CDESS study plan
CDESS · Give each week a clear focus.
8 weeks · About 6 hours per week
Data Engineering
Map the Data Engineering domain's weighting and enumerate every sub-topic you must be able to explain without notes
Set up a local practice environment with Spark, Hive-style tables, and a small dataset you will reuse all eight weeks
Write and run batch transformations covering joins, aggregations, window functions, and partitioning on that dataset
Diagnose at least three performance problems (skew, shuffle, small files) and record the fix you applied for each
Your checkpoint
A one-page topic inventory for Domain 1 with a self-rating for each item, plus a working local pipeline script that runs end to end
Data Engineering
Compare file formats (Parquet, ORC, Avro, JSON) and justify a choice for a given workload in writing
Explain partitioning versus bucketing trade-offs and demonstrate both on the practice dataset
Implement an incremental load pattern with idempotent reruns and a deduplication step
Document how schema evolution is handled across formats and table types
Your checkpoint
A written design note (600+ words) recommending storage format and layout for a sample workload, with benchmark timings from your local runs
Data in Motion / Streaming
Define the streaming vocabulary precisely: event time, processing time, watermark, trigger, window, checkpoint
Build a producer-consumer pipeline that ingests events and writes results with exactly-once semantics
Implement tumbling, sliding, and session windows and explain when each applies
Force late-arriving data through the pipeline and tune the watermark until results are correct
Your checkpoint
A running streaming job plus a test log showing correct results after deliberately injecting out-of-order and late events
Data in Motion / Streaming
Compare delivery guarantees (at-most-once, at-least-once, exactly-once) and identify what each requires from source and sink
Explain log-based broker fundamentals: partitions, offsets, consumer groups, retention, compaction
Design a schema governance approach for streams, including compatibility rules and a registry
Troubleshoot a simulated consumer lag problem and document the root cause and remedy
Your checkpoint
A streaming architecture diagram with annotated delivery guarantees and a lag-troubleshooting write-up, reviewed against your own checklist
Open Data Lakehouse and Apache Iceberg
Explain the Iceberg table format internals: metadata files, manifests, snapshots, and the catalog's role
Distinguish copy-on-write from merge-on-read and predict which one suits a given update pattern
Perform schema evolution, partition evolution, and hidden partitioning on a practice table
Demonstrate snapshot isolation, time travel, and rollback on the same table
Your checkpoint
A lab log with commands and outputs for each Iceberg operation, plus a short explanation of what changed in metadata after each one
Open Data Lakehouse and Apache Iceberg
Execute compaction and manifest rewriting, then measure the effect on query latency
Explain how delete files accumulate and how to keep them under control
Compare open table format behavior with a traditional warehouse table for updates and deletes
Describe how a lakehouse separates storage, catalog, and compute, and where governance attaches
Your checkpoint
A before-and-after maintenance report with file counts, snapshot counts, and query timings from your practice table
Cloudera Data Platform Administration and Security
Outline the platform's control plane versus data plane services and what each one manages
Describe authentication, authorization, and audit flows end to end, naming the component responsible at each step
Configure role-based access on a database, table, and column, then verify effective permissions from a second identity
Explain how workload isolation and resource queues affect multi-tenant clusters
Your checkpoint
A permissions matrix you built and validated by testing access as two different users, with the results recorded
Data Engineering & Data in Motion / Streaming & Open Data Lakehouse and Apache Iceberg & Cloudera Data Platform Administration and Security
Re-rate every topic from your week 1 inventory and close the three weakest gaps
Complete a timed mixed-domain practice set and analyze every miss by root cause, not by topic name
Re-derive the streaming and Iceberg internals from memory on a blank page
Write the operational checklist you would follow to secure and maintain a lakehouse table
Your checkpoint
A final gap report listing remaining weak areas with a concrete remediation action for each, plus your scored practice-set results
Milestones
Week 2: You can design a batch layout (format, partitioning, incremental load) on paper and defend it without looking anything up
Week 4: Your streaming job produces correct results with late data, and you can state the delivery guarantee at every hop in the pipeline
Week 6: You can draw Iceberg metadata from memory and show measured improvement after compaction
Week 8: Every Domain 4 access-control claim in your permissions matrix was verified by an actual test, and your practice-set score shows no domain below your target threshold