Examkin

Question

Difficulty: HardAzure Cosmos DB and NoSQL Solutions

You are designing a globally distributed data storage solution for a delivery tracking platform. The solution has the following requirements:
- A NoSQL database must store real-time package delivery telemetry. The telemetry ingestion is write-heavy (70%70\% writes, 30%30\% reads) with an expected load of millions of updates daily.
- The telemetry database must support a 99.999%99.999\% read and write availability SLA and sub-10 ms10\text{ ms} latency.
- Delivery confirmation images must be stored in object storage. External delivery agents must be granted temporary access to upload confirmation images directly to the storage.
- The entire solution must be resilient to regional datacenter outages.

Which of the following configurations should you select to meet these requirements? (Select TWO.)

  1. Configure Azure Cosmos DB with multi-region writes enabled across the target regions, and use OrderId as the partition key for the telemetry container.Answer
  2. Configure Azure Blob Storage with Geo-Zone-Redundant Storage (GZRS) and generate short-lived Shared Access Signature (SAS) tokens associated with a Stored Access Policy for image uploads.Answer
  3. C
    Configure Azure Cosmos DB with single-region writes, and use StoreId as the partition key for the telemetry container.
  4. D
    Configure Azure Blob Storage with Locally Redundant Storage (LRS) and generate long-lived Shared Access Signature (SAS) tokens with a one-year lifespan directly in the URI.

Answer

Configure Azure Cosmos DB with multi-region writes enabled across the target regions, and use OrderId as the partition key for the telemetry container; and configure Azure Blob Storage with Geo-Zone-Redundant Storage (GZRS) and generate short-lived Shared Access Signature (SAS) tokens associated with a Stored Access Policy for image uploads.
Enabling multi-region writes for Azure Cosmos DB is required to meet the 99.999%99.999\% write availability SLA, and using a high-cardinality partition key like OrderId prevents hot partitions under write-heavy workloads. For Blob Storage, using Geo-Zone-Redundant Storage (GZRS) ensures data is resilient against regional datacenter outages, and using short-lived SAS tokens with stored access policies aligns with security best practices.

Step-by-Step Solution

1
Analyze Azure Cosmos DB availability and latency requirements.
To achieve a 99.999%99.999\% write availability SLA, multi-region writes (multi-master) must be enabled. Single-region writes only offer up to 99.99%99.99\% write availability.
Establishing high write availability requires active-active write regions.
2
Determine the partition key for the write-heavy Cosmos DB container.
A high-cardinality key such as OrderId must be selected. Using a low-cardinality key like StoreId or DeliveryStatus would result in uneven partition sizes and hot physical partitions.
Ensuring horizontal scalability and preventing partition throughput limits.
3
Select the redundancy tier for the delivery confirmation images in Azure Blob Storage.
Geo-Zone-Redundant Storage (GZRS) is required because it protects against regional outages by replicating data across zones in the primary region and geo-replicating it to a secondary region. Locally Redundant Storage (LRS) is vulnerable to regional outages.
Meeting the regional disaster resilience requirement.
4
Define the secure access mechanism for external agents uploading images.
Short-lived Shared Access Signature (SAS) tokens associated with a Stored Access Policy should be used. Long-lived SAS tokens defined directly in the URI are insecure and difficult to revoke.
Ensuring secure, delegable access that adheres to the principle of least privilege.

Key Concept

Azure Cosmos DB partition key selection, multi-region write scalability, and Azure Blob Storage disaster resilience and security.
Rate this question

Topics

Azure Cosmos DB and NoSQL Solutions

More questions from this topic