0 views collected around this technical thread.
This article explains the materialized view pattern, describing why pre‑populated views are created when source data formats hinder efficient queries, how they improve performance as disposable caches, and the key considerations, trade‑offs, and scenarios for adopting this approach.
This article explains MySQL dual‑master architecture, covering dual‑master replication, master‑slave replication, master‑master with backup nodes, and ring replication, while discussing their advantages, common pitfalls such as loop replication and data inconsistency, and practical solutions to ensure high availability and data integrity.
This article explains the role of consensus algorithms in achieving high‑availability through redundancy and automatic failover, clarifies distributed consistency, explores the CAP theorem and its C component, and introduces the BASE theory as a practical complement for eventual consistency in modern distributed systems.
This article examines the unreliability of physical clocks in distributed systems, compares synchronous and asynchronous network timing, explains the roles of wall and monotonic clocks, and explores logical clocks, snapshot isolation, and practical solutions such as Google Spanner's TrueTime to ensure data consistency.
This article explains RocketMQ's transactional message mechanism, covering half‑message storage, three transaction states, status‑check procedures, key APIs, storage reliability, and the two‑phase commit process that guarantees eventual consistency in distributed systems.
The article analyzes data‑consistency challenges caused by RPC timeouts, especially when interfaces lack idempotency or idempotency fails, and presents a lightweight mock/spy utility that can intercept, mock, or spy on service calls to quickly restore consistency in distributed systems.
The article presents a lightweight tool to handle data consistency in RPC services during network timeouts by mocking or spying on service calls, ensuring data integrity without actual execution.
This article presents a comprehensive technical case study of a three‑layer product‑tagging platform that addresses the challenges of fine‑grained operations, ensures real‑time tag updates, guarantees data consistency, and eliminates read bottlenecks through traffic separation, event‑driven processing, deduplication MQ, and multi‑level caching.
The article explains how microservice architectures face data consistency challenges, introduces the CAP theorem's trade‑offs among consistency, availability and partition tolerance, and discusses practical solutions such as service registries, distributed transaction patterns, and cloud‑native strategies for maintaining reliable systems.
The article outlines Bilibili’s revamped recommendation data‑flow architecture that eliminates timing and calculation inconsistencies by snapshotting online features, unifying feature computation in a single C++ library accessed via JNI, and orchestrating label‑join and sample extraction through near‑line Kafka/Flink pipelines, with further performance gains and Iceberg‑based future extensions.
Xiaohongshu built a lock‑free, non‑disruptive data‑consistency verification tool that automatically selects optimal methods, handles heterogeneous sources and dynamic changes, performs full and incremental checks via chunked checksums or row‑by‑row comparison, quickly isolates mismatches, and supports automatic remediation, ensuring reliable migrations and sharding.
The paper proposes a multi‑level storage architecture for Bilibili’s comment service that replaces TiDB with a custom KV store (Taishan) and Redis caching, introduces unstructured indexes, CAS‑based consistency, real‑time and offline reconciliation, and a hedged degradation strategy to boost reliability, read throughput, and scalability during traffic spikes.
This article explains common multithreading problems such as race conditions, deadlocks, resource contention, and data consistency issues, and provides Python code examples that demonstrate synchronization techniques, lock ordering, connection pooling, thread pools, and transaction management to ensure correct and stable concurrent execution.
This article details Ctrip's data middle‑platform solution for international ticketing, covering background challenges, design principles, key technical practices such as version control, P2P distribution, data timeliness, robustness, consumption‑process optimization, overall architecture, achieved benefits, and future plans.
This article explains the four common Redis cache pitfalls—cache penetration, avalanche, breakdown, and data inconsistency—demonstrates how they can crash high‑traffic systems, and provides practical Java/Spring Boot solutions such as empty‑object caching, Bloom filters, distributed locks, and delayed double‑delete strategies.
This article presents two practical database migration schemes—dual‑write and a gray‑scale switch—detailing their workflows, code refactoring, data synchronization using Syncer, consistency verification with sync‑diff‑inspector, and operational trade‑offs for minimizing downtime and ensuring data integrity.
This article examines the challenges faced by Ctrip’s ticket reservation transaction system during flash‑sale events and details the architectural optimizations—including Redis caching, database load reduction, supplier integration, and multi‑layer traffic throttling—that ensure system stability, strong consistency, and high availability under extreme concurrency.
The Vivo Internet Storage R&D team's article describes an end‑to‑end MySQL data‑consistency verification tool for DTS that uses fixed‑size chunking and CRC32/MD5 fingerprint aggregation to quickly compare source and target tables, pinpoint mismatched rows, and enable automated or manual correction while minimizing impact on replication.
The article examines a cache‑consistency failure in Tmall International’s high‑concurrency push system, explains classic cache problems and mitigation techniques, analyzes the delete‑then‑update bug that caused null‑plan errors, and evaluates four corrective strategies ranging from double‑write to delayed double‑delete.
Cache Aside is a popular caching pattern that reads from cache first and writes through the database before invalidating the cache, but concurrent read‑write scenarios can cause stale data; the article explains these pitfalls and recommends the double‑delete strategy to keep cache and database consistent.