ACA Big Data study plan

Map the core Big Data concepts (batch vs. streaming, OLAP vs. OLTP, data warehouse vs. data lake) to the Alibaba Cloud product that implements each

Describe MaxCompute's architecture in writing: project/table model, storage-compute separation, and the role of units/quota

Create a MaxCompute project, load a sample table, and run five SQL queries covering filter, join, group-by, and window functions

Compare MaxCompute SQL syntax against standard SQL and list at least five concrete differences

Build partitioned and lifecycle-managed tables and demonstrate partition pruning with EXPLAIN

Implement a small ETL pipeline in MaxCompute using INSERT OVERWRITE across staging and target tables

Explain MaxCompute's data types, table types, and when to use external tables

Quantify how partitioning and file sizing affect scan volume and cost on your sample dataset

Describe DataWorks' component model (DataStudio, Data Integration, scheduling, operations center) in your own words

Create a DataWorks workspace and configure a synchronization task from a source into MaxCompute

Build a scheduled workflow with three dependent nodes and a parameter passed between them

Configure basic alerting and rerun/recover a failed instance in Operations Center

Implement a full batch pipeline in DataWorks: ingest, transform in MaxCompute, and publish a result table

Apply DataWorks scheduling parameters, node dependencies, and baseline/branch logic to a realistic scenario

Tune a slow MaxCompute query by rewriting joins and reducing data movement, and record the improvement

Document data quality checks (row counts, null checks, referential checks) placed between pipeline stages

Contrast batch, micro-batch, and true streaming models and identify which Alibaba Cloud services fit each

Set up a streaming ingestion path (e.g., DataHub or Log Service into a downstream store) and push test events through it

Write a windowed streaming computation and explain the effect of watermarks and late-arriving data

Identify failure and backpressure scenarios in a streaming job and describe the recovery behavior

Integrate the streaming output with the batch layer to form a lambda-style architecture diagram

Use DataWorks to orchestrate and monitor the streaming-adjacent jobs and set up an alert rule

Measure end-to-end latency of your streaming demo and document where the time is spent

List three production risks in real-time pipelines (schema drift, duplicates, exactly-once concerns) and a mitigation for each

Connect a BI/analytics tool to a MaxCompute or analytics-service dataset and build three chart types

Apply data security controls: describe RAM-based access control, column-level permissions, and data masking options

Configure audit/logging visibility for a project and interpret one audit log entry

Design a dashboard for a stated business question and justify each metric and filter

Re-derive the end-to-end data lifecycle across all five domains without notes and identify your weakest two areas

Complete timed practice questions weighted toward domains 2, 3, and 4 and review every miss to root cause

Produce a one-page cheat sheet of product-to-use-case mappings, SQL tuning rules, and security controls

Re-run one earlier lab from memory to confirm retention