This article explains what Netty is, how to design a distributed service framework using it, details the RPC call flow, discusses protocol design, showcases performance‑boosting features and best practices, and provides deep insights into Netty's architecture, threading model, and low‑level optimizations.
Nebula Graph is a high‑performance, open‑source distributed graph database that separates storage and compute, supports billions of vertices, and offers millisecond query latency, while its cloud‑native DBaaS deployment faces multi‑cloud resource, performance, and operational challenges addressed through Kubernetes, KubeSphere, and custom operators.
This article explains what a big data platform is, outlines its seven‑component overall architecture, details the technical stack from data sources to applications, and describes the key subsystems such as catalog management, data integration, governance, storage, processing, sharing, development, and analysis.
The article explains how to detect and prevent message loss, guarantee reliable delivery, handle duplicate consumption, manage message backlog, and implement distributed transaction messages in systems like Kafka, RocketMQ, and RabbitMQ, providing practical code examples and best‑practice guidelines.
This article provides a comprehensive overview of caching in large‑scale distributed systems, covering cache fundamentals, classification, major implementations such as CDN, reverse‑proxy, local, and distributed caches, detailed analyses of Memcached and Redis, common design challenges, and real‑world industry solutions.
This article explains the challenges of distributed consistency in decentralized environments and provides an in‑depth overview of classic consensus algorithms such as PBFT, Proof‑of‑Work, and Proof‑of‑Stake, including their mechanisms, limitations, and suitable application scenarios.
The article explains three frequent pitfalls when adopting microservices—complex remote communication, asynchronous processing challenges, and distributed transaction difficulties—and shows how fast‑fail, retries, timeouts, compensation, lightweight workflow engines, and idempotency can reduce complexity and improve resilience.
The article compares four consistent‑hashing algorithms—Ketama’s ring with virtual nodes, Rendezvous’s highest‑random‑weight method, Google’s Jump Consistent Hash, and Maglev’s lookup‑table approach—evaluating their balance, monotonicity, stability, scalability, and time complexity, and concludes that Ketama and Jump offer the best overall trade‑off.
This article explains the core mechanisms of the Raft consensus algorithm—including leader election, log replication, safety guarantees, message structures, state transitions, and key Go implementations in etcd-raft—providing code examples and detailed analysis of functions such as becomeLeader, tickElection, and appendEntry.
This article explains the design, architecture, and implementation of Transparent Multilevel Cache (TMC), a solution that adds application‑level hotspot detection and local caching to reduce cache hot‑spot pressure, improve consistency, and boost performance for Java services in high‑traffic scenarios.
This article explains the fundamental message exchange architectures—such as publish‑subscribe, fan‑out, unidirectional and bidirectional streaming, unicast, broadcast, multicast, and anycast—and how they are used in systems like Redis, Kafka, RabbitMQ, ZeroMQ and IBM MQ to simplify communication between architects and developers.
This article explains the evolution of message‑queue middleware by comparing the basic double‑ended queue implementation, Redis list usage, Kafka’s partitioned log architecture with segment files and sparse indexes, and Pulsar’s compute‑storage separation using BookKeeper, highlighting their designs, strengths, and trade‑offs.
This comprehensive guide explains why distributed locks are needed, walks through basic Redis lock implementations, reveals common dead‑lock and expiration issues, presents atomic solutions with SET EX NX and Lua scripts, evaluates the Redlock algorithm, examines expert debates, compares Zookeeper locks, and offers practical recommendations for safe usage.
The Apache SkyWalking Day offline meetup, co‑hosted by Tencent Cloud Tengyuanhui and the SkyWalking community on June 26 in Beijing, will feature industry experts presenting deep technical insights on distributed system observability and monitoring, while attendees can learn the latest ecosystem developments, performance‑management best practices, and enter giveaways such as AirPods Pro, Tencent figurines, and stickers.
This article compares popular message middleware such as Redis, Kafka, and Pulsar, explaining their underlying data structures, strengths and weaknesses, and how concepts like partitions, replication, cursors, and storage segmentation enable high performance, scalability, and reliability in modern distributed messaging systems.
This article explores the fundamental concepts of system architecture, scaling challenges, multi‑level caching, service decomposition, and monitoring techniques in massive web platforms, using Weibo’s evolution from LAMP to service‑oriented architecture as a detailed example.
The article presents the design and implementation of Transparent Multilevel Cache (TMC), a three‑layer caching solution that adds hotspot detection and local cache to reduce distributed cache pressure, explains its transparent Java integration, describes the sliding‑window hotspot discovery pipeline, and showcases performance gains in real‑world e‑commerce campaigns.
WE‑CAN, NetEase Cloud Communication's self‑developed large‑scale distributed transmission network, dramatically improves end‑to‑end data quality, reduces costs, and supports diverse real‑time communication scenarios through intelligent routing, layered decoupling, and global edge deployment, as demonstrated by extensive production metrics.
This article explains distributed consensus, compares Paxos, ZAB and Raft, details Raft's leader election, log replication and commit mechanics, introduces read‑optimisation techniques such as ReadIndex and Lease Read, and showcases the Java‑based SOFAJRaft library with its architecture, performance tricks and real‑world use cases.
The article presents a detailed overview of Tencent Advertising's real‑time strategy data framework, explaining its role in the ad system, the challenges of massive log volumes, and the architectural, performance, and high‑availability solutions implemented to achieve fast, reliable, and scalable ad decision making.
This article describes Dada Group’s evolution from a simple 1:1 test environment to a sophisticated machine‑labeling load‑testing solution, detailing core design, isolation techniques, custom testing platform, model construction, pre‑heat strategies, and post‑test analysis that ensure system stability during high‑traffic events.
The article introduces Transparent Multilevel Cache (TMC), a comprehensive caching solution that adds hotspot detection and local caching to existing distributed cache systems, explains its three‑layer architecture, transparent Java integration, real‑time hotspot discovery process, and demonstrates performance gains in high‑traffic e‑commerce scenarios.
The article introduces Youzan's Transparent Multilevel Cache (TMC), explains why hotspot cache access harms e‑commerce applications, describes its three‑layer architecture, details the Java client integration with Hermes‑SDK for automatic hotspot detection and local caching, and presents real‑world performance gains during large‑scale promotional events.
This article explains Redis Cluster's decentralized sharding design, the role of master‑slave replication for high availability, and how strong consistency is achieved using consensus protocols such as Raft, highlighting key concepts like global log indexes and commit pointers.
Gaode Ride‑Hailing created a comprehensive 360° observability platform—standardized logging, distributed tracing, multi‑domain metrics, visual dashboards, and an incident workflow—that transforms raw data into actionable insights, accelerates root‑cause analysis, and enables automated fault defense for its large‑scale cloud‑native microservice system.
This article explains the fundamentals of distributed system consistency, covering weak and strong consistency, the CAP theorem, ACID vs BASE, and detailed overviews of protocols such as 2PC, 3PC, Paxos, Raft, Gossip, NWR, Quorum, and Lease mechanisms.
This article examines the challenges of message loss in live‑streamed online classrooms, analyzes why pushes can fail, and proposes a comprehensive reliability strategy—including TCP fundamentals, multi‑channel redundancy, sequence‑based ordering, hole‑pulling, and configurable back‑end mechanisms—to achieve near‑100% delivery.
The article defines cloud‑native data warehouses as storage‑compute separated systems that elastically scale across clouds, outlines their key traits, describes a three‑layer architecture, compares Snowflake and OushuDB implementations, and illustrates a large bank’s migration to such a platform.
This article explains how to build a highly available, high‑performance, easily extensible and secure large‑scale website architecture by evolving from a single‑server setup through application‑data separation, caching, server clustering, read‑write separation, CDN/reverse‑proxy, distributed storage, NoSQL, service decomposition, and various architectural patterns.
This article explores the definition of software architecture, reviews essential technologies such as distributed systems, clustering, caching, queues, multithreading, and rate limiting, discusses security threats like SQL injection and XSS, and highlights common design mistakes to avoid.
This article explains RabbitMQ's core messaging patterns—including simple, work queues, publish/subscribe, routing, topics, RPC, and publisher confirms—along with the four main exchange types, providing practical use‑case examples for each to help developers choose the appropriate pattern for their backend systems.
This article explains the role of a service registry in micro‑service architectures, compares Zookeeper (CP) and Eureka (AP) in terms of design, consistency, and availability, and provides detailed code and configuration examples for deploying both solutions in a high‑availability cluster.
This article reviews the seven‑year evolution of Ctrip's microservice service‑registry from manual data maintenance through an etcd‑based solution to the self‑developed Artemis system, detailing its architecture, consistent‑hash data partitioning, high‑availability design, and second‑level instance up/down mechanisms.
The article explains Youzan's Transparent Multilevel Cache (TMC), a solution that automatically detects cache hotspots, adds an application‑level local cache, and provides hit statistics to reduce load on distributed caches, improve consistency, and boost performance for high‑traffic e‑commerce scenarios.
The article explains how Kafka 2.8 removes its dependency on Zookeeper, detailing Kafka's core concepts, the role of Zookeeper in broker registration, load balancing, and controller election, the operational drawbacks of this coupling, and how KIP‑500 with a Quorum Controller modernizes the architecture.
This article explains the evolution of software architecture toward microservices, defines microservice architecture, outlines its key characteristics, advantages and disadvantages, and then details ten essential design patterns—including database per service, event sourcing, CQRS, Saga, BFF, API gateway, Strangler, circuit breaker, externalized configuration, and consumer‑driven contract testing—while advising when each should be applied.
The article explains Youzan’s Transparent Multilevel Cache (TMC), detailing its architecture, hotspot detection, local caching, consistency mechanisms, and real‑world performance gains during flash‑sale events, showing how it reduces cache pressure and improves latency for Java‑based services.
This article explains Kafka's core concepts, the challenges of consumer failures in microservice architectures, why naïve retry loops or message skipping are insufficient, and presents a nuanced approach that distinguishes recoverable from unrecoverable errors, using back‑off retries and hidden topics to preserve ordering and data integrity.
This article explains high availability in distributed systems, covering its definition, design objectives, key metrics such as MTBF, MTTR, SLA, and practical design elements like redundancy, monitoring, failover, as well as common Q&A on cost, relationship with other architecture attributes, and implementation considerations.
This article explains how to build a highly available, high‑performance, easily extensible, scalable and secure web application system by describing the evolution of large‑site architectures, common patterns such as layering, clustering, caching, asynchronous processing, and the core architectural factors of performance, availability, scalability, extensibility and security.
This article visually breaks down Kafka’s core concepts—including producers, consumers, topics, partitions, consumer groups, and cluster architecture—explaining how messages flow, are stored, and replicated across partitions and nodes, while highlighting the role of ZooKeeper in managing metadata.
This article outlines the planning, design, and Java implementation of a distributed load‑testing framework called FunTester, focusing on two test‑case schemes (case transmission and execution), an HttpRequestBase‑based demo, case creation, transmission, execution, and the complete FunRequest class code.
This article breaks down the design and source‑code implementation of RocketMQ's NameServer, detailing its core responsibilities, component interactions, and the seven Java classes that together provide network handling, connection management, routing, configuration, and startup logic.
This article compares Redis list, Kafka, and Pulsar as messaging middleware, explaining their architectures, strengths, and weaknesses—including queue fundamentals, partitioning, cursor management, consumer groups, high‑availability mechanisms, storage strategies, and consumption models—to help readers choose the right solution for large‑scale systems.
The guide details how to build a distributed message‑governance platform for the Hello mobility service, covering unified SDK design, RocketMQ pitfalls, client and cluster health monitoring, risk mitigation, and a tiered microservice high‑availability architecture that uses circuit‑breaking, rate‑limiting, and pre‑heating to ensure resilient traffic handling.
This article explains the evolution of message‑queue middleware by comparing basic double‑ended queue implementations, Redis list usage, Kafka’s partitioned log design with cursor‑based consumption and high‑availability replication, and Pulsar’s compute‑storage separation using BookKeeper, while highlighting their strengths, limitations, and practical trade‑offs.
This article explains how Kafka 2.8 removes its dependency on Zookeeper, describes Kafka's core concepts and its interaction with Zookeeper, outlines the role of the Controller, discusses operational complexities and upgrade paths with KIP‑500, and highlights the benefits of the new KRaft‑based architecture.
This article outlines the evolution of software architecture—from monolithic to distributed, microservice, and serverless models—detailing each pattern’s structure, advantages, and drawbacks, helping developers understand how architectural choices impact scalability, maintenance, and innovation in modern software projects.
This article uses Taobao's backend as a case study to illustrate how a system evolves from a single‑server setup to a highly available, horizontally scalable architecture through stages such as separating Tomcat and database, adding caching, introducing load balancers, read/write splitting, sharding, microservices, ESB, containerization, and finally cloud‑native deployment, while summarizing key design principles.
Facing rising traffic, JD Daojia’s in‑house load‑testing platform was redesigned to automate script management, enable distributed JMeter execution, integrate real‑time monitoring, and support custom RPC protocols, dramatically lowering manual effort, cutting test cycles from an hour to fifteen minutes while ensuring system stability.
This article provides a comprehensive technical overview of large‑scale distributed website architecture, covering characteristics, goals, common patterns, high‑performance and high‑availability designs, scalability, extensibility, security, agility, a seven‑layer reference model, and the evolutionary steps of modern e‑commerce systems.
This article provides a comprehensive technical overview of large‑scale distributed website architecture, covering characteristics, goals, common patterns, high‑performance and high‑availability designs, scalability, extensibility, security, agility, a seven‑layer reference model, and a detailed e‑commerce case study with practical optimization steps.
Kafka 2.8.0 introduces a major improvement by implementing the Raft consensus algorithm, allowing the platform to operate without ZooKeeper, which reduces system complexity and improves efficiency, though the ZooKeeper‑less mode is still early‑stage and not recommended for production environments.
This article explains how JD constructs its OLAP platform from data ingestion to storage, querying, and management, describing the diverse data sources, real‑time and offline processing, scalability, consistency, fault tolerance, and future optimization plans, while addressing key technical challenges and solutions.
This article provides a comprehensive technical overview of large-scale distributed website architecture, covering characteristics, design goals, layered patterns, performance and high‑availability optimizations, scalability, extensibility, security measures, agile practices, and the evolutionary steps of e‑commerce systems.
The article presents a comprehensive overview of large‑scale distributed website architecture, detailing its characteristics, performance and availability goals, layered design patterns, high‑performance and high‑availability techniques, scalability, extensibility, security, and practical e‑commerce case studies.
This article provides a comprehensive technical overview of large‑scale distributed website architecture, covering characteristics, design goals, high‑performance, high‑availability, scalability, security, and detailed patterns such as layered design, caching, clustering, load balancing, database sharding, CDN, and service‑oriented decomposition.
The article explains why large internet firms adopt cross‑region multi‑active architectures for high availability, compares cold backup, hot standby, same‑city active‑active, and cross‑region active‑active solutions, discusses their trade‑offs, and presents practical design patterns and questions for implementing such systems.
The inaugural Techo TVP Developer Summit in Shenzhen gathered over 500 developers to explore the latest trends in databases, distributed systems, big data, and cloud‑native technologies, offering expert analyses, real‑world case studies, and career guidance for data professionals.
The article thoroughly examines Tars’s Java RPC client, explaining its initialization, proxy generation via JDK dynamic proxies, diverse load‑balancing strategies, NIO‑based Reactor networking, and both synchronous (CountDownLatch) and asynchronous callback mechanisms, supplemented with detailed source‑code illustrations of each component.
This article compares the fundamentals and limitations of Redis list‑based queues with Kafka and Pulsar, explaining how partitions, cursors, replication, segment storage, and consumption models overcome Redis’s drawbacks and enable high‑performance, scalable, and highly available messaging systems.
This article examines Ctrip’s Hermes messaging system, tracing its evolution from a simple Mongo‑based queue to a broker‑centric, MySQL/Kafka hybrid architecture, and explains the design choices, performance optimizations, cluster management via lease‑based meta‑server, and lessons learned for building high‑throughput, low‑latency MQ solutions.
This article provides a visual walkthrough of Kafka's fundamental concepts—including producers, consumers, topics, partitions, and cluster architecture—illustrated with diagrams to help readers clearly understand how messages flow and are stored in a distributed streaming system.
This article uses Taobao as a case study to trace the evolution of its server‑side architecture from a single‑machine setup to a cloud‑native, micro‑service ecosystem, detailing each technical milestone, the challenges addressed, and the core design principles for high‑concurrency systems.
This article provides a comprehensive overview of distributed system fundamentals, covering node models, replica concepts, various consistency levels, data distribution strategies, lease-based caching, quorum mechanisms, two‑phase commit, MVCC, Paxos consensus, and the CAP theorem, illustrating each with practical examples and diagrams.
This comprehensive guide explains Elasticsearch’s fundamentals, including its distributed architecture, indexing process, shard and replica mechanisms, query execution, near‑real‑time search, segment management, and practical optimization techniques, providing developers and engineers with the knowledge needed to design, operate, and troubleshoot large‑scale search clusters.
This article explains how consistent hashing, combined with Docker-based microservice architectures, reduces data movement during dynamic scaling by using a ring‑shaped hash space and intelligent node placement, making auto‑scaling more efficient and resilient.
Apache Kafka 2.8.0 introduces several significant enhancements, including a new group API, mutual TLS authentication for SASL_SSL listeners, JSON request/response logging, broker connection rate limiting, topic identifiers, self‑managed quorum replacing ZooKeeper, and numerous improvements to Streams and Connect APIs for more reliable real‑time data pipelines.
This article explains how to implement a true multi‑region active‑active architecture by enabling bidirectional data synchronization among three or more data centers, covering CAP trade‑offs, distributed ID generation algorithms, center closure strategies, final consistency mechanisms, and a disaster‑recovery design.
This article explains Apache Pulsar’s message storage architecture in BookKeeper, details the ledger and entry lifecycle, describes the multi‑layer caching read path, and outlines BookKeeper’s garbage‑collection process along with practical operational tips for avoiding disk‑heavy scenarios.
This article details the design and practical implementation of an AI-driven masked danmaku system that prevents comment overlay on video content, covering background, technology selection, instance segmentation methods, distributed task scheduling, mask generation, client rendering, performance optimizations, and future directions.
This article explains the fundamentals of consistent hashing, compares the classic ring‑based method with the more efficient jump consistent hash algorithm, provides reference implementations in C++, discusses their time‑complexity and practical trade‑offs, and shows how they are applied in systems such as Greenplum.
This comprehensive guide covers the core concepts of distributed systems—including CAP theory, consistency models, transaction protocols like 2PC/3PC and TCC, Redis features, persistence options, expiration strategies, memory eviction policies, distributed locking with Redis and Zookeeper, as well as an overview of Kafka's architecture and reliability mechanisms.
This comprehensive guide covers core distributed system theories, CAP consistency, transaction protocols, Redis features and persistence, Zookeeper coordination, message queue fundamentals, Kafka performance tricks, and practical solutions for caching, locking, and high‑concurrency scenarios.
This article explains microservice architecture, its key characteristics, advantages and disadvantages, and presents ten essential design patterns—including database per service, event sourcing, CQRS, Saga, BFF, API gateway, Strangler, circuit breaker, externalized configuration, and consumer‑driven contract testing—while offering guidance on when to apply each pattern.
This article outlines what middleware development is, the essential technical and personal qualities required, and provides a step‑by‑step learning roadmap for aspiring Java developers aiming to join or build middleware teams in modern backend environments.
This comprehensive guide explains microservice architecture, its key characteristics, advantages, disadvantages, and when to adopt it, then details ten crucial design patterns—including database per service, event sourcing, CQRS, Saga, BFF, API gateway, Strangler, circuit breaker, externalized configuration, and consumer‑driven contract testing—complete with pros, cons, use‑cases, and technology examples.
The Apache Pulsar Meetup in Shenzhen on April 17, 2021, co‑hosted by Tencent Middleware and StreamNative, will showcase Pulsar’s cloud‑native messaging, streaming, and storage capabilities through sessions on KoP migration, big‑data and IoT use cases, cloud‑native deployments, and the StreamNative Cloud Pulsar‑as‑a‑Service offering.
This article compares common globally unique ID generation methods—including database auto‑increment, batch services, UUID/GUID, timestamp‑based IDs, Redis INCR, and Twitter’s Snowflake—detailing their advantages, drawbacks, and improvement options to help developers select the most suitable approach for scalability and ordering needs.
Software developers need to understand the evolution of software architectures—from simple monolithic designs to distributed applications, microservices, and serverless platforms—because each model has distinct advantages and drawbacks that affect scalability, maintainability, deployment frequency, operational complexity, and overall cost.
This article provides a comprehensive, diagram‑driven explanation of Kafka’s architecture, covering producers, consumers, topics, partitions, replication, Zookeeper coordination, controller election, state machines, and the NIO‑based network model, helping readers understand the design philosophy and practical configuration implications.
This article provides a comprehensive tutorial on Zookeeper 3.5.10, covering client and server commands, node attributes, creation, querying, updating, deletion, listing children, checking node status, watcher usage, and detailed access control mechanisms with practical command examples.
Apache Kafka 2.8 replaces ZooKeeper with an internal Quorum controller (KRaft), enabling ZooKeeper‑free operation, reducing resource usage, improving performance, and supporting larger clusters while noting that early versions lack some security and management features and are not yet production‑ready.
This article introduces NSQ, an open‑source Go‑based distributed message queue, detailing its core components, scalable high‑availability deployment topology, backup and replay mechanisms, and monitoring practices to ensure reliable, fault‑tolerant message processing.
This article provides a detailed side‑by‑side analysis of four major distributed message‑queue systems—Kafka, RabbitMQ, RocketMQ, and ActiveMQ—examining them across seventeen criteria such as documentation, language support, protocols, storage, transactions, load balancing, clustering, management UI, availability, duplication, throughput, subscription models, ordering, acknowledgments, replay, retry, and concurrency.
This article explains how data islands hinder AI progress, introduces federated learning as a privacy‑preserving solution, details JD.com’s 9N‑FL platform—including its architecture, features, classification, privacy‑preserving techniques, and algorithm support—and demonstrates its successful application in e‑commerce advertising that yielded over 15% revenue growth.
Baidu’s Transaction Middle Platform Order System provides a unified, cloud‑native architecture that supports diverse Baidu products by orchestrating modular commands, handling order lifecycle, payment, refunds, and splitting, while employing caching, asynchronous processing, and sharded databases to achieve high throughput and low latency.
This article describes how to design and build a native distributed cluster called MyCluster without using any existing frameworks, covering master‑slave architecture, leader election, split‑brain handling, centralized configuration management, custom communication protocols, state transitions, and client interfaces.
The article explains the progression of software architecture from simple monolithic designs through distributed applications and microservices to modern serverless solutions, outlining each model's structure, advantages, disadvantages, and impact on development and operations for software engineers.
The article explores the fundamental concepts of software architecture, illustrating how massive platforms like Weibo handle millions of users through layered design, service decomposition, multi‑level caching, distributed tracing, and capacity planning to achieve high scalability and reliability.
Meituan evaluated open‑source distributed graph databases against strict latency, scale, and import criteria, selected NebulaGraph for its superior multi‑hop query and bulk‑load performance, and built a four‑layer, highly available platform that ingests petabyte‑scale data in real time, supports diverse business use cases, and provides interactive visualization.
This article provides a comprehensive overview of Zookeeper's cluster deployment, explains the four server states, details the leader election process—including initialization, voting, and decision logic—and presents key source code snippets to help developers understand and implement Zookeeper's high‑availability mechanisms.
This article explores why reliable inter‑service communication is essential in microservice architectures, illustrates common pitfalls with real‑world examples, and presents five practical retry and persistence solutions—including fast retry, in‑memory queues, persistent queues, retry services, and pre‑notification—to improve message delivery reliability.
The article outlines eBay’s practical scalability principles—including functional partitioning, horizontal sharding, avoiding distributed transactions, aggressive asynchronous decoupling, moving work to async pipelines, virtualization, and intelligent caching—demonstrating how large‑scale web systems can be designed for linear growth and high availability.
This article explores the fundamental principles of system architecture, illustrating how large‑scale services such as Uber and Weibo handle massive traffic through strategic abstraction, modularization, performance optimization, multi‑level caching, distributed tracing, and operational best practices to achieve scalability and reliability.
The article explains how Zeebe, a cloud‑native workflow engine, orchestrates microservices by separating task logic from coordination, detailing its BPMN‑based architecture (client, gateway, brokers, exporters), core features such as real‑time visibility, horizontal scalability, fault tolerance, message‑driven processing, and showcases Java job‑worker code and benchmark results demonstrating tens of thousands of workflow instances per second.
This article explains Youzan's Transparent Multilevel Cache (TMC), detailing its architecture, transparent Java integration, hotspot detection and local caching mechanisms, and demonstrates its real‑world performance gains during flash‑sale events and large‑scale marketing campaigns.
This article explains the challenges of Raft member changes, compares the two‑phase Joint Consensus approach with single‑step changes, discusses correctness and availability pitfalls, and offers practical engineering guidelines for implementing safe and efficient cluster reconfiguration.
This article details Baidu's Feed recommendation system architecture, explaining how a combination of dynamic retry scheduling, real‑time stop‑loss mechanisms, multi‑recall frameworks, ranking layer fallbacks, and IDC‑level multi‑master designs collectively ensure five‑nine availability across billions of daily requests.
This article provides an in‑depth technical overview of how Pinterest scales its distributed cache layer using Memcached and Mcrouter on AWS, covering architecture, performance, high availability, load balancing, trade‑offs, and future directions.
