Skip to content

DE Mastery GitHub

Start

Curriculum
Roadmap
How to study with this course

Tracks

FND Foundations of distributed data

Overview
01 Why distributed?
02 Partitioning & shuffles
03 Replication & consensus
04 Time
05 Delivery semantics
06 Storage formats
07 Back-of-envelope estimation

SPK Apache Spark

Overview
01 Spark architecture
02 Lazy evaluation and the DAG
03 The DataFrame API
04 Spark SQL and data sources
05 Catalyst and the physical plan
06 Joins and aggregations at scale
07 Partitioning, bucketing, and the small-file problem
08 Memory, caching, and Adaptive Query Execution
09 UDFs and the Pandas API
10 Structured Streaming I — the micro-batch engine and exactly-once
11 Structured Streaming II — watermarks, state, and streaming SCD2
12 Testing and production Spark
BYO Build your own mini columnar query engine

LKH Lakehouse & open table formats

Overview
01 The Iceberg metadata tree
02 Commits & concurrency
03 Catalogs, branching & write-audit-publish
04 Row-level changes — SCD2, CoW & MoR
05 Table layout & maintenance
06 Delta Lake
07 Hudi & the format landscape
08 Engines on the lake
BYO Build your own mini table format

JVM Java for data engineers

Overview
01 Java for Python developers
02 Collections and streams
03 Build and dependencies
04 Serialization
05 Concurrency and the JVM
06 Testing and debugging

KFK Apache Kafka

Overview
01 The log — events, state, retention and replay
02 Topics, partitions, replication — leaders, ISR, acks, durability
03 Producers — batching, idempotence, ordering under retries
04 Consumers & groups — manual commits, at-least-once, no loss across crashes
05 Transactions & exactly-once — read-process-write, read_committed, fencing
06 Schema Registry & contracts
07 Kafka Connect & CDC
08 Kafka Streams
09 Event & topic design
10 Testing & observability (apps)
11 Cluster configuration & partition reassignment
12 Security — SASL/SCRAM authentication & ACL authorization
13 Deployment architecture — multi-region DR
14 Operating & troubleshooting a cluster
BYO Build your own mini Kafka

FLK Apache Flink

Overview
01 Philosophy & architecture
02 DataStream API
03 Time & windows
04 Flink SQL & dynamic tables
05 Joins — regular, interval, temporal (point-in-time) & lookup
06 State, checkpoints & exactly-once — surviving a TaskManager crash
07 Process functions — timers, keyed state & broadcast rules
08 Table API & PyFlink — one pipeline, two languages
09 Streaming patterns — dedup, top-N, split/merge, MATCH_RECOGNIZE, OVER
10 CDC & lake sinks — Debezium changelog to Iceberg, commit on checkpoint, schema evolution
11 Performance & state at scale — backpressure, parallelism, redistribution, skew
12 Deploy & operate — upgrade a stateful job via savepoint without losing state
BYO Build your own mini stream processor

STH Streamhouse

Overview
01 Architecture history — Lambda → Kappa → lakehouse → streamhouse
02 Apache Paimon — primary-key tables, merge engines, partial-update
03 Apache Fluss — PK tables, lake tiering, streaming vs batch consistency
04 Iceberg as streaming storage — commit interval vs small files
05 Streaming dimensional modeling — SCD2 dim + fct with temporal (as-of) join
06 Streaming data quality — in-flight DQ, quarantine / DLQ, completeness
07 Serving layer — OLAP on the streamhouse
08 Latency tiers & cost — seconds, minutes, hours

Foundations of distributed data · FND-07

Back-of-envelope estimation

3 h
requires FND-01
requires FND-06

From your world. Sizing the commerce mart; translating 'bytes scanned' into an Athena bill; capacity-planning a feed

Lesson Exercises 1 Quiz Interview drills Explain it back

Practice quiz with instant feedback. For a scored run that counts toward your local progress, use uv run de quiz FND-07.

1A feed runs at 1 event/sec sustained. Roughly how many events per year? A~31.5 million B~86 thousand C~3.15 billion D~1 million

1/sec x ~3.15e7 s/year ~ 31.5M/year (and ~86k/day). You should memorize these constants.

Source: https://static.googleusercontent.com/media/research.google.com/en//people/jeff/stanford-295-talk.pdf

22,000,000 events/sec at 100 bytes/event is what raw throughput? A~200 MB/s B~2 MB/s C~20 GB/s D~2 GB/s

2e6 x 100 = 2e8 B/s = 200 MB/s.

Source: https://static.googleusercontent.com/media/research.google.com/en//people/jeff/stanford-295-talk.pdf

3Which assumptions change a 'bytes stored per year' estimate the MOST, and so you must state them? ACompression ratio and replication factor BThe programming language CThe color of the dashboard DThe number of engineers on the team

Stored bytes = raw x (1/compression) x replication. Each of these two factors moves the number by 3-10x. So you must state them.

Source: https://parquet.apache.org/docs/file-format/

4What drives the cost of an Athena-style query? ATB scanned, which column and partition pruning reduce BNumber of columns in the table CRows returned to the client DQuery length in characters

Serverless engines bill per TB scanned. Pruning (columns, partitions, row groups) directly lowers the bill. The number of rows returned is not the metric.

Source: https://aws.amazon.com/athena/pricing/

5A feed stores Y bytes/year. For a 7-year SOX retention estimate, the storage cost is approximately: AY x 7 x ($/GB-month) x 12 (converting bytes to GB) BY x 7 x ($/TB-scanned) CY / 7 DY x 7 (bytes equal dollars)

Storage cost = GB stored x $/GB-month x months. Over 7 years, the number of months is 7x12. Scanned-TB pricing is for queries, not for storage.

Source: https://aws.amazon.com/s3/pricing/

6What is the goal of a back-of-envelope estimate? AThe right order of magnitude with stated assumptions, fast BA number accurate to within 1% CTo replace load testing permanently DTo avoid writing down any assumptions

An estimate aims for the correct order of magnitude (~2x) with explicit assumptions that people can challenge. More precision is wasted, because the inputs are rough.

Source: https://static.googleusercontent.com/media/research.google.com/en//people/jeff/stanford-295-talk.pdf

7Why can RAW event bytes mislead a STORAGE estimate? (select all) (select all that apply) AColumnar compression typically shrinks stored size 3-10x BReplication multiplies stored bytes (e.g. 3x) CRaw bytes already include indexes and small-file overhead DRaw bytes ignore the compression and replication factors

Stored size differs from raw size because of compression (down) and replication or overhead (up). Raw bytes do NOT include indexes or overhead, so C is false.

Source: https://parquet.apache.org/docs/file-format/

PreviousFND-06 · Storage formats NextSPK-01 · Spark architecture

DE Mastery · source on GitHub · Found an error? Report it on GitHub