2.5 Practice Drills and Readiness Markers

Supported Cloud Platforms and Regions

Snowflake is cloud-agnostic and runs natively on three providers:

Each Snowflake account is hosted in exactly one cloud platform and one region (for example, AWS us-east-1). The platform behaves the same across clouds — the SQL and features are identical — but data does not automatically move between clouds or regions. To span clouds or regions you use Database Replication and Cross-Cloud / Cross-Region Data Sharing, and you can fail over between regions with database failover (Business Critical and above). Choosing a region matters for latency, data residency, and compliance.

The AI Data Cloud

The AI Data Cloud is Snowflake's vision and branding for a global network of Snowflake accounts that can share and collaborate on data without copying or moving it. Because storage is centralized and access is governed by the cloud services layer, a provider can grant a consumer live, read-only access to data that always reflects the latest values — no ETL, no duplication, no stale copies. The "AI" emphasis reflects native capabilities like Cortex (LLM and ML functions), document AI, and vector search built directly on governed data.

Secure Data Sharing and the Marketplace

Secure Data Sharing is the mechanism behind the AI Data Cloud. A provider creates a share (an object that grants access to specific databases, schemas, and tables) and a consumer account accesses it as a read-only database. Crucially:

• No data is copied — the consumer queries the provider's storage directly through Snowflake's services layer.
• The consumer uses their own virtual warehouse (and pays their own compute) to query shared data.
• Updates by the provider are immediately visible to consumers.

The Snowflake Marketplace is the public storefront built on this capability. Providers publish listings — free, paid, or personalized data sets, plus Native Apps — that consumers can discover and access live in their own account. A Reader Account lets a provider share data with a party that is not itself a Snowflake customer; the provider manages and pays for that reader account's compute.

Interfaces, Connectors, and Drivers

Snowflake is accessed through several interfaces the exam expects you to recognize:

• Snowsight — The modern web UI for running queries, building dashboards, monitoring warehouses, and administration (it replaced the Classic Console).
• SnowSQL — The command-line client for scripting and bulk operations.
• Drivers — JDBC and ODBC for standard tool connectivity.
• Connectors — Native connectors for Python, Spark, Kafka, Node.js, .NET, Go, and PHP.
• Snowpark — A developer framework to run Python, Java, and Scala code (DataFrames, UDFs, stored procedures) inside Snowflake's compute.
• REST API / Snowpipe — Programmatic access and continuous, serverless data ingestion.

Readiness Markers for This Domain

Because this domain is ~31% of the exam, drill until any scenario maps instantly to the right concept:

You are ready when, after a day away, you can read a feature scenario with the domain label hidden and still name the correct layer, edition, cache, or sharing mechanism and justify why the distractors fail.

Replication, Failover, and Cross-Cloud Reach

The AI Data Cloud spans clouds and regions through a few related features that are easy to confuse:

• Database Replication copies a database to another account, region, or cloud, kept in sync on a schedule — useful for read scaling and migration.
• Failover/Failback (Business Critical+) promotes a replica to primary for disaster recovery.
• Cross-Cloud / Cross-Region Sharing lets a consumer in one cloud access a provider's data in another; under the hood this uses replication because storage is region-bound, so the consumer reads a replicated copy rather than the original.

The takeaway: within one region/cloud, sharing is no-copy and live; across regions or clouds, Snowflake replicates data so it is local to the consumer, adding storage and a sync interval.

Native Apps and Cortex on Governed Data

The "AI" in AI Data Cloud is concrete. Snowflake Cortex exposes LLM functions (summarize, translate, sentiment, complete) and ML functions that run on data without it leaving Snowflake's governance boundary. The Native App Framework lets providers package application logic with data and distribute it through the Marketplace, where it runs in the consumer's account. Document AI extracts structured fields from unstructured documents, and the VECTOR data type plus vector functions support similarity search for retrieval workloads.

You do not need deep AI knowledge for SnowPro Core, but you should recognize that these capabilities are built into the platform and operate on the same governed storage — the "bring compute and AI to the data" theme that defines this domain.

Final Drill Pattern

For each readiness marker, write a two-column sheet: left, a one-line scenario ("identical query reruns instantly, no credits"); right, the exact concept ("result cache, 24h, cloud services"). Self-test on shuffle until every scenario maps without hesitation — that recall is the sign this 31% domain is exam-ready.