Cloud Pro: AWS Analytics: Choose, Compare and Launch

What's inside

1. Level 1: The AWS Analytics Map: Every Service on One PageFree
   You build a single grouped map of every AWS analytics and data service, covering warehouses and lakehouses, query engines, integration and pipelines, streaming and messaging, big data processing, catalogues and governance, and BI. The map records current and former names (Kinesis Data Analytics is now Managed Service for Apache Flink, Kinesis Data Firehose is now Amazon Data Firehose, and QuickSight has become Quick Sight inside Amazon Quick Suite, later shortened to Amazon Quick). It also marks what is closed to new customers, such as AWS Data Pipeline, Timestream for LiveAnalytics, IoT Analytics and Glue Ray jobs.
2. Level 2: The Standard Analytics ProfileFree
   You learn a reusable profile that captures a workload's needs before any service is chosen. It records sources and volume, batch or streaming, how fresh results must be, query patterns and concurrency, who the users are, whether open formats such as Iceberg or Delta are required, governance and lineage, data residency, team skills and the preferred cost model.
3. Level 3: How AWS Analytics Is Billed: The Cost Levers
   You learn the handful of billing models that recur across every service: provisioned nodes or capacity units, serverless compute units, data scanned per query, storage, requests, and streaming throughput by shard or partition. You learn to price each one with a calculator and dated example numbers.
4. Level 4: Who Operates What, and the Nine-Part Service Lesson
   You learn to sort services by who runs what: self-managed on EC2, managed clusters, serverless, and fully managed SaaS. You also learn the nine-part template every later service lesson follows, and set up a safe hands-on account with budgets, tags, IAM and a clean-up routine.
5. Level 5: Amazon Redshift Provisioned (RA3 Clusters)
   You cover provisioned Redshift on RA3 nodes from end to end: managed storage, separate compute and storage, concurrency scaling, data sharing and Spectrum. The lesson includes dated limits, node and reserved-instance pricing, and a quick start that creates a cluster, loads data and then deletes it.
6. Level 6: Amazon Redshift Serverless
   You learn how Redshift Serverless differs from provisioned clusters: namespaces and workgroups, RPU-based billing, base and maximum capacity, usage limits, and current dated limits. A quick start creates a workgroup, queries data and removes it.
7. Level 7: Amazon Athena SQL: Querying the Lake
   You learn Athena's SQL engine: per-terabyte-scanned pricing versus provisioned capacity reservations, workgroups and their scan limits, the dated service quotas that bite, and how partitioning, columnar formats and partition projection change both cost and speed. A quick start runs a query on S3 data and cleans up.
8. Level 8: Athena Beyond Basic SQL: Federation, Iceberg, Spark
   You go beyond basic Athena SQL: federated queries through Lambda connectors, reading and writing Iceberg tables (including MERGE and time travel), and Athena for Apache Spark notebooks, each with its own limits and costs.
9. Level 9: Amazon EMR on EC2
   You learn classic EMR clusters: primary, core and task nodes; instance fleets and Spot; release labels; managed scaling; and running Spark, Hive, Trino and HBase. The lesson covers dated limits, the EMR uplift on top of EC2 pricing, and a quick start that runs one Spark step on a cluster that terminates itself.
10. Level 10: Amazon EMR Serverless
    You learn EMR Serverless applications, pre-initialised capacity, per-vCPU and per-GB billing, maximum capacity limits and job-run quotas. A quick start submits a Spark job and deletes the application.
11. Level 11: Amazon EMR on EKS
    You learn how EMR on EKS runs Spark on an existing Kubernetes cluster through virtual clusters and job runs. You see the extra platform work it brings, its pricing, and when it makes sense for teams that already run EKS.
12. Level 12: AWS Glue Data Catalog and Crawlers
    You learn the Glue Data Catalog as the shared metastore for Athena, Redshift Spectrum, EMR and Lake Formation. The lesson covers databases, tables, partitions and partition indexes, crawlers and their costs, catalog quotas and storage or request pricing, plus a quick start that crawls an S3 dataset.
13. Level 13: AWS Glue ETL Jobs and Pipelines
    You learn Glue's Spark and Python shell jobs, worker types and DPU-hour billing, Flex execution, job bookmarks, Glue Studio visual jobs, workflows and triggers, and data quality rules. The lesson notes that Glue Ray jobs went into maintenance and closed to new customers from 30 April 2026. A quick start runs a CSV-to-Parquet job.
14. Level 14: AWS Lake Formation Permissions
    You learn how Lake Formation layers fine-grained access on top of the Data Catalog and S3: data lake administrators, registered locations, database, table, column, row and cell-level permissions, LF-Tags, cross-account sharing, and hybrid access mode. It finishes with a quick start that grants a column-restricted view.
15. Level 15: Amazon S3 Tables: Managed Apache Iceberg
    You learn S3 table buckets, namespaces and tables, automatic compaction and snapshot management, how they integrate with the catalog and analytics services, and their dated limits and pricing (storage, requests, monitoring and compaction). A quick start creates a table and queries it.
16. Level 16: SageMaker Unified Studio and Amazon DataZone
    You learn how Amazon DataZone and SageMaker Unified Studio (part of the next generation of Amazon SageMaker) provide domains, projects, a business data catalogue, publish and subscribe sharing, governed access and lineage across analytics services, along with their limits and pricing. A quick start publishes and subscribes to a dataset.
17. Level 17: Amazon Kinesis Data Streams
    You learn shards, partition keys, provisioned versus on-demand capacity modes, retention, enhanced fan-out, the per-shard throughput limits from the docs, and shard-hour versus on-demand pricing. A quick start puts records into a stream, reads them back and deletes the stream.
18. Level 18: Amazon Data Firehose
    You learn Firehose (formerly Kinesis Data Firehose) as the managed delivery service: sources, destinations including S3, Redshift, OpenSearch, Splunk, HTTP endpoints and Iceberg tables, buffering hints, format conversion, dynamic partitioning, Lambda transforms, and its quotas and per-GB pricing. A quick start delivers to S3.
19. Level 19: Amazon Managed Service for Apache Flink
    You learn the managed Flink service (formerly Kinesis Data Analytics): applications, KPUs and parallelism, checkpoints and snapshots, Studio notebooks, supported Flink versions, limits and KPU-hour pricing. A quick start runs a windowed aggregation over a Kinesis stream.
20. Level 20: Amazon MSK Provisioned Clusters
    You learn provisioned Amazon MSK for Apache Kafka: Standard and Express brokers, sizing brokers and storage, tiered storage, authentication options, the dated partition and throughput limits, and broker-hour and storage pricing. A quick start creates a cluster, produces and consumes a topic, then deletes it.
21. Level 21: Amazon MSK Serverless and MSK Connect
    Learn when MSK Serverless's per-cluster and per-partition limits and its throughput pricing beat provisioned brokers, and how MSK Connect runs managed Kafka Connect connectors so you don't have to run your own Connect workers.
22. Level 22: OpenSearch Service Managed Domains
    Work through the nine parts for OpenSearch Service managed domains: sizing data and dedicated master nodes, shard and storage limits, UltraWarm and cold storage tiers, and the cost levers of instance hours and storage.
23. Level 23: OpenSearch Serverless Collections
    Learn how OpenSearch Serverless collections bill in OpenSearch Compute Units (OCUs), how search, time-series and vector collection types differ, and where Serverless costs more or does less than a managed domain.
24. Level 24: Quick Sight BI (formerly QuickSight)
    Work through the nine parts for AWS's BI service: QuickSight became the Quick Sight part of Amazon Quick Suite in October 2025 and has since been shortened to Amazon Quick. Covers SPICE versus direct query, per-user and capacity pricing, embedding, and row-level security.
25. Level 25: Lakehouse Design with Open Table Formats
    Design a lakehouse on S3 by comparing Apache Iceberg, Delta Lake and Hudi on how each AWS engine supports them, how compaction and snapshot expiry work, and the choice between S3 Tables and self-managed Iceberg.
26. Level 26: Governance and Lineage Across the Stack
    Combine Lake Formation, the Glue Data Catalog and DataZone/SageMaker Unified Studio into one governance design that covers fine-grained access, cross-account sharing, data quality and end-to-end lineage.
27. Level 27: Streaming Architecture Design
    Design end-to-end streaming systems: choosing between Kinesis and MSK for ingestion, Flink or Firehose for processing, exactly-once versus at-least-once delivery, ordering and partition keys, late data and replay.
28. Level 28: Worked Cost: A Reporting Workload
    Cost one batch reporting workload end to end (ingest, Glue ETL, storage, a warehouse or query engine, BI) under three designs: Redshift provisioned, Redshift Serverless, and Athena on Iceberg.
29. Level 29: Worked Cost: A Streaming Workload
    Cost one real-time pipeline (ingest, processing, delivery to a lake and to search) under a Kinesis-based design and an MSK-based design, including the cross-AZ data transfer and retention charges that are easy to miss.
30. Level 30: Building a Weighted Selection Scorecard
    Turn the standard analytics profile into a weighted scorecard whose criteria, weights and 1–5 scoring scales can be defended and repeated by other people.
31. Level 31: Applying the Scorecard Across AWS Services
    Use the scorecard to rank candidate AWS services and combinations of services for one profile, using the limits, costs and operating models from earlier lessons as evidence for each score.
32. Level 32: Gathering Dated Evidence for Another Cloud
    Collect comparable, dated evidence for the equivalent analytics services on a second cloud (for example Google Cloud's BigQuery, Dataflow and Pub/Sub, or Azure's Fabric and Event Hubs): their limits, pricing units and operating models.
33. Level 33: Scoring AWS Against Another Cloud
    Apply the same weighted scorecard to AWS and to the second cloud for one profile, accounting for migration, egress, skills and ecosystem factors beyond the core service features.
34. Level 34: The Selection Record and Launch Plan
    Write a one-page selection record (an architecture decision record): context, profile, options, scores, decision, consequences, review date. Pair it with a staged launch plan that includes guardrails and a clean-up plan.
35. Level 35: Final Challenge: The Decision Workshop Exam
    Run a full decision workshop for a new scenario that draws on the whole course: build the profile, shortlist services, cost the options, score them on AWS and on another cloud, and produce a defended one-page selection record.

Access

The first 2 levels are free with a free account. Every level, the podcast edition and the AI tutor come with All Access at £4.99/month or any Creator plan — see pricing.

Start free →