Plan and manage an Azure AI solution 1.2 Set up AI solutions in Foundry
A European insurer must keep model inferencing within the EU but wants pay-per-token…
Study guide skill: Choose appropriate deployment options
Question B1-002
A European insurer must keep model inferencing within the EU but wants pay-per-token pricing and more capacity than a single region can offer. Which deployment type should you use for the model?
Select one answer.
Answer: C
Data Zone deployment types process data only within the Microsoft-defined data zone (US, EU or Asia Pacific), and Data Zone Standard is billed per token. Global types can process prompts in any region, provisioned types reserve capacity rather than pay per token, and Developer is meant for evaluating fine-tuned models.
Deep dive
How it works
When you deploy a model in Foundry, the deployment type decides three things: where your data is processed, how you pay (pay-per-token or reserved capacity), and the performance you get. There are three main categories: standard (pay-per-token), provisioned (reserved capacity) and batch (discounted asynchronous processing). Within standard and provisioned, you then pick global, data zone or geography-based processing to match your compliance needs.
Data Zone types process data only within the Microsoft-specified data zone: US, EU or Asia Pacific. The EU zone follows the Azure EU Data Boundary. Inside a zone, processing isn't tied to one region; for the US zone, for example, the service processes data anywhere within the US. Data Zone Standard is billed per token, and the docs list it as the fit for EU, US or APAC data zone compliance. That's the combination in the question: EU-only inferencing with pay-per-token billing, processed across the EU zone rather than in one region. Data stored at rest stays in the designated Azure geography for every deployment type.
The other options, and when you'd use them
- Global Provisioned: Global types may process data in any Azure region, and provisioned types bill for reserved PTUs. Use it for predictable high throughput when data location isn't restricted.
- Global Standard: This is the default starting point. It has the lowest price and broadest region coverage, but data may be processed in any Azure region, so it doesn't meet an EU-only rule.
- Developer: Developer deployments are for fine-tuned model evaluation only. They have a 24-hour lifetime and no SLA or data-residency guarantee.
Exam tip
Read the residency requirement first. "Anywhere" means Global, "within the EU, US or APAC zone" means Data Zone, and "within one Azure geography" means Standard or Regional Provisioned. Then pick Standard for pay-per-token or Provisioned for reserved capacity.
References
Last reviewed: