Presented by Yi Chai, Varun Shandilya, and Amit Srivastava at Microsoft Ignite 2025, this session explores Azure IaaS optimization for high performance and global scalability.

Azure IaaS Best Practices to Enhance Performance and Scale

This advanced (level 300) session from Microsoft Ignite 2025 dives into proven techniques for boosting performance and scalability on Azure Infrastructure as a Service (IaaS).

Speakers: Yi Chai, Varun Shandilya, Amit Srivastava

Session Overview

Azure IaaS enables scalable application deployments by leveraging features like high-throughput storage, low-latency networking, and intelligent auto-scaling. The session covers:

• Optimizing compute, storage, and networking resources
• Performance tuning strategies
• Cost reduction and automation methods

Key Topics and Enhancements

• Azure Boost: Performance enhancements and faster VM operations
• Compute Fleet: Efficient workload management across multiple VM types and sizes
• Azure Virtual Machines: Latest automation and schedule actions for cost control
• Storage Solutions: Latest with Azure Ultra Disk, improved latency and IOPS, distributed storage solutions for AI inferencing
• Networking: Introduction to QUIC and HTTP/3 support for better throughput and lower latency
• Application Gateway for Containers: Demo of load-aware routing, enabling intelligent traffic management in containerized workloads
• Hyper-V Advancements: Direct virtualization for improved performance gains

Demos and Use Cases

• Load-aware routing in Application Gateway for Containers
• VM automation scheduling to optimize for cost and resource needs
• Distributed storage approaches for AI and highly concurrent workloads
• Relaunch of Azure Ultra Disk with improved capabilities

Learning Resources

• Migration and modernization plans
• Microsoft Ignite 2025 sessions on-demand

Chapters

• Session Introduction and Overview
• Deep Dive into Networking: Ingress and Product Suite
• New Networking Enhancements: QUIC and HTTP/3
• Demo: Application Gateway for Containers
• Compute Optimization and Fleet Management
• VM Automation: Schedule Actions for Cost Optimization
• Hyper-V: Direct Virtualization Improvements
• AI Inferencing and Distributed Storage
• Ultra Disk Relaunch & Performance

Best Practice Summary

• Right-size VMs for workload needs
• Use Azure Boost for intensive compute
• Adopt QUIC/HTTP3 for network improvements
• Automate VM schedules to save costs
• Choose Ultra Disk and distributed storage for high concurrency

Reference Materials

Links are provided for related Microsoft resources for deeper learning.