Azure Storage: Blob, Data Lake, Files & Tables

What type of data is Azure Blob Storage designed to store?

• AMassive amounts of unstructured data as binary large objects✓
• BGraph data made up of vertices and edges
• CKey-value pairs organised into partitions and rows
• DStructured data organised into strictly enforced relational tables

Answer: Azure Blob Storage stores massive amounts of unstructured data as binary large objects, or blobs, in the cloud. Applications read and write these blobs through the Blob storage API rather than through structured tables or rows. Relational tables, partitioned key-value rows, and graph structures describe other Azure data stores, not Blob Storage.

A team needs to store large binary objects that change infrequently, with an individual blob potentially approaching 190 TiB in size. Which Azure Blob Storage blob type should they use?

• AAzure Queue Storage
• BBlock blobs✓
• CPage blobs
• DAppend blobs

Answer: Block blobs are handled as a set of blocks and can hold up to roughly 190.7 TiB, making them best suited to large, discrete binary objects that change infrequently. Page blobs are organised into fixed 512-byte pages and are optimised for random read and write access such as virtual machine disks. Append blobs only support adding data to the end of the blob, and Azure Queue Storage is a separate messaging service, not a blob type.

Which Azure Blob Storage blob type is organised as a collection of fixed 512-byte pages and is used by Azure to implement virtual machine disk storage?

• AAppend blobs
• BAzure Files shares
• CPage blobs✓
• DBlock blobs

Answer: Page blobs are organised as a collection of fixed 512-byte pages and are optimised to support random read and write operations, which is why Azure uses them to implement virtual machine disk storage. Block blobs are built from variable-size blocks better suited to large objects that change infrequently, and append blobs only support adding data to the end of the blob. Azure Files shares are a separate file-storage service, not a blob type.

Which Azure Blob Storage access tier is used by default for blobs that are accessed frequently?

• ACool
• BArchive
• CCold
• DHot✓

Answer: The Hot tier is the default access tier and stores blob data on high-performance media for blobs accessed frequently. The Cool and Cold tiers cost less to store but more to access, and are intended for infrequently accessed data. The Archive tier has the lowest storage cost but the highest access latency.

A blob has been moved into the Archive access tier. What must happen before an application can read that blob again?

• AIts access tier must be changed to Hot, Cool, or Cold so the blob can be rehydrated✓
• BA lifecycle management policy must be deleted from the storage account
• CThe blob must be recreated as an append blob, then moved back into the same container
• DThe storage account's redundancy option must be changed to GZRS

Answer: Blobs in the Archive tier are stored in an offline state, so before one can be read its access tier must be changed to Hot, Cool, or Cold and the blob rehydrated, a process that can take up to 15 hours. Recreating the blob as a different blob type does not rehydrate archived data, and neither changing the redundancy option nor deleting a lifecycle policy affects whether an archived blob can be read.

What does Azure Data Lake Storage Gen2 add on top of Azure Blob Storage's scalability and storage tiers?

• AA dedicated messaging queue for streaming telemetry
• BA hierarchical file system with POSIX-compliant access control lists✓
• CA relational schema with foreign keys and stored procedures
• DA built-in Power BI reporting workspace

Answer: Azure Data Lake Storage Gen2 combines the scalability and cost control of Azure Blob Storage with a hierarchical file system, whose hierarchical namespace enables POSIX-compliant access control lists for fine-grained file and folder permissions. It does not add relational features such as foreign keys and stored procedures, a messaging queue, or a built-in Power BI workspace.

A company upgrades an existing Azure Storage account to enable the hierarchical namespace needed for Azure Data Lake Storage Gen2. What happens if the company later wants to go back to a flat namespace?

• AThe account reverts automatically after 90 days
• BThe upgrade cannot be reversed, so the account stays on a hierarchical namespace✓
• CAn administrator can revert it through Azure role-based access control
• DThe account can be reverted at any time from the Azure portal

Answer: Enabling the hierarchical namespace is a one-way upgrade, so once an Azure Storage account has been upgraded it cannot be reverted to a flat namespace. There's no automatic reversion after a fixed period, and neither Azure role-based access control nor the Azure portal offers a way to undo the upgrade.

What is Microsoft OneLake?

• AA single, unified, logical data lake for an entire organisation✓
• BA relational database engine, built into Microsoft Fabric, used for OLTP workloads
• CA stand-alone blob access tier for archived data
• DA dedicated Cosmos DB API for graph data

Answer: OneLake is a single, unified, logical data lake designed for an entire organisation, provisioned automatically with every Microsoft Fabric tenant and serving as the central repository for all analytics data. It's not a relational database engine, a Cosmos DB API, or a blob access tier.

What underlying file format does OneLake use to store data, built on top of Azure Data Lake Storage Gen2?

• AAvro with a relational schema
• BDelta Parquet✓
• CCSV with fixed-width columns
• DBSON

Answer: OneLake is built on Azure Data Lake Storage Gen2 and stores data in Delta Parquet format, an open, efficient file format widely used for analytics data. BSON is a binary JSON format used elsewhere in Azure, and fixed-width CSV or relational-schema Avro are not the format OneLake uses.

What is Azure Files primarily used for?

• ACreating cloud-based network file shares that multiple users can access✓
• BRunning graph queries over vertices and edges
• CStoring key-value rows partitioned by a partition key
• DHosting virtual machine disk images as page blobs

Answer: Azure Files lets organisations create cloud-based network shares, similar to on-premises file shares, so documents and other files can be made available to multiple users without the hardware and maintenance costs of local servers. Graph queries, partitioned key-value rows, and virtual machine disk images describe other Azure data services, not Azure Files.

A Linux-based application needs to mount an Azure file share using the NFS protocol. What does the unit say is required?

• AAn HDD-tier storage account and Microsoft Entra ID authentication
• BThe Azure File Sync agent installed on the Linux host
• CA hierarchical namespace and POSIX-compliant access control lists
• DAn SSD-tier storage account and a virtual network through which access is configured✓

Answer: NFS Azure file shares require an SSD-tier storage account, and you must create and configure a virtual network through which access to the share is controlled; NFS is supported on Linux kernel 4.3 or later but not on Windows or macOS. An HDD-tier account, a hierarchical namespace, and the Azure File Sync agent are not what the unit describes as NFS requirements.

What does the Azure File Sync service do?

• ASynchronises locally cached copies of shared files with the data in Azure File Storage✓
• BReplicates blob data asynchronously to a secondary Azure region
• CAutomatically moves ageing files between Blob Storage access tiers
• DConverts append blobs into block blobs for archival storage

Answer: Azure File Sync lets you synchronise locally cached copies of shared files with the data held in Azure File Storage, so on-premises users keep fast local access while the authoritative copy stays in the cloud. Replicating blob data to a secondary region, converting blob types, and moving files between access tiers are unrelated Azure Storage features.

What two-part key does every row in an Azure Table Storage table require?

• AA primary key and a foreign key
• BA partition key and a row key✓
• CA container name and a partition key
• DA row key and a timestamp

Answer: Every row in an Azure Table Storage table has a unique key made up of a partition key, which identifies the partition containing the row, and a row key, which is unique within that partition; a timestamp column is recorded separately whenever a row is modified. Primary and foreign keys are relational database concepts that Azure Table Storage tables don't have, and a container name isn't part of the row key.

For a new workload that needs key-value storage at scale, which service does the unit recommend over Azure Table Storage?

• AAzure Cosmos DB for Table✓
• BAzure Data Lake Storage Gen2
• CAzure Files with the SMB protocol
• DAzure Blob Storage with the Archive tier

Answer: Azure Cosmos DB for Table stores data in the same key/value format as Azure Table Storage but adds greater performance and global availability, so it's the recommended choice for new workloads. Azure Data Lake Storage Gen2, Azure Files, and Azure Blob Storage's Archive tier serve different purposes and aren't presented as alternatives to Azure Table Storage.