What's changed: Deepened AZ-305 Chapter 2 (added comparison tables, scenarios, FAQs, exam traps, deep-dive paragraphs to each section; localized figures to Japanese)
2.3Designing Data Integration, Analytics, and Protection
Understand using and protecting data—Data Lake Storage, Synapse Analytics, Data Factory, caching (Azure Cache for Redis), and encryption (at rest/in transit). Integrate/analyze data and keep it secure.
Turn data into value via integration and analytics, and protect it with encryption. Data Factory ingests; Synapse analyzes.
2.3.1Integration, analytics, and protection
- Data Factory: orchestrate ETL/ELT to ingest and transform data.
- Data Lake Storage Gen2: large-scale hierarchical data lake (on Blob).
- Synapse Analytics: unified data warehouse/big-data analytics.
- Redis cache/encryption: Azure Cache for Redis for low latency, protected by at-rest encryption + TLS.
Common on AZ-305: ingest/ETL orchestration = Data Factory, data lake = Data Lake Storage Gen2, DWH/big-data analytics = Synapse Analytics, low-latency cache = Azure Cache for Redis, at-rest key management = Key Vault (customer-managed keys). Storage is encrypted at rest by default.
AZ-305 data-utilization design centers on "connecting ingest → store → analyze → serve with services that fit requirements." Azure Data Factory orchestrates ETL/ELT from diverse sources in pipelines, with codeless mapping data flows and schedule/event triggers (combine with Stream Analytics or Event Hubs for real time). Storage standardizes on Data Lake Storage Gen2 (hierarchical namespace on Blob, ideal for large-scale analytics), and Synapse Analytics unifies dedicated SQL pools (DWH), serverless SQL, and Spark—with Microsoft Fabric (OneLake/Direct Lake) now also an option. For performance, Azure Cache for Redis caches hot data to cut DB load and latency, combined with Front Door/CDN to accelerate delivery. Protection is layered: Storage is encrypted at rest (SSE) by default, with Key Vault customer-managed keys (CMK) for stricter control, TLS in transit, private endpoints to limit exposure, and RBAC + managed identities for access. Classify and protect sensitive data with Microsoft Purview governance. The key is to build a pipeline and protection that are neither over- nor under-engineered, considering batch vs stream, scale, latency, cost, and regulation.
| Stage | Service | Key point |
|---|---|---|
| Ingest/transform | Data Factory | ETL/ELT orchestration; triggers |
| Store (data lake) | Data Lake Storage Gen2 | Hierarchical namespace; analytics-scale |
| Analyze (DWH/big data) | Synapse Analytics | SQL pool/serverless/Spark unified |
| Speed/protect | Redis; SSE/CMK; TLS | Cache + layered encryption + private endpoints |
Scenario: Ingest on-prem DB and SaaS data nightly, aggregate into a data lake, and run BI analytics. Sensitive data requires self-managed encryption keys. → Schedule ingestion with Data Factory pipelines into Data Lake Storage Gen2, analyze with Synapse Analytics. Encrypt Storage with Key Vault customer-managed keys (CMK), access via managed identity + RBAC, and keep traffic private with private endpoints.
FAQ: At-rest encryption is default—why use Key Vault CMK? Azure Storage is encrypted by default with platform-managed keys (PMK), sufficient for many requirements. When regulation/compliance requires self-controlling key lifecycle, revocation, and audit, switch to Key Vault customer-managed keys (CMK). The difference is "who manages the keys."
Exam trap: Handling real-time stream processing with Data Factory alone is a weak design. Data Factory is mainly batch/scheduled ETL orchestration; low-latency streams suit Stream Analytics / Event Hubs. Also, don’t solve "analytics speed-up" via a storage-tier choice—the cache layer is Azure Cache for Redis.
2.3.2Section summary
- Integrate/analyze = Data Factory + Data Lake Gen2 + Synapse
- Speed = Redis / protect = at-rest encryption + Key Vault + TLS
Sign in to track progress — Log in.
Quick check
(just a quick review)Q1. You want to orchestrate pipelines that ingest/transform data from diverse sources. What?
Q2. Which analytics platform unifies data warehousing and big-data analytics?
Q3. You want to cache hot data in front of a database to cut read latency. What?