Trung Hieu Nguyen
Adaptive Radix Tree
A practical walkthrough of radix trees and adaptive radix trees, covering node layouts, path compression, binary-comparable keys, algorithms, and synchronization techniques.
Writing
55 articles
Trung Hieu Nguyen
Split-Ordered Lists: Lock-Free Extensible Hash Tables
How split-ordered lists build a lock-free, extensible hash table: ordering items by the bit-reversed value of their hash, using dummy nodes as stable bucket anchors, and a segmented bucket array to grow without moving items.
Trung Hieu Nguyen
TDSQL: Tencent Distributed Database System
How Tencent built a MySQL-compatible distributed database serving 30,000 enterprises with physical replication, Raft-based consistency, lock-free internals, and horizontal scaling explained.
Trung Hieu Nguyen
Robin Hood Hashing – Fairness in Collision Resolution
A write-up about Robin Hood hashing - fairness in collision resolution.
Trung Hieu Nguyen
Time, Clocks and Ordering in Distributed Systems
A write-up about time, clocks, and ordering in distributed systems.
Trung Hieu Nguyen
Understanding Pratt Parsing: Top‑Down Operator Precedence
How Pratt parsing uses binding power to build expression trees efficiently for expression evaluation, with a step‑by‑step Go implementation and extensions for unary minus, parentheses, exponentiation, and variables.
Trung Hieu Nguyen
TiDB: A Raft-based HTAP Database
A write up about TiDB’s Raft-based HTAP architecture: exploring multi-Raft storage, learner replicas, DeltaTree columnar engine, MVCC transactions, read optimizations, and how TiKV and TiFlash deliver scalable, consistent, and isolated OLTP + real-time OLAP in one system.
Trung Hieu Nguyen
Percolator: Large-scale Incremental Processing Using Distributed Transactions and Notifications
An engineer’s walkthrough of Google Percolator: from MapReduce to incremental processing, exploring transactions, lazy lock cleanup, observers, and performance optimizations on Bigtable.
Trung Hieu Nguyen
Multi-version Concurrent Control
A practical deep dive into MVCC: concurrency protocols, version storage, garbage collection, and index design.
Trung Hieu Nguyen
Google Spanner: Google’s Globally-Distributed Database
Explore how Google Spanner combines SQL, global distribution, TrueTime, and Paxos to deliver externally consistent transactions at planet scale.
Trung Hieu Nguyen
SegCache: A memory-efficient and scalable in-memory key-value cache for small objects
SegCache is an in-memory key–value cache built for small objects, improving memory efficiency, throughput, and multi-core scalability with TTL-bucketed segments, compact hash metadata (bulk chaining), and segment-level proactive expiration plus merge-based eviction.
Trung Hieu Nguyen
Google MapReduce: Simplified Data Processing on Large Clusters
A clear, concise walkthrough of Google’s MapReduce, explaining how it simplified large-scale data processing while revealing its key design insights and limitations.
Trung Hieu Nguyen
Distributed Transaction At Scale In Amazon DynamoDB
How DynamoDB delivers ACID at scale: timestamp ordering instead of locks, a distributed 2PC, writeless reads, buffered writes, and recovery that keeps latency predictable.
Trung Hieu Nguyen
Amazon DynamoDB: A Scalable, Predictably Performant, and Fully Managed NoSQL Database Service
A practical tour of DynamoDB’s architecture: how it partitions data, controls throughput with GAC, balances hot spots, and keeps availability and durability high.
Trung Hieu Nguyen
Google Bigtable: A Distributed Storage System for Structured Data
Understand Bigtable end-to-end: sparse data model, tablet assignment and lookup (root→metadata→user), Chubby coordination, GFS/SSTables, compactions, caching, Bloom filters, commit logs, and recovery.
Trung Hieu Nguyen
Wormhole: Reliable Pub-Sub to support Geo-replicated Internet Services
Explore how Facebook’s Wormhole powers real-time data propagation at massive scale. This post breaks down its core design - flows, caravans, data-markers, and multi-datacenter failover, and explains how it achieves trillions of reliable, in-order updates daily with millisecond latency.
Trung Hieu Nguyen
Introduction To ZooKeeper
A practical tour of ZooKeeper’s data model, common recipes (locks, barriers, group membership), and the internals (ZAB, snapshots) that make it reliable.
Trung Hieu Nguyen
Chubby: Lock Service For Loosely-Coupled Distributed Systems
Chubby explained: why a lock service (not a library), how caching/invalidations work, master failover and sessions, plus partitioning, proxies, and name-service use
18 of 55 articles