Examkin

Question

Difficulty: HardData Integration and Analytical Storage

A smart grid utility provider is designing an analytical data platform on Azure to monitor electricity consumption. The platform must ingest real-time telemetry from 500,000 smart meters at a velocity of 10,000 events per second, totaling 12 TB of new data monthly. The architecture must satisfy the following requirements:
- The raw telemetry data must be stored in a data lake and remain highly available and resilient to regional datacenter outages.
- Data analysts must perform ad-hoc, exploratory SQL queries on years of historical Parquet files in the data lake without provisioning or paying for idle compute resources.
- External partners must be granted read access to the historical files securely, with the ability to revoke access immediately if needed.

Which two options should you recommend in the architectural design to meet these requirements?

  1. Azure Data Lake Storage Gen2 configured with Geo-Redundant Storage (GRS) to store the raw telemetry data.Answer
  2. A Serverless SQL pool in Azure Synapse Analytics to perform the ad-hoc exploratory queries on the historical Parquet files.Answer
  3. C
    A Dedicated SQL pool in Azure Synapse Analytics to query the historical Parquet files.
  4. D
    Azure Data Lake Storage Gen2 configured with Locally Redundant Storage (LRS) to store the raw telemetry data.
  5. E
    An ad-hoc Shared Access Signature (SAS) token configured with a two-year lifetime directly in the URI for client data access.

Answer

Configure Azure Data Lake Storage Gen2 with Geo-Redundant Storage (GRS) to ensure regional resilience, and use a Serverless SQL pool in Azure Synapse Analytics to query historical Parquet files on-demand without provisioning idle resources.
Configuring Azure Data Lake Storage Gen2 with Geo-Redundant Storage (GRS) satisfies the requirement for regional disaster resilience by replicating data to a secondary region. Implementing a Serverless SQL pool in Azure Synapse Analytics satisfies the requirement to run ad-hoc, exploratory SQL queries on Parquet files without provisioning or paying for idle compute, as it uses a pay-per-query model based on the volume of data processed.

Step-by-Step Solution

1
Analyze the storage redundancy requirement.
Locally Redundant Storage (LRS) only protects against local hardware failures within a single datacenter. To survive a regional outage, Geo-Redundant Storage (GRS) must be chosen.
Ensures that data is replicated to a secondary paired region for disaster recovery.
2
Analyze the query engine requirement.
A Dedicated SQL pool is a provisioned resource that incurs hourly compute costs regardless of active querying. A Serverless SQL pool charges only per query based on data processed, making it ideal for ad-hoc, intermittent exploration.
Satisfies the cost-efficiency constraint of avoiding payment for idle compute resources.
3
Verify security and access control practices.
Providing a long-lived ad-hoc SAS token is insecure because it cannot be revoked without rotating storage keys, which affects all other applications using those keys. Best practice dictates using stored access policies or short-lived tokens.
Ensures the ability to revoke external access immediately.

Key Concept

Selecting optimal Azure storage redundancy and Synapse SQL compute types based on availability and cost requirements.
Rate this question

Topics

Data Integration and Analytical Storage

More questions from this topic