Today, we are excited to announce the latest updates to vSphere 8. This release includes not only Version 8 Update 2 of our enterprise workload platform, but also includes a new cloud service to which users of vSphere+ will soon have access. These updates will help enhance the operational efficiencies of IT admins, supercharge the performance of demanding workloads, and accelerate the pace of innovation for DevOps engineers, developers, and anyone else that can benefit from self-service access to infrastructure services.
Enhance Operational Efficiency
With this latest release, we placed a heavy emphasis on lifecycle management. Updating and upgrading vSphere can sometimes be challenging in large environments because it often extends across the entire on-premises footprint, running on possibly hundreds or thousands of systems. With enhancements, patches, and security fixes coming out regularly, VMware vSphere needs to be quick and easy to update. This release takes a huge leap forward in accomplishing that goal. In addition to lifecycle management, we wanted to make it easier for IT admins to centrally manage authentication by continuing to expand our support for 3rd party identity providers.
ESXi Lifecycle Management Service – An extremely impactful enhancement is the new cloud service in vSphere+ that enables admins to centrally orchestrate updates across their entire fleet of ESXi hosts. The time and effort needed to manage the lifecycle of ESXi hosts can be considerable, as these represent the vast majority of vSphere infrastructure. Numerous lifecycle operations (such as patches and upgrades) must be performed as dictated by VMware vCenter and cluster boundaries. This new cloud service, available to vSphere+ customers, will allow IT admins to update an entire fleet of similarly configured ESXi hosts, across multiple vCenters and clusters, in one operation and centrally monitor progress from the Cloud Console. An example can be used to show the benefit more clearly. With traditional vSphere, one update is required for every cluster of ESXi hosts, and the maximum number of hosts per cluster is 32. Suppose a user has 32,000 ESXi hosts spread across 1,000 clusters. To update their entire ESXi fleet, the user would need to perform 1,000 separate update operations. With the ESXi lifecycle management service, only one update is required for each standard hardware configuration, regardless of the number of clusters. Most IT organizations standardize their hardware configurations, so in this case it is reasonable to assume this customer might only have around 3-5 standard hardware configurations (depending on how many server vendors they purchase from). This means they could update their entire ESXi fleet in as few as 3-5 update operations, a tiny fraction of the 1,000 updates needed with traditional vSphere. With the ESXi lifecycle management service, upgrades will require far less time and effort and can be performed more often, allowing customers to stay more secure and take advantage of the latest capabilities of ESXi.
Reduced Downtime for vCenter Upgrades – Originally introduced for vSphere+ customers last year, we have expanded the availability of this functionality to all editions of vSphere. During an upgrade, vCenter downtime is reduced roughly from an hour to a few minutes (your mileage may vary). This is essentially the time required to shut down services on the old vCenter and start services on the new one. So planned maintenance windows needed are much shorter. Upgrades may be done more frequently because it is so much easier to do, thereby benefiting from the latest vCenter features.
Support for Microsoft Entra ID (formerly Azure AD) Identity Federation – For the last couple years we have been steadily expanding our support for 3rd party identity providers, including Microsoft Active Directory Federation Service (ADFS) and recently Okta Identity Service. With this release we are expanding this support to include Microsoft Entra ID (formerly Azure AD). When authentication is handled by a 3rd party identity provider like Entra ID, authentication is no longer processed by vSphere. Since logins and passwords are no longer stored within vSphere, security audits are easier, faster, or not necessary.
Supercharge Workload Performance
With the sensational introduction of ChatGPT in November 2022, there has been an industry-wide surge of interest in generative AI (GenAI). As reported in the press, by January 2023 ChatGPT had become the fastest-growing consumer software application in history, gaining over 100 million users. As a result, GenAI is now a strategic priority for many organizations. vSphere has been on the forefront of AI ever since the introduction of our AI-Ready Enterprise Platform in March 2021, and with this latest release we continue to push the AI envelope by continuing to scale and refine our GPU virtualization technology. Along with AI-related enhancements, we continue to expand the availability of Data Processing Unit (DPU) technology across more hardware platforms so customers can realize these performance benefits.
Increase maximum vGPU devices per VM from 8 to 16 – Larger workloads (especially AI workloads) continue to require more and more GPU power. With this latest release, we have increased the maximum number of vGPU devices that can be assigned to a single VM to 16, doubling the upper performance limit for larger workloads. For AI, this means you can reduce AI/ML model training time and run higher end models with larger data sets.
GPU-aware DRS workload placement and load balancing – We have discovered scenarios where workloads may not fully utilize the GPU resources available. To address these scenarios, we have enhanced DRS. DRS is now aware of vGPU profile sizes and consolidates equivalent size vGPUs on the same host. This also helps with initial placement during power-on of GPU-enabled VMs, which avoids wasting GPU capacity caused by fragmentation. Workload placement and load balancing is now GPU aware, and DRS will try to place workloads with similar profile requirements on the same host. This improves GPU resource utilization, which reduces cost, as fewer GPU hardware resources must be acquired to achieve a desired level of performance.
Expand DPU support to include Lenovo and Fujitsu server hardware – We introduced support for DPUs a year ago with the introduction of vSphere 8, allowing customers to shift infrastructure workloads away from CPUs to a dedicated DPU, thereby improving the performance of business workloads. With this release, customers utilizing Lenovo or Fujitsu servers will now be able to take advantage of vSphere DPU integration and its performance benefits.
Accelerate Innovation for DevOps
For infrastructure consumers, such as DevOps engineers or developers, self-service access is paramount. Having to email someone or open a ticket to get access to infrastructure services such as servers, storage, or networking, is a great way to slow things down. Ever since we introduced our Kubernetes integration in 2020, we have continued to raise the bar for self-service access and this release is no exception. In addition to giving infrastructure users faster access, we have also made it easier for IT admins or DevOps engineers to set up their Kubernetes environment.
New VM Image Registry Service – For DevOps engineers or other users who need to create virtual machines, they need a way to store those VM images so they can be reused or shared. This release provides a new VM Registry Service allowing consumers to publish, modify, and delete images using Kubernetes APIs. Images can be used for the deployment of VM Service VMs. With this new registry, DevOps teams, developers, and other VM consumers can publish and manage VM images in a self-service manner allowing VMs to be created more quickly without requiring assistance from an IT Admin.
VM Service support for Windows and GPUs – The VM service is a great way to provision VMs in a self-service manner, but in the past it has been limited to Linux VMs and select configurations. This release removes these limitations. VM Service can be used to provision Windows VMs alongside Linux VMs. VMs can also be deployed with VM Service with any hardware, security, device, custom or multi-NIC support, and passthrough devices that VMs on vSphere support, achieving complete parity with traditional vSphere VMs. Importantly, VM Service can now be used to provision workload VMs that require GPUs.
Import and Export Supervisor Cluster Configuration – Setting up a supervisor cluster can take a long time due to the many settings that are required. If you have many clusters to set up, this complexity compounds. We have added enhancements in this release to alleviate this complexity. A Supervisor cluster configuration can now be saved and exported to other Supervisor clusters. You can now easily export an existing supervisor cluster configuration and import into a new supervisor cluster, bypassing the manual configuration process. In addition, this will allow you to more easily replicate a standard configuration across multiple supervisor clusters.
Next Steps
This article highlights the key capabilities in this update, but there are many more enhancements to learn about. Ready to get started on your journey with vSphere? Check out these helpful resources:
