• Express Storage Architecture performance optimizations: In vSAN U2, customers can expect to achieve up to 30% greater performance on ESA through means like: single object streaming write performance improvements and enhancements to the DOM client.
• Increased per-host VDI VM density: vSAN’s per-hot VDI VM density has been increased to support 500 VDI VMs. With this improvement, ESA customers can consolidate 2.5x more VMs per host, provided the underlying hardware supports it, driving down costs and creating favorable TCO.
• Support for medium-sized witnesses: Support for a 16GB witness is now available in the Express Storage Architecture, creating greater scalability options in vSAN ESA, and accommodating smaller clusters or workload demands.
• Adaptive write path for vSAN Max: With an adaptive write path for disaggregated storage deployments on vSAN ESA, vSAN Max users can expect to see improved streaming writes on ESA, achieving higher throughput for sequential workloads.
• Support for vSAN file services: Support for file services is now available in vSAN Express Storage Architecture, enabling customers to easily provision a file share with a single workflow and use vSAN ESA as a unified, multi-protocol storage control plane for both block and file storage.
• Support of VCF management and workload domains: VMware Cloud Foundation supports vLCM, enabling you to run vSAN Express Storage Architecture in management domain, in addition to workload domain. Standardize and simplify operations through the co-existence of vSAN OSA and vSAN ESA in the same VCF and workload domain instance.

Simplify operations, issue detection, and remediation with vSAN

• Intuitive detection of VMs and disks consuming the most resources: The cluster-level health dashboard in the vSAN Skyline module has been enhanced with a Top Contributor view, simplifying the detection of VMs and disk consuming the most resources. This update also gives the ability to transpose other top contributors in the same performance view, whether it be IOPS, throughput, or latency. The highly customizable view enables administrators to quickly evaluate the VMs placing the most demand on the cluster, and determine if other cluster host resources are disproportionally consuming resources in comparison, accelerating and simplifying detection and remediation of performance issues.
• Customized, guided remediation: Triggered health findings now provide tailored guidance based on the version of vSAN running, and the topology that is configured.  The recommendations to resolve an issue are much more prescriptive, helping administrators resolve identified issues in a fast and effective way.
• Improved detection of performance bottlenecks in stretched clusters: It’s now easier than ever to identify the source of a performance bottleneck with the vSAN I/O Trip, allowing analysis of workloads running in a vSAN stretched cluster. Easily determine where the primary source of latency is occurring in a vSAN stretched cluster, as well as latencies in other parts of the stack that may be contributing to the overall latency seen by the VM.
• Auto-policy management improvements in vSAN ESA: The dynamic defaults storage policy feature in vSAN ESA has been expanded to support existing clusters. As hosts are added to or removed from an existing vSAN ESA cluster, the auto-policy management feature reconfigures the datastore default policy, as need, to the optimized configuration, applying either RAID1 or RAID 5/6.
• Simplified storage device management at scale: A new vSAN ESA prescriptive disk claim capability better integrates with HCL and gives administrators greater flexibility when claiming disks and accommodating customized hardware. Enjoy an automated, simplified disk-claim process with greater granularity, and simplify the management of storage devices for each host in a vSAN cluster.
• Easier configuration for 2-Node and stretched clusters: Workflow enhancements in vSAN 8 U2 give administrators a simple, automated UI option for setting up and configuring witness traffic separation in stretched clusters, simplifying operations, and improving user experience.