An online multiplayer gaming platform is designing the compute infrastructure on Azure. The architecture must support two workloads:
| Workload | Type | Scaling Requirement | Eviction Tolerance | CPU-to-Memory Ratio |
|---|---|---|---|---|
| Workload 1 (Session Hosting) | Stateless | Dynamic scaling | Can tolerate unexpected evictions | At least (e.g., per vCPU) |
| Workload 2 (Matchmaking Engine) | Stateful | Steady-state | Cannot tolerate unexpected evictions | At least (e.g., per vCPU) |
Which two configuration options should you select to meet these requirements?
- For Workload 1, deploy a Virtual Machine Scale Set using Azure Spot VMs with a Dsv5-series VM size.Cevap
- For Workload 2, deploy Azure Virtual Machines using an Esv5-series VM size.Cevap
- CFor Workload 2, deploy Azure Spot VMs using an Esv5-series VM size.
- DFor Workload 2, deploy Azure Virtual Machines within a single availability zone using an Availability Set to provide zone-redundant high availability.
Cevap
Deploy a Virtual Machine Scale Set using Azure Spot VMs with a Dsv5-series size for Workload 1, and deploy Azure Virtual Machines using an Esv5-series size for Workload 2.
Deploying the session hosting layer on a Virtual Machine Scale Set using Azure Spot VMs of the Dsv5-series meets the requirements of cost savings, eviction tolerance, and the CPU-to-memory ratio. Deploying the matchmaking database on standard Azure Virtual Machines of the Esv5-series satisfies the steady-state SLA requirement (no unexpected evictions) and the memory-optimized ratio.
Adım Adım Çözüm
Anahtar Kavram
Selecting VM sizes based on memory-to-CPU ratios, and choosing between standard VMs and Spot VMs based on workload eviction tolerance and SLA constraints.