Tag: design-system
All the articles with the tag "design-system".
-
System Design - How Futexes Work (The Foundation of Every Lock You Use)
How Linux futexes work: the hybrid userspace/kernel synchronization primitive that powers every mutex, condition variable, and semaphore in modern programs. Covers the fast-path CAS, slow-path syscall, kernel hash table, the three-state mutex optimization, and a source code walkthrough from the Linux kernel and glibc.
-
System Design - How Apache ZooKeeper Works
How Apache ZooKeeper works: the coordination service behind Kafka, Hadoop, and HBase. Covers ZAB consensus protocol, znodes, watches, ephemeral nodes, leader election, sessions, and a source code walkthrough from the apache/zookeeper repository.
-
AI/ML - How BPE Tokenization Works in Large Language Models
How Byte Pair Encoding (BPE) tokenization works in large language models: the training algorithm, encoding and decoding, tiktoken and sentencepiece implementations, vocabulary construction, and why subword tokenization strikes the right balance between characters and words.
-
System Design - How the Linux Kernel Network Stack Works
How a packet travels from NIC hardware to your application's socket: interrupts, NAPI, sk_buff, protocol layers, and socket receive queues — a walkthrough of the Linux kernel network stack.
-
System Design - How Protocol Buffers (Protobuf) Encoding Works
How Protocol Buffers encoding works internally: wire types, varint encoding, field tags, length-delimited records, and a byte-level walkthrough showing why protobuf messages are compact and fast to parse.
-
System Design - How Database Connection Pooling Works
How database connection pooling works: the cost of raw connections, pool lifecycle, idle/active management, connection multiplexing modes, HikariCP internals, and PgBouncer architecture with source code walkthrough.