0 views collected around this technical thread.
This article examines the challenges of high‑concurrency flash‑sale scenarios in multi‑active architectures, analyzes fairness issues caused by geographic latency, and presents practical solutions such as data‑sharding and global‑clock ordered queues, complemented by a Redis‑based implementation example.
The article examines the write‑after‑read consistency problem in multi‑active cross‑region systems, compares single‑write‑single‑read routing, quorum‑based multi‑write‑multi‑read, and read‑while‑copy methods, explains why primary‑secondary replication is preferred, and proposes a four‑step framework—scenario flagging, data marking, latency evaluation, and near‑by asynchronous replication—to meet WAR requirements efficiently.
This article explains the motivations, application scenarios, architectural patterns (same‑city, cross‑city, and cross‑country), and concrete design techniques for building multi‑active cross‑region systems that ensure high availability and graceful degradation during extreme failures.
To achieve disaster-recovery and cross-data-center resilience, Didi progressed from a simple proxy double-write scheme to a sophisticated MQ-free multi-active Redis design that uses a dedicated syncer, shard-based loop prevention, op-id replay protection, conflict detection, and incremental AOF durability, ensuring low latency, no data loss, and consistent availability.
Bilibili guarantees 99.99% availability for its live‑gift ecosystem by pre‑loading panel data, using circuit‑breakers and cached responses, sharding databases, auto‑scaling Kubernetes pods, employing Redis hot‑key caching and distributed rate limiting, handling order timeouts and idempotent MQ processing, and deploying multi‑active cross‑region services with intelligent discovery.
The article explains various high‑availability solutions for stateful backend services, comparing cold backup, dual‑machine active/standby, same‑city active‑active, cross‑city active‑active, and cross‑city multi‑active approaches, and discusses their trade‑offs, implementation details, and real‑world examples from large internet companies.
This article chronicles Bilibili's five‑year evolution of Site Reliability Engineering, detailing the introduction of SRE culture, the construction of high‑availability and multi‑active architectures, capacity management with Kubernetes, VPA/HPA, incident case studies, and the ongoing transformation of SRE practices across the organization.
The article visually details Bilibili’s multi‑active database architecture, showing system layout, data replication, failover strategies, and operational workflows that enable high availability across multiple data centers by synchronizing data, balancing load, and maintaining continuous service during node or regional failures.
This article explores the evolution of high‑availability strategies for stateful backend services, comparing cold backups, active/standby, same‑city and cross‑city active‑active setups, and discusses the trade‑offs, design considerations, and real‑world implementations of multi‑active and multi‑active architectures.
This article explains TiDB's multi‑active high‑availability architectures—including same‑city dual‑center, triple‑center, and two‑region three‑center deployments—detailing hard requirements, RPO/RTO goals, placement‑rule configurations, and practical disaster‑recovery recommendations for distributed database clusters, and how adaptive sync modes affect failover performance.
This article explains what geo‑active multi‑active (异地多活) architecture is, why it is needed for ultra‑high availability, and walks through the step‑by‑step evolution from a single‑node system to sophisticated multi‑data‑center designs that use redundancy, disaster‑recovery, data synchronization, routing, and conflict‑resolution techniques.
Covering essential high‑availability techniques, the article examines disaster‑recovery architectures from same‑city dual‑center to cross‑country active‑passive deployments, compares five patterns, details three multi‑active models, outlines required traffic‑scheduling, replication, and database layers, and provides design methodology, practical safeguards, and key HA metrics.
This article examines stateful backend services and various high‑availability strategies—including cold backup, active/standby hot backup, same‑city active‑active, cross‑region active‑active, and multi‑active architectures—detailing their advantages, limitations, and practical implementation considerations, and includes real‑world examples from major e‑commerce platforms.
This article explains the evolution of high‑availability deployment architectures—from simple cold backups and hot standby to same‑city active‑active, cross‑city active‑active, and finally multi‑active solutions—detailing their advantages, drawbacks, and practical design considerations for large‑scale internet services.
This article explains how Ele.me built a multi‑active, geographically distributed system that enables elastic scaling and data‑center‑level disaster recovery by partitioning services, routing traffic, replicating data in real time, and enforcing strict consistency and availability principles.
Multi‑active high‑availability architectures—ranging from same‑city dual‑active and two‑site three‑center setups to fully remote multi‑active deployments—provide continuous 24/7 service by replicating data across sites, but introduce latency, consistency, routing, and cost complexities that require careful unit‑based design, synchronized storage, and sophisticated traffic management.
This article describes JDHBase’s large‑scale KV storage architecture, its HBase‑based asynchronous replication mechanism, multi‑active cluster design, client‑side routing via Fox Manager, automatic failover strategies, dynamic replication tuning, and serial replication techniques to ensure data consistency across geographically distributed data centers.
This article describes JDHBase’s large‑scale KV storage deployment, its HBase‑based asynchronous replication mechanism, the multi‑active architecture with active‑standby clusters, client interaction via Fox Manager, automatic failover strategies, dynamic replication tuning, and serial replication techniques to ensure data consistency across data centers.
Wang Xiaobo explains how public‑cloud services can simplify designing and implementing active‑active architectures, covering data‑center redundancy, real‑time synchronization, fault‑tolerant networking, micro‑service migration, and cost‑benefit trade‑offs, while urging incremental, cloud‑assisted approaches rather than full 100% multi‑active deployments.
The article describes how a DBA team at Ele.me transformed manual database management into a fully automated, platform‑driven operation covering monitoring, alarm handling, MHA automation, resource pooling, large‑scale migration, SQL review, and multi‑active DDL release using Go‑based tools and custom workflows.