Happy Wednesday! At some point every growing product hits the same wall: one database, one machine, and a table that no longer fits in memory. The fix almost always involves today’s word — shard — and the vocabulary around it shows up constantly in design docs, database vendor talks, and any interview question that starts with “how would you scale this to a billion rows.”
📖 Word of the Day: Shard
- Word: shard
- IPA: /ʃɑːrd/
- Vietnamese meaning: mảnh dữ liệu — một phần của tập dữ liệu lớn được chia nhỏ và lưu trên một node riêng biệt để dàn trải tải
- Example sentences:
- “We sharded the users table by region, so European users live on one set of servers and Asian users on another.”
- “Picking the wrong shard key early on cost us a six-month migration later.”
- “Each shard holds about 50 million rows, and we add a new one whenever that number climbs too high.”
Listen and practice pronunciation:
Vocabulary Table — Data Partitioning
| Phrase | Vietnamese | Example |
|---|---|---|
| Partition key | khóa phân vùng | ”We chose user_id as the partition key so all of one user’s data lands on the same shard.” |
| Hot partition | phân vùng quá tải | ”One celebrity account turned into a hot partition and slowed down the whole shard.” |
| Consistent hashing | băm nhất quán | ”Consistent hashing means adding a new node only reshuffles a small slice of the data, not all of it.” |
| Quorum | số phiếu tối thiểu để xác nhận | ”We need a quorum of two out of three replicas to agree before a write counts as successful.” |
| Rebalance | cân bằng lại dữ liệu | ”Rebalancing kicked off automatically after we added the fourth shard.” |
🗣 Pronunciation Guide
“Shard” breakdown — /ʃɑːrd/:
- One syllable: SHARD — rhymes with “hard” and “card”
- Starts with the “sh” sound, like in “share” or “shark,” not an “s” sound
- The “ar” is a long, open vowel — hold it slightly longer than you would in casual speech
- Ends with a clear “d,” don’t let it trail off into “shart” or “shar”
Practice sentence (read aloud 3x):
“If you pick the wrong shard key today, you’ll be rebalancing a hundred million rows at 2 a.m. six months from now.”
Read it slowly the first time to nail the “shard” and “rebalancing” sounds, emphasize “wrong” and “2 a.m.” the second time, then read it at natural speed the third time like you’re warning a teammate.
✍️ Exercise 1: Fill in the Blank
Fill each blank with one term: shard, partition key, hot partition, consistent hashing, quorum, rebalance.
- “We use _______ so scaling out only moves a small percentage of the data instead of everything.”
- “Traffic to one customer ID turned their data into a _______, slowing down every other tenant on the same node.”
- “A write only succeeds once we get a _______ of replicas confirming it.”
- “Choosing the right _______ up front saves you a painful migration later.”
- “After adding two new nodes, the cluster had to _______ automatically.”
Show answers
- consistent hashing
- hot partition
- quorum
- partition key
- rebalance
✍️ Exercise 2: Translate to English
Translate these Vietnamese sentences using today’s vocabulary.
- “Chúng tôi chia database thành nhiều mảnh (shard) để xử lý hàng triệu người dùng.”
- “Chọn sai khóa phân vùng khiến một node bị quá tải dữ liệu.”
- “Hệ thống cần số phiếu tối thiểu từ ba bản sao trước khi xác nhận ghi dữ liệu thành công.”
Show answers
- “We sharded the database into multiple pieces to handle millions of users.”
- “Choosing the wrong partition key caused one node to get overloaded with data.”
- “The system needs a quorum from three replicas before confirming a write.”
💬 Idiom of the Day: “Spread too thin”
- Vietnamese meaning: dàn trải quá mức, chia nhỏ nguồn lực đến mức không còn hiệu quả
- Example 1: “We sharded into twelve tiny pieces before we actually needed to, and now our ops team is spread too thin just keeping track of them.”
- Example 2: “Don’t shard prematurely — you’ll spread your data (and your team) too thin for a problem you don’t have yet.”
📺 Recommended Watching
- ByteByteGo — visual walkthroughs of sharding strategies and real-world partitioning decisions
- Hussein Nasser — deep, spoken-word explanations of databases, replication, and partitioning internals
- YouGlish — hear “shard” and “partition” pronounced across real technical talks
🎯 Daily Challenge
Think of one app you use daily that must be sharded behind the scenes — a banking app, a food delivery app, a messaging app. Say out loud, in one sentence, what you think its partition key probably is and why. Thirty seconds, out loud, no notes.