Dell intros updated VxRail and dedicated AI platform with VMware

HCI platform now supports vSphere 8, vSAN 8, and VCF 4.5

VMware Explore Dell has announced its latest kit co-developed with VMware, including updated VxRail HCI systems with support for network offloading to DPUs, support for VMware's Tanzu Kubernetes Grid in its Apex subscription platform, and a turnkey platform for operating AI workloads.

The announcements from VMware's former parent were timed to coincide with the VMware Explore event in San Francisco this week, where the cloud and virtualization biz is unveiling vSphere 8.0, the latest release of its platform.

This timing is no accident because the updated VxRail, Dell's hyperconverged infrastructure (HCI) platform for operating private clouds, now supports vSphere 8, vSAN 8, and version 4.5 of VMware's Cloud Foundation (VCF) hybrid cloud stack.

Dell VxRail

One of the new features in vSphere 8 is support for data processing units or DPUs, alternatively known as smartNICs, which can offload some of the processing tasks currently performed by the host CPU, especially network-related tasks such as firewalls, overlay networks, and security functions.

VMware is calling this feature vSphere Distributed Services Engine on DPU.

"DPUs are really intended to lay the groundwork for the future and will be a game changer for increasing workload demands on servers and the need for distributed scalable virtualized networking and security services," said Shannon Champion, Dell VP of Product Marketing.

This will also be supported on Dell PowerEdge systems, Champion said, but VxRail is the first running it, supplied as a ready-integrated solution.

Initially at least, DPU support only covers Nvidia Bluefield-2 and the Pensando DSC-25 DPUs, both of which are based on Arm processor cores.

One feature of vSAN 8 is vSAN Express Storage Architecture, described as a single-tier HCI storage solution, intended to deliver optimized performance, greater efficiency, and resilience, Champion claimed.

"It's up to 4x greater performance, and all NVMe VxRail models have been optimized for VSAN Express Storage Architecture, which will enable users to run more high performance applications in VxRail environments," she said.

Dell also announced VxRail ruggedized modular nodes designed for edge deployments. These are scalable from two to 64 nodes and specifically designed with half-width blades, but the most notable feature, according to Champion, is an integrated hardware witness capability that can store virtual machine metadata.

"The hardware witness removes the need for remote datacenter infrastructure power and administration, so this is especially useful in high-latency, low-bandwidth locations," she said.

These VxRail ruggedized modular nodes are targeting sectors such as retail, military, maritime, and manufacturing, and will be available globally sometime between now and the end of 2022, Dell said.

Dell also talked up its dedicated AI and machine learning platform, using VxRail as the base infrastructure.

The Dell Validated Designs for AI – the Automatic Machine Learning platform - uses automated machine learning models to assist customers in developing AI-powered applications, according to Dell.

Available now globally, it integrates VxRail with Nvidia GPUs, Dell PowerScale storage, and PowerSwitch networking, while the software layer comprises the H20.ai platform and Nvidia's AI Enterprise Suite.

"It's kind of like AI to do AI," said Champion. "And this integrated solution includes H20 Ai, which is a driverless AI to automate machine learning, together with Nvidia AI Enterprise Suite for cloud native AI development and deployment, delivered on VMware vSphere with Tanzu on an engineer validated and optimised Dell infrastructure stack."

Dell has also expanded its Apex portfolio of IT-as-a-service infrastructure solutions, specifically its Apex Cloud Services with VMware Cloud, which now has the option to add VMware Tanzu Kubernetes Grid services (TKG), managed by Dell, which enables them to provision and operate Kubernetes clusters through the vSphere user interface and run containerized workloads as well as in virtual machines.

"Organizations will be able to speed development efforts by building, testing and running cloud-native applications alongside traditional applications on the same platform," said Chad Dunn, VP of Product Management for Dell Apex.

"Often we see customers who are balancing the infrastructure needs of more modern development of container technologies, versus more traditional development of VM based workloads. And managing that the balance between those two can sometimes be tricky."

This option is available now in the US, UK, France, Germany, Australia, and New Zealand with additional global availability to come.

Apex Private Cloud and Apex Hybrid Cloud also now have compute-only options that allow customers to independently scale compute and storage resources, Dunn said. These can be connected to Dell's Apex Data Storage Services instead.

"We see customers who want a mix of the hyperconverged flavor using vSAN as the underlying storage, but for some workloads, they want to be able to leverage external storage, whether that's traditional arrays, PowerStore, PowerMax, or very specifically, in our case, Apex Data Storage Services," Dunn said.

This will be available in the US, UK, France, and Germany from September with other territories to follow. ®