2: Core Concepts of Distributed Systems

Learning Objectives

• Define cloud computing using the NIST model and explain its five essential characteristics: on-demand self-service, broad network access, resource pooling, rapid elasticity, and measured service.
• Trace the historical evolution of distributed systems — from mainframes and cluster computing, through grid computing, to modern cloud — and explain how each stage built upon the previous one.
• Explain the "cluster as a single computer" concept, including the warehouse-scale computer model, single system image, and how clusters achieve fault tolerance and high availability through redundancy and distributed protocols.
• Describe workload management and resource alignment strategies in data centers, including scheduling algorithms, priority-based preemption, oversubscription, workload placement, autoscaling, and multi-tenancy.
• Explain how Google's Borg and Kubernetes orchestrate distributed workloads, including their architectures, scheduling approaches, fault recovery mechanisms, and how Kubernetes brought cluster management concepts to the broader industry.
• Describe the Hadoop ecosystem's approach to distributed data processing, including how HDFS provides fault-tolerant storage through replication and how MapReduce exploits data locality and parallelism for large-scale computation.
• Analyze Netflix's microservices architecture as a cloud-native design pattern, identifying key components (Eureka, Ribbon, Hystrix, Zuul) and explaining how they enable service discovery, load balancing, fault isolation, and elastic scaling.
• Explain how Meta (Facebook) addresses distributed systems challenges at social network scale, including layered caching with eventual consistency, database sharding, polyglot persistence, and global data replication across regions.