utilization: OceanBase Database's large cluster mode significantly improves resource utilization. After the business is migrated to OceanBase Database, the resource cost of the database is reduced by more than ��% on average. Improved O&M efficiency: OceanBase Database significantly improves the O&M efficiency of database administrators (DBAs) based on its kernel-based online DDL and intelligent diagnostics features and the cloud console. �� Case Studies Ctrip Founded in ����, Ctrip is an online ticketing service company headquartered in Shanghai, China. As a leading platform for hotel booking in China, Ctrip has partnered with more than ���,��� hotels all over the world, and went public on Nasdaq in December ����. On October ��, ����, the company officially rebranded as Trip.com Group. On April ��, ����. Trip.com was listed on the Hong Kong Stock Exchange. Challenges The large-scale business traffic of customer service exceed the data storage limitation in MySQL, which can only covers the data storage for two months. The active and standby architecture of MySQL is unable to expand storage by horizontal scale out. The active and standby architecture of MySQL cannot guarantee the data consistency and rapid failover in case of an IDC failure. Solution OceanBase replaces MySQL with nearly ��% data compression capability to cover the data storage for one year, compared with � months before. The horizontal scale out capability of OceanBase can deal with the business by adding or reducing machines without business interruption. The multi-replica consistency by Paxos consensus guarantees RPO=�, RTO<�mins. The data migration and synchronization capabilities of OceanBase Migration Service (OMS) ensures the smooth and stable database switchovers. �� IDC � Original Architecture Business Application IDC � Business Application IDC � Primary WriteOnly Database Primary WriteOnly Database Primary WriteOnly Database Business Application Business Application Business Application Business Application Business Application Business Application Primary/standby Write-Only Databases Primary/standby Write-Only Databases Write-Only OceanBase Database Replica � Write-Only OceanBase Database Replica � Write-Only OceanBase Database Replica � Primary/standby Write-Only Databases IDC � New Architecture Business Application Write-Only OceanBase Database Replica � IDC � IDC � Write-Only OceanBase Database Replica � Write-Only OceanBase Database Replica � Business Application Benefits More than one year of data storage capability can be obtained under the same hardware investment. Automatic failover to improve the rapid recovery capability for nodes or IDC failure, and ensure SLA. Strong data consistency, and reduce the risk of data loss. The ultra-high compression ratio, multiple replicas, and strong consistency of OceanBase improve the quality of customer services and reduce the cost. �� Case Studies DANA DANA is one of the leading wallet providers in Indonesia. DANA provides digital payment and financial services to more than �� million users of BBM, the second largest chat tool in Indonesia. Challenges With rapid growth of DANA Wallet's business, its initial small-scale cluster architecture has urgent needs to increase capacity. However, the asynchronous replication in the MySQL architecture cannot guarantee a high success rate and data consistency in the high-availability switchover. Solution Based on OceanBase Database's �-replica architecture, DANA deployed � replicas in the local IDC and � replica on the public cloud to build a hybrid cloud infrastructure, based on OceanBase Database's �-replica architecture. OceanBase Database's native horizontal scalability enables the upgrade from the �-node cluster (�-�-�) to the �-node cluster (�-�-�) without business interruption, significantly improving performance and capacity. OceanBase Migration Service (OMS) migrated hundreds of instances from the MySQL database to OceanBase Database on the public cloud. Application Application Rollback OceanBase Database (Primary) Oracle/MySQL Database (Standby) OceanBase Database (Standby) Oracle/MySQL Database (Primary) �� Benefits OceanBase Database helps DANA make business breakthrough with its unique features such as elastic horizontal scalability, high availability, and financial-grade data consistency. Since its business migration from MySQL to OceanBase Database in ����, DANA no longer suffers from capacity and availability issues and is able to fully cope with its rapid business growth. OceanBase Database now supports tens of millions of DANA users. Since ����, DANA has witnessed zero database failure and zero data loss, as well as an overall business system availability of over ��.��%. DANA's third IDC under construction is built on the public cloud, which truly realizes a tri-IDC hybrid cloud deployment and further improves disaster recovery capability and service availability. With a comprehensive OceanBase Database training program based on theory and practice, DANA has developed a local team of OceanBase Database administrators (DBAs). �� Case Studies Industrial and Commercial Bank of China Industrial and Commercial Bank of China (ICBC) was established on January �, ����, and was successfully listed on both the Shanghai Stock Exchange and the Hong Kong Stock Exchange on October ��, ����. ICBC is the No.� Bank for multiple consecutive years in The Banker's Top ���� World Banks and Forbes Global ����. Challenges High disaster recovery requirements: ICBC Asset Under Management (AUM) is trillions of dollars, and therefore, ICBC requires ��/� services and Level � disaster recovery capabilities. High hardware requirements to operate legacy database: The existing business systems run on conventional mainframes and DB� databases which are costly to procure and maintain. Monolitic architecture leads to poor utilization of computing resources. As ICBC business grow, it requires increasing over provisioning of standby resources to cater for disaster recovery purpose. Due to the monolithic architecture from legacy databases, the over-provisioning leads to extremely high cost in wastage and equipment operating cost. Solution OceanBase Database provides native support for multiple data replicas. Data is replicated across nodes based on the Paxos consensus to achieve high availability of the cluster and geo-disaster recovery. Replicas with different zones can be combined to deploy as three IDCs across two regions or five IDCs across three regions to achieve IDC and region level disaster recovery capabilities. The solution implements region-level disaster recovery with RPO of less than two minutes and RTO of less than two hours when it is deployed with primary/standby geo-disaster recovery mode. OceanBase Database auto balances the work load to the least utilized and most suitable node in the cluster based on unified real time resource scheduling. The fault management service automatically detects the faulty nodes and route the transactions to healthy nodes, ensuring global strong consistency of distributed transactions without a need for manual intervention. �� IDC - � OCP-related Domain Name Resolution Primary City Standby City IDC - � IDC - � OCP-meta OBServer � OCP-meta OBServer � OBProxy OBProxy OCP OCP Business Cluster Business Cluster OBServer � OBServer � (Primary) (Primary) Business Cluster Business Cluster OBServer � OBServer � (Primary) (Primary) OBProxy � OBProxy � OBProxy � OCP-meta OBServer � Business Cluster Business Cluster OBServer � OBServer � (Primary) (Standby) OBProxy � OBProxy � F�/SLB F�/SLB F�/SLB APP APP APP Benefits OceanBase deployment in three IDCs deployed across two regions meets the level � disaster recovery requirements of ICBC and supports ��/� non-stop services.The traditional mainframe computers are replaced with Advanced RISC Machine (ARM) servers, which reduces the the system cost. OceanBase database restructures the original database systems of ICBC into a distributed system, increasing the resource utilization of the database nodes to up to ��%. When the processing capacity of the system is severely bottlenecked, the system can be automatically scaled out to increase the resources of the cluster to improve the processing capacity. �� Case Studies Cainiao Cainiao Smart Logistics Network Limited, formerly known as China Smart Logistics Network, is a Chinese logistics company launched by Alibaba Group, jointly with eight other companies, on �� May ����. As of May ����, Cainiao was one of the largest unicorn companies in China, valued at ��� billion yuan. Cainiao ensures delivery within �� hours to any region of China. The company also shares resources with other logistics companies. Challenges Cainiao Delivery Management System serves as an order processing and delivery platform for merchants to record, pull, review orders and track packages. During order review, multiple orders from the same buyer or recipient need be automatically merged. This makes it convenient for the merchant to decide whether to pack the products into one delivery package to reduce logistics costs, especially during major promotion campaigns. The management page needs to display the orders that can be merged and allow the merchant to search by various meta-data. In early version of Cainiao Delivery Management System, the transaction and order data is synchronized to the corresponding tables stored in an AnalyticDB for MySQL (ADB) database. When a query is initiated, the database joins the two tables to obtain and aggregate the waybill numbers based on the buyer ID, recipient name, recipient contact information, recipient address, transaction type, and associated package. Then, the database returns a list of aggregation results and statistics. However, with the rapid business growth, this solution reaches its performance bottleneck. After rounds of hybrid I/O optimization, the aggregate query performance can reach only around ��� QPS, which is too slow to support the business growth. In addition, the cost to operate the database is too high. To ensure the data write and aggregate query performance, the cluster uses ��� nodes. Each node is configured with � CPU cores and costs more than RMB ��,��� a year, which means that the database cluster costs about RMB � million (approx ~US$���K) a year. To meet the revenue goals of the next fiscal year, a new solution is required. Solution The new solution meets the following requirements: a. The aggregate query performance reaches �,��� QPS under a data write concurrency of higher than ��,��� TPS. b. The real-time processing of incremental data delivers acceptable performance. c. Data can be filtered and searched based on multiple metrics or attributes. d. A cost effective solution which can satisfy the above requirements �� OceanBase Database delivers the expected read/write performance in order merger. OceanBase Database stores the incremental data in a MemTable in the memory and the baseline data in an SSTable in the SSD. Data modifications generate incremental data, which is written only to the memory. In addition, subpartitioning splits data by user ID and order creation time, to improve read performance. OceanBase Database supports three replicas, Paxos-based log synchronization, strong-consistency reads specified by hints, and accurate statistics at low costs. Cainiao Delivery Management System allows valuation application to query the historical order statistics of merchants or buyers during a major promotion campaign, supports the valuation application to calculate the shipping prices, classifies the historical orders as required by the merchant, and displays the historical order statistics in real time. Benefits This upgrade lays the groundwork for business promotion in the future, and breaks the performance bottleneck of order merger. The aggregate query performance is improved by ���% from ��� QPS to more than ���� QPS, and the time consumption of a list query is reduced from ��� ms to �� ms, which brings better experience to users. Also, the estimated annual costs are reduced by ��% from � million CNY of the ADB-based solution to �.� million CNY of a solution based on OceanBase Database and ElasticSearch. After Cainiao deploys OceanBase Database, the system of Cainiao has experienced significant performance improvement, and has demonstrated stable and efficient package handling capabilities both in major promotion campaigns and daily transactions. Users spend less time for package delivery and pick-up. The new solution offers much better user experience and is more efficient. �� Case Studies Ziel Home Ziel Home Furnishing Technology Co., Ltd. is a cross-border e-commerce market leader from China. Ziel Home sells home furnishing products to more than �� countries, such as Western Europe, North America, and Japan. The company has garnered excellent customer reviews across various marketing channels. At the moment of writing, Ziel Home is the top seller of home furnishings on Amazon Europe and is one of the largest B�C cross-border e-commerce exporters from Chinese mainland. In ����, the company achieved a revenue of nearly RMB � billion, a ���% increase year-on-year.x Challenges As Ziel Home business grows, its current legacy database becomes a technical bottleneck. Ziel Home’s old database cannot cope with high concurrency and high transaction traffic. Vertical scaling of its database is impossible as database have reached its performance limit. Ziel Home generates data from multiple platforms across the globe and Ziel Home has difficulty in aggregating data for analysis and decision-making. Ziel Home realized that it requires a technology stack that can help them to aggregates data from multiple cloud platforms. Ziel Home requires a database management solution which can scale across different countries as the business expands to more sites around the world with high data availability and reliability for its database solution. Database O&M becomes more complex and expensive due to numerous types of databases and large number of instances. The variety of business systems Ziel Home operates results in high O&M cost and administrative strain on its engineers. Database O&M costs are high. Ziel maintains database instances across multiple platforms such as AWS Europe, AWS North America, Alibaba Cloud Hong Kong (China), Alibaba Cloud Europe, and self-built IDCs. The architectures and resources of database instances vary with the platform. The geographical and temporal differences make the O&M management more complex in both technical and financial areas. Ziel Home maintains database instances across multiple cloud platforms such as Europe, North America, Alibaba Cloud Hong Kong (China), Alibaba Cloud Europe, and self-managed IDCs. Due to multiple platform operations, across multiple regions, Ziel Home operates different architecture and database specifications and this adds O&M complexity, cost and overhead burden. Solution OceanBase database’s scalability can achieve linear performance development of business by adding nodes horizontally, without sharding database. �� OceanBase multi-tenants architecture supports multi-cloud hybrid deployment and enable unified management of cloud or on-premises databases through OceanBase Cloud Platform (OCP) console. OceanBase operates with high availability and is based on architectures which can operate high disaster recovery capabilities. Data are stored on a high compression ratio due to OceanBase encoding technologies and through utilizing multi-tenant architecture, OceanBase can achieve higher cost efficiency and lower O&M burden. New business systems can now be deployed and launched quickly, and legacy systems can be migrated seamlessly. Integrate data from multiple cloud platforms using OceanBase Migration Service (OMS). RDS Europe AWS PostgreSQL MongoDB MySQL Self-developed OB OMS OMS SQLServer RDS PostgreSQL MySQL Hong Kong Alibaba Cloud Cloud service OceanBase Self-developed OceanBase OMS North America AWS MongoDB Hong Kong Alibaba Cloud Self-developed OB OMS OMS SQLServer Benefits With high compatibility with MySQL, adopting OceanBase allows Ziel Home to combine data from multiple MySQL and RDS instances and integrate into one cluster. Furthermore, using data compression it allows operation and maintenance cost and administrative strain to be reduced. OceanBase database’s horizontal expansion capability enables elastic scalability which allows Ziel Home to handle rapid growth of its business traffic. With HTAP capability, OceanBase can reduce synchronization problem, migration of data for analysis, and enable real-time data reporting. �� About OceanBase OceanBase is a cost effective, high performance, and highly scalable native distributed relational database, ���% independently developed by Ant Group. It has been supporting Alibaba and key e-commerce events such as Double �� over the past years, and has been widely deployed in core systems among customers in various industries. OceanBase supports deployment model including on-premise, cloud SaaS and multi-cloud, hybrid model, and can handle Hybrid Transactional/Analytical Processing (HTAP) with one database engine. OceanBase is highly compatible with both Oracle DB and MySQL DB, users can easily migrate from conventional databases to OceanBase. OceanBase en.oceanbase.com contactus@oceanbase.com Twitter: @OceanBaseDB