The article traces the evolution from monolithic Java apps to RPC services, explains why sharding and database partitioning cannot alone achieve limitless scaling due to connection limits, and introduces unitization—assigning each application to a specific database shard—to overcome the bottleneck.
This paper presents the design, implementation, and experimental evaluation of OceanBase's unitization architecture for large‑scale online map services, demonstrating superior disaster‑recovery, high‑throughput OLTP/OLAP performance, and storage efficiency compared with competing distributed databases.
This article explains ByteDance’s multi‑region unitization approach, covering its core concepts, motivations, architectural challenges, traffic routing, data synchronization, cut‑over strategies, and future evolution for large‑scale, resilient services. It also discusses operational optimizations, risk controls, and the impact on cost and development efficiency.
To meet exploding transaction volumes, rapid innovation cycles, and strict regulatory demands, large banks are replacing mainframe core systems with distributed, horizontally‑scalable architectures, and Baidu’s GaiaDB‑X database—offering strong ACID consistency, zero‑RPO disaster recovery, and automated operations—has successfully powered core banking migrations for institutions such as Bank of China and state‑owned banks.
This article explains the concept of unitized architecture, its characteristics and types, the performance and reliability challenges it solves for large‑scale microservice systems, and how Tencent Cloud’s TSF platform implements unit routing, gray release, and disaster‑recovery to achieve efficient, cross‑region, high‑availability deployments.
ShopeePay replaces its traditional active‑standby design with a unit‑based multi‑active data‑center architecture—splitting user, public, and order data across independent units, employing routing tables, caches, and a dual‑active model that enables seamless failover, reduced latency, and continuous payment processing across two or more IDC sites.
The article explains multi‑active deployment modes—same‑city, cross‑city, and cross‑city multi‑active—along with routing rules (random, user‑ID, region) and unitization, describing how sharding IDs map traffic to central or unit data centers, and detailing middleware solutions for storage, messaging, SOA, and Snowflake ID generation to achieve scalable, highly available architecture.
This article explains how Ant Financial’s Alipay scales to hundreds of thousands of transactions per second during Double 11 by using logical data centers (LDC), unitized architecture, CAP theorem analysis, OceanBase database, and multi‑region disaster‑recovery strategies.
The article explains how Ant Financial scales Double‑11 payment traffic to hundreds of thousands of TPS by employing logical data centers (LDC), unit‑based system design (RZone, GZone, CZone), database sharding, CAP theorem analysis, Paxos‑based consensus, and the OceanBase distributed database, while also detailing disaster‑recovery and traffic‑shifting mechanisms.
The article traces the evolution of backend services from monolithic applications through RPC-based scaling and sharding, highlights the limitations of database connection counts, and proposes a unitized architecture that partitions databases per service to achieve true unlimited scaling while acknowledging added complexity.
The article explains how traditional scaling methods like sharding and RPC eventually hit connection limits, and introduces unitization as a strategy to achieve unlimited service expansion by limiting each application to a single database instance.
The article explains why simple sharding and database partitioning cannot achieve unlimited scaling in backend systems, analyzes the problem of excessive RPC‑to‑DB connections, and proposes a unitization approach that limits each service to a single database to enable true horizontal expansion.
This article analyzes the typical evolution from monolithic to RPC-based services, explains why sharding and database partitioning cannot alone achieve unlimited scaling due to connection limits, and introduces a unitization approach that assigns each application to a specific database shard to overcome those constraints.
This article traces the evolution of backend services from monolithic applications through RPC and sharding, explains why sharding and database partitioning cannot alone achieve unlimited scaling due to connection limits, and proposes a unitization approach that isolates database connections per service to enable true horizontal expansion.
This article analyzes the massive traffic handling of Alipay during Double 11, explaining the LDC (Logical Data Center) unit‑based design, the RZone‑GZone‑CZone hierarchy, traffic steering, disaster‑recovery mechanisms, and how OceanBase and Paxos enable CAP compliance for ultra‑high‑availability payments.
This article explains how Ant Financial’s logical data center (LDC) architecture, based on user‑sharded RZones, GZones, and CZones, combined with OceanBase’s Paxos‑based consensus, enables massive horizontal scaling, high availability, and disaster‑tolerant processing of over half a million payment transactions per second during Double‑11.
The article explains how Ant Financial’s Logical Data Center (LDC) and unit‑based architecture, combined with sharding, CAP analysis, and OceanBase’s Paxos‑based consensus, enable the payment platform to sustain tens of millions of transactions per second during Double‑11 events while ensuring high availability and disaster recovery.
The article examines why traditional sharding and database partitioning cannot alone achieve limitless scaling, explains the problem of excessive RPC‑to‑DB connections, and proposes a unitization approach that limits each service to a single database shard to enable true horizontal expansion.
The article explains how Alipay achieves massive payment throughput during Double‑11 by using logical data centers (LDC), unit‑based system design, multi‑active disaster‑recovery, and CAP‑theorem analysis, highlighting the role of OceanBase and PAXOS in ensuring consistency and availability.
This article explains how Ant Financial’s massive Double‑11 payment traffic is handled through logical data centers (LDC), unit‑based architecture (RZone, GZone, CZone), traffic routing, disaster‑recovery strategies, and a CAP analysis that highlights the role of OceanBase’s Paxos‑based consensus in achieving high availability and eventual consistency.
The article explains how Alipay’s Logical Data Center (LDC) architecture, with its RZone, GZone, and CZone unitization, combined with OceanBase’s Paxos‑based consensus, enables massive TPS growth, traffic diversion, and disaster‑recovery while navigating the CAP theorem constraints.
The article explains how traditional monolithic, RPC, and sharding architectures eventually hit MySQL connection limits, why simple database partitioning cannot guarantee unlimited scaling, and proposes a unitization approach that restricts each service to a single database shard to overcome connection‑count bottlenecks.
This article details Gaode's practical experience in building a unit‑based service architecture, covering challenges like request routing, unit isolation, and data synchronization, and explains the design choices, deployment strategies, performance metrics, and future optimization plans.
The article explains why traditional sharding and database partitioning cannot guarantee limitless scalability, examines the connection‑limit problems of RPC‑based services, and introduces unitization as a practical approach to achieve true unlimited scaling while managing database connections efficiently.
The article traces the evolution of backend services from monolithic applications through RPC-based scaling and database sharding, explains why sharding alone cannot achieve unlimited scalability due to connection limits, and proposes a unitized architecture that partitions databases per service to overcome these constraints.
