Path 4 · Module 1
Design a data storage solution for non-relational data
Read the module on Microsoft Learn, then come back here: test yourself with the quiz and apply it in the design exercise. Write down what you want to remember with the ✎ Note buttons.
Step 1 · Read the module
Open Design a data storage solution for non-relational data on Microsoft Learn. The units:
- Introduction
- Design for data storage
- Design for Azure storage accounts
- Design for data redundancy
- Design for Azure Blob Storage
- Design for Azure Files
- Design for Azure managed disks
- Design for storage security
- Module assessment
- Summary and resources
Close the Learn tab. Answering from memory is what makes it stick; looking it up just measures how fast you can scroll.
Quiz
21 exam-style questions, roughly in the order of the units. Click an answer to see the explanation. Your best score is saved and shown on the path page; use Retry at the end to go again.
1. Images, videos and log files that don't follow a schema are which kind of data?
C. Unstructured data has no defined model; blob storage is the usual home. JSON and XML are semi-structured.
2. Which storage account type is recommended for most scenarios, supporting blobs, files, queues and tables?
C. GPv2 supports all services, access tiers and redundancy options.
3. An app needs consistently low latency for small, frequent blob writes. Which account type?
A. Premium block blob accounts use SSDs for high transaction rates and low latency.
4. Why would you create several storage accounts instead of one?
B. Settings like redundancy and networking apply to the whole account, so data with different needs goes in different accounts.
5. Which redundancy keeps three copies in one datacenter only?
A. LRS protects against drive and rack failures, not datacenter loss.
6. Data must survive a zone failure and stay in the same region (data residency). Which redundancy?
B. ZRS replicates synchronously across three zones in one region.
7. Data must survive a zone failure and a region failure, and reads must be possible from the secondary during an outage. Which option?
B. GZRS = ZRS in the primary + LRS in the secondary region; RA adds read access to the secondary.
8. With GRS (not RA-GRS), when can you read data in the secondary region?
B. GRS replicates but the secondary isn't readable until you (or Microsoft) fail over. RA-GRS makes it readable.
9. Log files are written daily, read for a month, then kept for 7 years for compliance and almost never read. Cheapest design?
D. Lifecycle policies move blobs between tiers automatically based on age or last access.
10. What must you do before reading a blob in the Archive tier?
D. Archive is offline storage; rehydration takes up to 15 hours (standard priority).
11. Records must be kept unchangeable for 5 years for a financial regulator (WORM). What do you configure?
A. Immutable storage (WORM) blocks modification and deletion for the retention period.
12. Big-data analytics needs a hierarchical namespace with directory-level ACLs on blob storage. What do you enable?
D. ADLS Gen2 is a storage account with the hierarchical namespace enabled.
13. Windows file servers on-premises must keep serving local users while files are also stored in Azure. What do you design?
A. File Sync turns Windows Servers into caches of an Azure file share; cloud tiering keeps hot files local.
14. Linux workloads need an NFS file share with low latency. Which Azure Files option?
C. NFS on Azure Files is available on premium (FileStorage) accounts.
15. Users must access an Azure file share with their on-premises AD credentials and NTFS permissions. What's required?
D. Azure Files supports AD DS, Entra Domain Services and Entra Kerberos for SMB, with share- and file-level permissions.
16. A production SQL Server VM needs sub-millisecond latency and up to 160,000 IOPS on one disk. Which disk type?
C. Ultra Disk goes up to 400,000 IOPS per disk with sub-millisecond latency. Premium SSD v2 (up to 80,000 IOPS) is the cheaper step below it.
17. Which disk type is cheapest for dev/test VMs and infrequently accessed data?
C. Standard HDD has the lowest cost but variable performance.
18. Disks must be encrypted with keys the company controls and can rotate. What do you use?
B. Disk encryption sets point managed disks at a customer-managed key.
19. Clients need temporary, limited access to one container without the account key. What do you give them?
D. A user delegation SAS is scoped and time-limited and is the most secure SAS type.
20. A storage account must only be reachable from a VNet over a private IP, with public access disabled. What do you configure?
B. A private endpoint gives the account a private IP in your VNet; service endpoints still use the public endpoint.
21. Which setting should be disabled so apps must use Entra ID instead of keys?
D. Disabling Shared Key forces Entra ID authorization, which supports RBAC and auditing.
Design exercise
On paper · about 25 minutes
Northgate Media produces video content.
- Raw footage (20 TB/month) is edited for 2 weeks, then rarely touched; it must be kept 10 years.
- Published videos are streamed worldwide and must stay readable if the primary region goes down.
- Editors in the office use a Windows file server; the company wants to drop its local storage arrays but keep fast access.
- A rendering cluster of Linux VMs needs a shared NFS file system with high throughput.
- Legal contracts must be stored unchangeable for 7 years.
-
Task
Choose storage services and accounts
For each data set, choose the service, the account type and the redundancy.
Done when- Each data set has service, account type and redundancy.
- Accounts are split where settings differ.
Model answer
- Raw footage: GPv2, blob, LRS or ZRS (cost), lifecycle tiers.
- Published videos: GPv2, blob, RA-GZRS, fronted by a CDN/Front Door.
- Office files: Azure Files (standard, ZRS) + Azure File Sync with cloud tiering.
- Rendering: Premium file shares with NFS (or Azure NetApp Files).
- Contracts: separate GPv2 account with immutable storage.
-
Task
Write the lifecycle and retention rules
Write the rules for raw footage and contracts.
Done when- Tier moves have ages.
- The 10-year and 7-year rules are met.
Model answer
Raw footage: Hot → Cool after 14 days → Archive after 60 days (or Cold) → delete after 10 years. Contracts: time-based immutability policy of 7 years (locked), versioning, soft delete.
-
Task
Secure the accounts
List the security settings for all accounts.
Done when- Network access and authentication are covered.
- Encryption and data protection are covered.
Model answer
Private endpoints for internal accounts, public access disabled where possible; Shared Key disabled, Entra ID + RBAC; user delegation SAS for partners; secure transfer + TLS 1.2; CMK in Key Vault if required; soft delete and versioning; Defender for Storage.