What’s New in vSphere 8 – Part 2

In the latest VMware Explore US 2022 (previously called VMworld) event, VMware announced the new vSphere 8. The vSphere 8 is officially expected to be available by October 28, 2022.

As discussed in the article What’s New in vSphere 8? Part 1, there are many new features and enhancements in this new vSphere 8, and I can’t go through them all, but I will try to focus on the most important ones since we couldn’t fit them in one article. It was divided into two parts, and this is vSphere 8 What is New? – Part 2.

Table of Contents

1. What is new in vSphere 8?
2. VMs, Guest OS, and Workloads
3. Virtual Hardware(vHW) Version 20
4. What is the Virtual Trusted Platform Module Provisioning Policy(vTPM)?
5. High latency sensitivity with hyperthreading
6. Simplified vNUMA configuration:
7. Compute Maximums
8. Unified Management for AI/ML Hardware Accelerators
9. Device Virtualization Extensions(DVX)
10. Simplified Hardware Consumption with Device Groups
11. Resource Management – Monitoring
12. vSAN 8
13. Optional Next Generation Architecture is built into vSAN
14. A New Way to Process and Store Data Efficiently
15. Conclusion

What is new in vSphere 8?

In this second part, we will show the following new features:

• VMs, Guest OS, and Workloads
• Resource Management
• Artificial Intelligence & Machine Learning
• Resource Management – Monitoring
• Security and Compliance
• vSAN

VMs, Guest OS, and Workloads

I will show and detail by section the changes in vSphere 8 regarding VMs, Guest OS, and Workloads.

In the Virtual Machines and guest operating segment, there are also some good changes in vSphere 8. The latest virtual machine hardware version, particularly in vNUMA.

• Virtual Hardware version 20
• Virtual TPM Provisioning Policy
• Migration aware applications
• High latency sensitivity with hyperthreading
• Simplified vNUMA configuration
• Compute Maximums

Virtual Hardware(vHW) Version 20

We have already discussed some of the changes to the maximum in the first article. There are new maximums for vGPU(8) and DirectPath I/O devices(32).

This new vHW brings updated support for the latest Intel and AMD CPUs and guest operating systems.

What is the Virtual Trusted Platform Module Provisioning Policy(vTPM)?

Windows TPM is an encryption method that protects and ensures the privacy of your system operations. Using cryptography protects software and authentication through encryption and decryption using the UEFI firmware version on your VM. Windows 11 requires vTPM devices to be present in virtual machines.

The problem in previous Virtual Hardware versions is that when you have vTPM enabled on your VMs if you clone a VM or use VM Templates, you can risk security as TPM secrets are cloned.

In vSphere 8, vTPM devices can be automatically replaced during clone or deployment operations, providing best security practices and ensuring that each VM contains a unique TPM.

Migration-aware applications were and still are always a problem when we migrate some VMs using vMotion. Some applications cannot tolerate the stuns associated with vMotion.

I have often seen this problem with VMs using communications (mobile core networks) and VoIP applications. Or when using Cluster Applications or Databases.

This is the new process task when starting vMotion for this kind of application using Migration aware:

1. Applications register using VMware Tools and can be notified of impending vMotion tasks.
2. The application receives a notification indicating that the VM will be migrated
3. The application can respond to this notification as desired. This could be done by gracefully stopping services or by performing a failover in the case of a clustered application, for example
4. The application sends an acknowledgement that the migration can proceed
5. The VM is migrated using vSphere vMotion
6. The application receives a migration end notification and can respond as needed based on the preparation work performed before the migration

I hope this new process dramatically improves the migration of this application since vMotion significantly impacts production for this type of application.

High latency sensitivity with hyperthreading

We could say that the high latency sensitivity with the hyperthreading feature is a continuation of the previous feature-Migration aware, because, like the previous feature, it was created to support highly sensitive applications and servers.

As I said, I have worked with some very sensitive VM Telco applications. For example, when we do a vMotion of one of those VMs with those Telco applications, some have a cluster or mirror in the Guest OS, and in the migration, if they have any latency, the mirror is out of sync and breaks the DB or the application.

So this feature is to support latency in those sensitivity applications. In this case, they work at the virtual machine vCPU level(using and scheduling the same hyperthreaded physical CPU core), delivering improved performance on those workloads.

To enable High latency sensitivity with hyperthreading, select in VM advanced settings.

Note: To use this feature on a VM requires 100% reservation on vCPU and vMemory.

Simplified vNUMA configuration:

Another feature is simplifying some of the settings for some VMs, particularly the Monster VMs with Business Critical Applications.

Monster VMs, for example, Databases, with large memory and vCPUs, Migration aware, High latency sensitivity with hyperthreading, and this simplified vNUMA configuration arrives to improve vMotion and performance with less disruption.

In previous vSphere releases, to control vNUMA topology, we needed to set parameters in the advanced VM settings. vSphere 8 with vHW v20 can be done directly in the VM since there is an option in the VM summary tab, and the CPU topology tile is visible.

Compute Maximums

I already showed the table for the new vSphere 8 maximums in the first article. This is just a resume of what it changed.

Unified Management for AI/ML Hardware Accelerators

• Device Virtualization Extensions(DVX)
• Simplified Hardware Consumption with Device Groups

Device Virtualization Extensions(DVX)

In previous releases, VMs consuming physical hardware devices using DirectPath IO or Dynamic DirectPath IO were limited in their mobility.

In vSphere 8, VMware takes DVX to another level by introducing a new framework API for vendors to create hardware-backed virtual devices.

VMware provides a new Dynamic DirectPath IO to vendors to support virtual hardware devices that can be used when using vSphere vMotion, suspending, or resuming any VM. It also supports disk and memory snapshots.

This new DVX and the new Dynamic DirectPath IO, a virtual machine that uses a physical hardware device on a specific ESXi host, can now be migrated with vMotion to another ESXi host and still use the virtual device, as long as the ESXi host supports it. And for an ESXi host to support and share the virtual hardware device(VDX) between ESXi hosts in a cluster, it needs to have the device virtualization extension driver installed.

The Device Virtualization Extensions support:

• vSphere DRS and vSphere HA
• Live migration using vSphere vMotion
• VM suspend and resume
• Disk and memory snapshots

Simplified Hardware Consumption with Device Groups

Device groups can be composed of two or more hardware devices that share a common PCIe switch or devices that share a direct interconnect between each other. Device groups are discovered at the hardware layer and presented to vSphere as a single unit representing the group.

In vSphere 8, after you create a device group, you can just add a device to the VM using the add PCI Device and select your group, and you will have devices that are shared with a direct interconnect between each other.
DRS and vSphere HA are aware of device groups and will place VMs appropriately to satisfy the device group.

Resource Management -Monitoring

• Enhanced DRS Performance
• Monitor Energy and Carbon Emissions

Enhanced DRS Performance is vSphere Memory Monitoring and Remediation v2 (vMMR2). It is an upgrade of a new feature introduced in vSphere 7.0U3 called vMMR.

vMMR provides bandwidth statistics for VM and host levels, default alerts, and the ability to configure custom alerts based on VM workloads.

In vSphere 8, vSphere Memory Monitoring, and Remediation v2, DRS performance can be significantly improved when Intel Optane PMEM is present by leveraging memory statistics, resulting in optimal placement decisions for VMs without affecting performance and resource consumption.

vMMR2 allows an administrator to properly troubleshoot performance issues and any possible memory bottlenecks between DRAM and PMem.

With Monitor Energy and Carbon Emissions, administrators check the power consumption of hosts and VMs. This allows administrators to monitor the energy and carbon emissions of all vSphere infrastructure.

In a green world where many enterprises care about their carbon footprint, these new reports can help companies check their energy efficiency and save money on power while reducing their carbon footprint by powering off unnecessary VMs or hosts, not in use but consuming power.

The three new metrics track :

• Power consumption of a hosta system activities; how much power the host is using and is not attributed to VMs
• Power consumption of a hosta idle activity; how much power the host is using when it’s not doing anything except being on
• Power consumption of a host due to VM workloads; how much power the host is using to run VM workloads

When integrated with vRealize Operations Manager, these metrics illustrate carbon intensity and emission rates

Security and Compliance
Secure out-of-the-box and security changes in vSphere 8 are mainly set in 4 steps:

• Prevent Untrusted Binaries: Activating the execInstalledOnly option by default prevents untrusted binaries from running. Only binaries installed via signed VIBs are trusted
• TLS 1.2 Only: vSphere 7 previously deactivated TLS 1.0 and TLS 1.1 by default. In vSphere 8, these protocols are removed, and only TLS 1.2 is supported
• SSH Timeout: SSH access is deactivated by default, and in vSphere 8, a default timeout is introduced to prevent SSH sessions from lingering
• Sandboxed Daemons: ESXi 8.0 daemons and processes run in their own sandboxed domain where only the minimum required permissions are available to the process

vSAN 8

Regarding vSAN 8, there are also a lot of changes and new features. But for that, I would need a complete new vSAN article just to detail all the new features, enhancements, performance, and options.

So I will only add the new main features sections:

Optional Next Generation Architecture is built into vSAN

vSAN Express Storage Architecture™ (ESA) in vSAN 8.

• Next generation storage architecture for the workloads of today and tomorrow
• Optional, alternative architecture to the vSAN original storage architecture
• Available when running on qualified hardware in ESA-approved vSAN ReadyNodes
• Built into the software you know

A New Way to Process and Store Data Efficiently

• vSAN Express Storage Architecture
• Highly adaptable to the hardware of today and tomorrow
• Stores data efficiently while meeting performance requirements
• Higher default levels of resilience
• Native, scalable snapshots
• Simplifies administration and serviceability

• Improved Efficiency through Optimized Data Handling
• New high-performance RAID – 5/6 erasure coding
• Intelligent I/O traffic management for vSAN network traffic
• Adaptive RAID-5 to accommodate cluster conditions
• Storage policy-based compression

…and much more features.

Conclusion

With this vSphere 8 Part 2, we finish detailing almost all the changes that will arrive in the new vSphere 8.

In this part 2, we discussed one of the most important and that had a lot of changes, Virtual Hardware v20. With many changes and improvements, both in performance and security, and how VMs and applications now react to a migration using vSphere vMotion.

For someone that worked a lot with Monster VMs and highly sensitive applications and servers in some Critical Business Applications, Migration awareness and High latency sensitivity with hyperthreading together with the vNUMA changes is an outstanding improvement and good help for many Enterprise customers, even when running this type of VMs in a Hyperscaller.

As VMware stated, some of the features are still in development and should have more improvements or changes in future releases. But until vSphere 8 is officially GA, VMware may make some changes and improvements.

Related Posts:

What’s New in vSphere 8 – Part 1
What’s New in vSphere 8 Update 1

Follow our Twitter and Facebook feeds for new releases, updates, insightful posts and more.