Microsoft has announced several enhancements for Azure Virtual Machines at their Microsoft Ignite Conference held last November. Here is a quick rundown of the changes to expect.
Azure Generation 2 Virtual Machines general availability
Improved features compared to generation 1 VMs
- Increased Memory
- Intel Software Guard Extensions (Intel SGX)
- Virtualized Persistent Memory (vPMEM)
- Uses the new UEFI-based boot architecture
Virtual Machine scale sets enhancement
- Customers will now be able to create an empty virtual machine scale set and add various virtual machines belonging to different series
- Provision VMs with custom images using a shared image gallery
- New scale-in policy that lets customers determine the order in which VMs should be scaled in, or deprovisioned
- New termination notifications with up to 15 minutes buffer before VMs are deprovisioned
Advanced security for SQL Server hosted on an Azure Virtual Machine
It supports the following features
- Vulnerability assessment. A database scanning service that can discover, track, and help you remediate potential database vulnerabilities
- Advanced threat protection. A detection service that continuously monitors your database for suspicious activities and provides action-oriented security alerts on anomalous database access patterns.
In addition, Microsoft also announced new Virtual Machines which are now available.
Dav4-series and Das v4-series Virtual Machines
- Workloads: For general purpose workloads
- CPU: up 96 vCPUs
- RAM: up to 384 GiBs
- Disk: 2,400 GiBs SSD-based temporary storage
- Based on AMD EPYC™
Eav4-series and Eas v4-series Virtual Machines
- Workloads: Memory intensive
- CPU: up 96 vCPUs
- RAM: up to 672 GiBs
- Disk: 2,400 GiBs SSD-based temporary storage
- Based on AMD EPYC™
NVv4 Azure Virtual Machines
- Workloads: GPU Visualization
- CPU: up 32 vCPUs
- RAM: up to 112GB
- GPU: up to 16GB of GPU memory
- Disk: Support for Azure Premium SSD Disks
- Based on AMD EPYC 7002 processors and virtualized Radeon MI25 GPU
NDv2-series Virtual Machines
- Workloads: Distributed HPC, AI, and Machine Learning
- CPU: up 40 vCPUs
- RAM: up to 672GiB
- Disk: 2,400 GiBs SSD-based temporary storage
- GPU: 8 NVIDIA Tesla V100 NVLINK interconnected GPUs with 32 GB of memory each
