InstaVerse

Led the end-to-end platform design from requirements gathering to deployment, ensuring enterprise-level scalability and performance.

• Requirements Analysis: Comprehensive business requirement gathering and technical feasibility assessment
• Technology Selection: Strategic technology stack selection based on performance and scalability requirements
• Architecture Design: Microservices architecture design ensuring high scalability, availability, and performance
• Project Setup: Complete project scaffolding and deployment pipeline configuration
• API & Database Design: RESTful API design and optimized database schema for massive user loads

Developed reusable components and utilities to improve code maintainability and development efficiency.

• User Context Component: Centralized user session and context management
• Logging Component: AOP-based interface logging with performance monitoring
• Jackson Configuration: Automated JSON serialization/deserialization configuration
• Global Exception Handler: Centralized exception handling with custom error responses
• Custom Validation: Parameter validation annotations for business logic
• Response Utilities: Standardized API response wrapper and utilities

Designed and implemented microservices architecture based on domain-driven design principles.

• Gateway Service: API routing, load balancing, and centralized authentication
• Authentication Service: User authentication, authorization, and session management
• User Service: User profile management with high-concurrency caching
• Blog Service: Content publishing, editing, and management with async processing
• Relation Service: Social relationship management (follow/unfollow)
• Counter Service: High-performance counting for likes, views, followers
• OSS Service: Multi-cloud object storage with strategy pattern
• ID Generator Service: Distributed ID generation using Leaf framework
• KV Storage Service: Short text key-value storage using Cassandra

Built and optimized the API Gateway as the single entry point for all client requests.

• Unified Routing: Centralized routing management for all microservices
• Load Balancing: Intelligent traffic distribution across service instances
• SaToken Integration: Unified authentication and authorization
• Global Filters: Token parsing and user ID propagation to downstream services
• Rate Limiting: API rate limiting and traffic control

Implemented comprehensive authentication system based on RBAC model with multiple login methods.

• RBAC Model: Role-based access control for fine-grained permissions
• JWT Authentication: Stateless authentication with token-based security
• Multi-Login Support: Username/password and SMS verification login
• User Registration: Self-registration with email/SMS verification
• SMS Integration: Alibaba Cloud SMS service with async processing
• Anti-Spam Protection: SMS interface rate limiting and abuse prevention

Integrated Meituan's Leaf framework to build high-performance distributed ID generation service.

• Dual Mode Support: Segment-based and Snowflake algorithm ID generation
• Feign Interface: RESTful API for seamless service integration
• High Performance: 22,000+ TPS with 4ms average response time
• Massive Scale: 1.9+ billion daily ID generation capacity
• Performance Testing: JMeter load testing for single-node optimization

Designed flexible object storage service supporting multiple cloud providers.

• Multi-Cloud Support: Alibaba Cloud OSS and MinIO integration
• Strategy Pattern: Factory + Strategy pattern for extensible storage backends
• Dynamic Configuration: Nacos-based dynamic storage strategy registration
• File Management: Upload, download, and metadata management
• Performance Optimization: Async processing and CDN integration

Built high-performance key-value storage service for short text content using Cassandra.

• Cassandra Integration: Apache Cassandra with RocksDB storage engine
• Short Text Storage: Optimized for blog content and comments
• High Throughput: Massive read/write operations support
• Data Modeling: Efficient data modeling for query patterns

Implemented comprehensive user management service with advanced caching strategies.

• User Management: Complete CRUD operations for user profiles
• Security: BCrypt password encryption with random salt
• L2 Caching: Redis + Caffeine local cache for high-concurrency reads
• Cache Protection: Prevention of cache avalanche, penetration, and breakdown
• Performance: Optimized for millions of concurrent user queries

Built comprehensive blog management system with advanced performance optimizations.

• Content Management: Publishing, editing, querying, pinning, and permission control
• Async Processing: CompletableFuture for concurrent downstream service calls
• L2 Caching: High-concurrency read support with cache invalidation
• Data Consistency: RocketMQ broadcast for distributed cache invalidation
• Like System: Redis Bloom filter (later Roaring Bitmap) for high-performance like operations
• Traffic Shaping: RateLimiter token bucket for database protection

Designed social relationship management system for follow/unfollow operations at scale.

• Follow/Unfollow: High-concurrency social operations with ordered messaging
• Redis Optimization: Lua scripts for atomic operations and performance
• Idempotency: Unique indexes for relationship record deduplication
• List Queries: Redis ZSET for follower/following list pagination
• Hot Data Caching: Different strategies for followers vs following lists

Built specialized counting service for likes, follows, and view statistics at massive scale.

• BufferTrigger: Traffic aggregation (1000 records/batch, 1s intervals)
• Redis First: Write to Redis cache before async database persistence
• Data Alignment: XXL-Job scheduled tasks for eventual consistency
• Shard Broadcasting: Parallel processing for large data volumes
• Real-time Updates: Immediate counter updates with background persistence

Implemented real-time search service with full-text capabilities and data synchronization.

• Index Design: Elasticsearch indexes for blogs and users
• Bulk Indexing: Logstash for full data indexing
• Real-time Sync: Canal CDC for MySQL binlog processing
• Search Features: Chinese tokenization, highlighting, filtering, custom scoring
• Data Consistency: Guaranteed consistency between MySQL and Elasticsearch

Built comprehensive comment system with advanced reliability and performance features.

• Data Design: Redundant fields for query optimization (earliest reply ID, sub-comment count)
• MQ Reliability: Spring Retry + fallback storage + async retry for guaranteed delivery
• Async Publishing: High-throughput comment publishing with async operations
• Batch Processing: Bulk MQ consumption and database writes
• Pagination: Efficient L1/L2 comment pagination with L2 caching
• Like System: Memory operation merging for high-performance like processing