Microsoft Certified: Azure AI Engineer AssociatePlan and manage an Azure AI solutionHard
A media company is developing an Azure AI solution to automatically generate captions for video content using Azure Video Indexer. They anticipate varying workloads throughout the day but need to ensure predictable performance for high-priority, time-sensitive content. The solution architects want to optimize costs while guaranteeing service levels for critical tasks. Which pricing model and capacity allocation strategy should they implement?
- AA hybrid approach using pay-as-you-go for standard content and dedicated capacity for high-priority content.
- BPay-as-you-go for all Video Indexer operations.
- CDedicated capacity for all Video Indexer operations.
- DUtilize the Free tier for all operations and scale up manually as needed.
Show answer & explanationAnswer & explanation
Correct answer: A. A hybrid approach using pay-as-you-go for standard content and dedicated capacity for high-priority content.
A hybrid approach combines the cost-effectiveness of pay-as-you-go for variable, less critical workloads with the predictable performance and guaranteed throughput of dedicated capacity for high-priority, time-sensitive content. This optimizes both cost and service levels.
Why the other options are wrong
- B. Pay-as-you-go alone might not guarantee predictable performance during peak times for high-priority content.
- C. Dedicated capacity for all operations would be expensive and inefficient for varying workloads, leading to over-provisioning during low-demand periods.
- D. The Free tier is not suitable for production workloads and cannot guarantee service levels for any content, let alone high-priority tasks.
Hybrid Azure AI Capacity Strategy
Combining pay-as-you-go pricing for variable workloads with dedicated capacity for critical, predictable workloads to optimize cost and performance.
- Leverages the flexibility of pay-as-you-go.
- Guarantees performance for high-priority tasks with dedicated capacity.
- Optimizes cost by avoiding over-provisioning for all traffic.
Memory trick: Hybrid capacity balances cost and critical performance.