Bookshelf, Reading Tracker & Core Mental Models
Foundational computer science, distributed systems, and low-level engineering literature distilled into actionable architectural mental models. Every canonical book is directly linked to chapter notes in this repository and interactive 3D flashcards.
Key Mental Models Derived from Canonical Literature
Architectural invariants and trade-offs that distinguish senior from staff-level systems design.
LSM-Tree vs B-Tree Duality
LSM-trees turn random writes into sequential disk writes at the cost of compaction overhead and read-amplification. B-trees optimize point reads with deterministic tree depths at the cost of random write I/O.
🃏 Practice FlashcardFutexes & Fast User-Space Locking
Uncontended locks should execute via hardware CAS without a syscall. Kernel context switching into sleep queues only occurs when contention is detected, saving thousands of CPU cycles.
🃏 Practice FlashcardHappens-Before & Memory Visibility
Synchronization isn't just about atomic mutual exclusion. Without happens-before edges (volatile writes or lock releases), CPU registers and out-of-order execution leave state changes invisible to other cores.
🃏 Practice FlashcardFencing Tokens & Lease Invalidation
Distributed locks with TTLs (e.g. Redlock) are unsafe against GC pauses unless target storage enforces strictly monotonic fencing tokens on every mutation (`version > prev_version`).
🃏 Practice Flashcard