9. The Trouble with Distributed Systems
9.13 Terminology introduced here
Terminology introduced here
65- partial failure
- nondeterminism
- shared-nothing
- asynchronous packet network
- congestion control / flow control / backpressure
- network partition / netsplit
- RST / FIN
- ICMP Destination Unreachable
- false positive / false negative in failure detection
- unbounded delay
- network congestion
- head-of-line queueing
- steal time
- noisy neighbour
- jitter
- Phi Accrual failure detector
- circuit switching vs packet switching
- bounded delay
- ATM / InfiniBand / QoS / L4S
- time-of-day clock
- monotonic clock
- epoch
- NTP
- slewing
- stepping
- clock drift (ppm)
- leap second
- smearing
- PTP
- confidence interval
- TrueTime / ClockBound
- commit wait
- logical clock vs physical clock
- lease
- process pause
- stop-the-world GC pause
- live migration
- thrashing
- SIGSTOP / SIGCONT
- hard real-time
- RTOS
- quorum
- majority
- zombie
- fencing
- STONITH
- fencing token / sequencer / epoch number / ballot number / term number
- conditional write / CAS
- Byzantine fault
- Byzantine Generals Problem
- two generals problem
- system model
- synchronous / partially synchronous / asynchronous model
- crash-stop (fail-stop)
- crash-recovery
- stable storage
- limping node / gray failure / fail-slow
- safety property
- liveness property
- model checking
- TLA+
- fault injection
- Jepsen
- chaos engineering
- deterministic simulation testing (DST)