Understanding Distributed Systems

Copyright
About the author
Acknowledgements
Preface
	Who should read this book
Introduction
	Communication
	Coordination
	Scalability
	Resiliency
	Operations
	Anatomy of a distributed system
I Communication
	Reliable links
		Reliability
		Connection lifecycle
		Flow control
		Congestion control
		Custom protocols
	Secure links
		Encryption
		Authentication
		Integrity
		Handshake
	Discovery
	APIs
		HTTP
		Resources
		Request methods
		Response status codes
		OpenAPI
		Evolution
II Coordination
	System models
	Failure detection
	Time
		Physical clocks
		Logical clocks
		Vector clocks
	Leader election
		Raft leader election
		Practical considerations
	Replication
		State machine replication
		Consensus
		Consistency models
			Strong consistency
			Sequential consistency
			Eventual consistency
			CAP theorem
		Practical considerations
	Transactions
		ACID
		Isolation
			Concurrency control
		Atomicity
			Two-phase commit
		Asynchronous transactions
			Log-based transactions
			Sagas
			Isolation
III Scalability
	Functional decomposition
		Microservices
			Benefits
			Costs
			Practical considerations
		API gateway
			Routing
			Composition
			Translation
			Cross-cutting concerns
			Caveats
		CQRS
		Messaging
			Guarantees
			Exactly-once processing
			Failures
			Backlogs
			Fault isolation
			Reference plus blob
	Partitioning
		Sharding strategies
			Range partitioning
			Hash partitioning
		Rebalancing
			Static partitioning
			Dynamic partitioning
			Practical considerations
	Duplication
		Network load balancing
			DNS load balancing
			Transport layer load balancing
			Application layer load balancing
			Geo load balancing
		Replication
			Single leader replication
			Multi-leader replication
			Leaderless replication
		Caching
			Policies
			In-process cache
			Out-of-process cache
IV Resiliency
	Common failure causes
		Single point of failure
		Unreliable network
		Slow processes
		Unexpected load
		Cascading failures
		Risk management
	Downstream resiliency
		Timeout
		Retry
			Exponential backoff
			Retry amplification
		Circuit breaker
			State machine
	Upstream resiliency
		Load shedding
		Load leveling
		Rate-limiting
			Single-process implementation
			Distributed implementation
		Bulkhead
		Health endpoint
			Health checks
		Watchdog
V Testing and operations
	Testing
		Scope
		Size
		Practical considerations
	Continuous delivery and deployment
		Review and build
		Pre-production
		Production
		Rollbacks
	Monitoring
		Metrics
		Service-level indicators
		Service-level objectives
		Alerts
		Dashboards
			Best practices
		On-call
	Observability
		Logs
		Traces
		Putting it all together
	Final words