How a Four‑Layer Cloud‑Native Architecture Powers Scalable Web, Mobile, and IoT Services
This article details a four‑layer cloud‑native system architecture that separates presentation, communication, service, and data layers, leveraging Spring Cloud microservices, Alibaba Cloud networking, multiple databases, JWT authentication, ELK logging, and HTTP DNS to achieve scalability, reliability, and optimal user experience across web, app, and IoT platforms.
System Architecture Diagram:
Four‑Layer Architecture Design
Presentation Layer
Web front‑end built with HTML5, Vue, CSS3, compatible with major browsers. The presentation layer is fully decoupled from the data layer and communicates with the backend via cross‑origin requests.
APP: Native Android and iOS apps with HTTPS link optimization, anti‑hotlinking and DNS hijack handling.
WeChat Official Account / Mini‑Program: Certain business data are displayed via H5 on the official account; modules that control hardware devices are exposed through a WeChat mini‑program to improve user experience.
RESTful API: Business‑specific services are exposed through standard RESTful interfaces.
Communication Layer
Static content acceleration via Alibaba Cloud CDN.
Server load balancing using Alibaba Cloud SLB.
Data exchange over TCP, HTTP, HTTPS; TCP is implemented with Netty.
Service Layer
Core business functions are implemented as microservices on Spring Cloud.
Spring Cloud provides configuration management, service discovery, circuit breaking, intelligent routing, distributed session, and cluster state management for JVM‑based cloud applications.
Key components include:
Netflix Eureka : Service registry and discovery.
Netflix Hystrix : Circuit breaker for fault tolerance.
Netflix Zuul : Edge service for dynamic routing, monitoring, and security.
Netflix Archaius : Dynamic configuration API.
Spring Cloud Config : Centralized configuration server supporting local, Git, and SVN backends.
Spring Cloud Bus : Event/message bus for propagating state changes across the cluster.
Spring Cloud Sleuth : Distributed tracing with Zipkin/HTrace integration.
Spring Cloud Task : Management of short‑lived microservice tasks and scheduled jobs.
Data Layer
MongoDB – stores unstructured, weakly related data such as device sensor readings.
MySQL – stores transactional and strongly related data like orders and financial records.
HDSF – stores images, videos, and report files uploaded by monitoring devices.
Elasticsearch – part of the ELK stack for log storage and analysis.
Additional Components
Authentication System
Dual‑token JWT mechanism: accessToken for user authentication, refreshToken for obtaining a new accessToken after expiration.
Login flow diagrams illustrate token validation, token refresh, and complete token expiry scenarios.
Log System
Centralized log management using the ELK solution (Elasticsearch, Logstash, Kibana).
Session Management
Custom Netty session handling for monitoring, timeout, and reconnection.
DNS Hijack Mitigation
Addresses issues caused by LocalDNS caching and NAT forwarding by adopting HTTP DNS, which queries an HTTP web server directly, providing accurate IP resolution, better traffic steering, and extensibility.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
ITFLY8 Architecture Home
ITFLY8 Architecture Home - focused on architecture knowledge sharing and exchange, covering project management and product design. Includes large-scale distributed website architecture (high performance, high availability, caching, message queues...), design patterns, architecture patterns, big data, project management (SCRUM, PMP, Prince2), product design, and more.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.