Azure Cobalt 100-based VMs
The new Azure Cobalt 100-based Virtual Machines (VMs) are now readily available to the general public. The 64-bit Arm-based Azure Cobalt 100 CPU, which Microsoft first developed entirely in-house, powers these virtual machines. In optimizing and customizing each layer of the infrastructure stack from silicon to servers to services, they mark a critical turning point in planning and advancing cloud architecture. Azure Cobalt 100-based vms, which combine hardware and software vertically, are one of Microsoft’s most recent innovations in improving and optimizing its cloud infrastructure using an end-to-end systems approach, to provide its customers with the ideal balance of performance, power efficiency, and scalability.
This includes its memory-optimized Epsv6-series VM series as well as its new general-purpose Dpsv6-series and Dplsv6-series VMs. For a variety of scale-out and cloud-native Linux-based workloads, such as data analytics, web and application servers, open source databases, caches, and more, they are an appealing alternative since they provide up to 50% better pricing performance than our previous generation Arm-based virtual machines.
Comparison of Azure Arm-based and new Azure Cobalt 100-based virtual machines
Comparing the new Azure Cobalt 100-based vms to earlier generations of Azure Arm-based virtual machines, the former provide superior performance on a variety of workloads: up to 1.4x CPU performance, up to 1.5x performance on workloads based on Java, and up to 2x performance on web servers,.NET applications, and in-memory cache applications. Compared to the earlier generation of Azure Arm-based virtual machines, these VMs can handle up to 1.5 times network bandwidth and 4 times local storage IOPS (with NVMe).
Availability
North Europe, Southeast Asia, Sweden Central, Switzerland North, UAE North, West Europe, Japan East, Mexico Central, Canada Central, Central US, East US 2, East US, Germany West Central, and West US 2 are all places where the new virtual machines are widely accessible. Australia East, Brazil South, France Central, India Central, South Central US, UK South, West US 3, and West US will be additions to the list in 2024 and beyond.
Consumer acceptance and situations
While the preview period was in effect, Azure worked with several internal and external clients. As an illustration, IC3, the platform that facilitates billions of customer discussions in Microsoft Teams, is improving customer service for its expanding clientele by up to 45% on virtual machines (VMs) based on Cobalt 100.
It also provides Cobalt 100-based VMs to a large number of its independent software vendor (ISV) partners that provide software as a service (SaaS) and platform as a service (PaaS) via Microsoft Azure.
Embracing innovation and customer benefits on the path to Arm
The integration of Arm technologies and contributions to Arm architecture are not new to Microsoft. Its development of significant industry standards that readied the Arm architecture for datacenter-scale computing was made possible by this experience. Along with this, it has been collaborating closely with Arm on industry projects like ServerReady and SystemReady, both of which have earned industry accolades.
The goal of its foray into Arm-based virtual machines is to provide exceptional power efficiency and price-performance. Because of these advantages, the Cobalt 100-based virtual machines (VMs) represent this hope. Microsoft customers may now enjoy a unique blend of cost-effectiveness and performance with its adoption of Arm-based virtual machines.
Ecosystem for developers
With significant advancements over the past few years, Arm’s developer ecosystem is still thriving. Major programming languages and platforms, including Java, C++, and.NET, offer Arm-native versions. For each of these platforms and languages, it has made investments in Arm-specific optimizations to make the most of the Arm architecture’s potential.
With several well-known infrastructure and deployment options now offering native Arm compatibility, the broader ecosystem has embraced Arm. To continually create, test, and deploy apps, many developers utilize GitHub Actions, the company’s continuous integration and continuous delivery (CI/CD) workflow engine. Both GitHub-hosted runners and self-hosted runners which may be installed on local Arm hardware or in an Arm virtual machine are now available for Arm.
Because of their efficient resource use, portability, reproducibility, isolation and security, and faster development workflow, containers are a popular deployment target. It is now possible to create Arm agent nodes and combine x86 and Arm architecture nodes in a cluster using Microsoft Azure Kubernetes Service (AKS).
Characteristics
With Azure Virtual Machines, you may choose from a variety of memory ratios for a specific vCPU size, allowing you to customize the setup that best suits your workloads in terms of memory requirements and CPU performance. To enable you to implement the choice that best suits your workload, all of these virtual machine series are offered with and without local drives.
- With a memory-to-vCPU ratio of 4:1, the new Dpsv6-series and Dpdsv6-series general-purpose virtual machines have up to 96 vCPUs and 384 GiB of RAM. They operate well with cloud-native systems like AKS, web servers, application servers, small to medium open-source databases, and scale-out workloads. Arm developers can use these virtual machines (VMs) in development, test, and CI/CD pipelines.
- 192 GiB of RAM and up to 96 vCPUs are available in the new Dplsv6-series and Dpldsv6-series virtual machines (2:1 memory-to-vCPU ratio). Media transcoding, microservices, gaming servers, small databases, and workloads that don’t require a lot of RAM per virtual CPU are all ideal uses for them.
- The new memory-optimized virtual machines (VMs) from the Epsv6 and Epdsv6 series have up to 96 virtual CPUs and 672 gigabytes of random-access memory (up to 8:1 memory-to-vCPU ratio). These virtual machines are intended for tasks requiring a lot of memory, like data analytics, in-memory caching apps, and big databases.
Standard SSD, Standard HDD, Premium SSD, and Ultra Disk storage are among the remote disk formats that the new virtual machines support. See Azure managed disk type for further information on different disk types and their availability by location. Unlike virtual machines, disk storage is priced separately. You can use the command-line interface (CLI), PowerShell, SDKs, Azure portal, and APIs to deploy these new virtual machines.
Costing
The Azure Virtual Machines pricing and Pricing calculator pages provide additional information regarding the cost of Azure Cobalt 100-based VMs.
You can also reduce your expenses by using Spot Virtual Machines, Reserved Instances, and Azure’s compute savings plan. Because reserved virtual machine instances are paid for in advance for one or three years, they can lower expenses and enhance your budget forecasts. Get one-year Azure Reserved Virtual Machine (VM) Instances for specified Linux VMs and save 15% for a limited time.
From October 1, 2024, to March 31, 2025, this bargain is valid. Details can be found here. You can choose how much you want to spend on Azure VMs and other Azure services with the Azure savings plan for compute. For applications that can withstand disruptions and have variable execution times, Spot Virtual Machines can further optimize your cloud expenditures while also drastically lowering the cost of operating in Azure.
Power efficiency and pricing performance in a new era
With the release of Azure Cobalt 100-based VMs, a new era in Azure’s architecture has begun. Through its proprietary silicon program, it provides its clients with outstanding value, performance, and power efficiency.
