A Practical Path to Enterprise Cloud Computing: Insights on Microservices and DevOps

The author, Gu Wei, a chief architect at PuYuan, recounts his experience over more than a decade working on cloud‑related projects such as OpenStack, Kubernetes, microservices, DevOps, CloudFoundry, and ELK, aiming to provide a fast, feasible path for implementing microservices and DevOps in enterprises.

Four parts of the talk:

Personal cloud‑computing journey and market‑level thoughts.

Understanding of microservices and DevOps.

Platform‑level practice and support based on a microservice‑driven DevOps platform.

References and summary.

Cloud‑computing journey: Starting in 2009 with a traditional PaaS platform, moving to OpenStack/CloudStack in 2011, participating in hybrid‑cloud projects in 2013, and by 2015 adopting container‑based microservice architecture to build an enterprise‑grade DevOps digital platform.

The current market is described using a UCOBP view rather than the classic IaaS/PaaS/SaaS distinction, highlighting stable core advantages for operators and hardware vendors, while integration and service providers face intense competition.

Key success factors for entering broad fields such as cloud computing, big data, AI, etc., are clear positioning, a “thick+thin” approach (complex enterprise processes combined with selective use of open‑source technologies), and strong organizational DNA.

Microservices and DevOps understanding: The author explains the evolution from traditional monoliths to vertical architecture, SOA, and finally microservice‑centric DDD‑driven services, and defines DevOps as a movement that tightly couples development and operations to achieve rapid, reliable releases.

Relationships among cloud infrastructure, DevOps, and microservices are illustrated: containers or VMs provide the resource pool; DevOps offers automation and workflow; microservices coexist with legacy systems (Bi‑modal) and are introduced gradually.

Platform practice: The platform uses Kubernetes for container orchestration, Flannel for networking, Ceph for storage, SpringBoot + Netflix components for microservices, and integrates documentation, mock, automated testing, journald + Fluentd for logging, and time‑series databases for monitoring.

Key microservice capabilities highlighted include isolation, scaling, upgrade/rollback, circuit‑break, service discovery, and registration (etcd). The runtime model follows Kubernetes concepts of Pods, Replication, and Services, emphasizing isolation, scalability, and external access.

DevOps systems built into the platform (15 subsystems) cover IAM, SPM, SCM, SRM, SEM, QAF, UMC, VCS, CI, BPR, DPR, PM, IM, TM, and MKT, forming a complete lifecycle management suite.

Challenges encountered on public cloud (e.g., CoreOS upgrade issues, security‑group configuration) and private cloud pilot experiences are also discussed.

References and summary: The author lists valuable materials such as Gartner’s DevOps assessment model, the 12‑Factor App, Google’s Borg design, and emphasizes the importance of quality attributes, cloud‑native goals, and continuous learning.

Finally, readers are invited to join the author’s cloud‑computing discussion group for deeper exchange.