VMware builds vSAN Max, consolidates multicloud and online deployment

vSAN Max SDS extends storage independently of VMs

Compared with vSAN software-defined storage, which is integrated with server clusters that run VMs, vSAN Max provides a cluster dedicated to storage. A key benefit is financial, because vSAN Max allows customers to extend the number of disks from less powerful servers – so, less costly – when vSAN requires additional servers with the same CPU performance.

vSAN was developed to allow easy extension to projects by adding identical nodes as required, as in the hyper-converged principle. For its part, vSAN Max is better adapted to scenarios in which applications consume more data over time. vSAN Max can handle capacity up to 8.6PB with NVMe flash supported and 3.6 million IOPS per dedicated storage cluster.

Commercially speaking, vSAN Max isn’t an extension of vSAN but a separate product. So, vSAN is invoiced on the number of processors in the cluster while vSAN Max is charged according to storage capacity managed. In the two cases, the systems simulate a storage array separate from servers in a cluster.

“You can say that vSAN in its Max version is a storage array like our partners offer,” said John Gilmartin, VMware director. “But the aim of vSAN is always to facilitate storage utilisation, via our vCenter admin console, which also manages VMs and which will also be the case with Max.

“We offer an admin environment that resolves the problems of storage for you, taking into account the same priorities for each volume as the volumes on the VMs they’re associated with,” he added.

“The other point, above all, is that we claim excellent performance that compares to other SAN solutions. And that is made possible by a total reworking as part of the vSAN 8 update.”

Quick cloud restores in case of an attack

The key new feature of Ransomware Recovery is to put into production several restored copies at once instead of one after the other. That simple improvement will be critical in the sense that it potentially drastically reduces the time taken to restore enterprise activity after a ransomware attack.

Ransomware Recovery, launched last year, determines the health of backup copies before restoration. It is an extended function of Disaster Recovery, which is a long-established VMware service that carries out regular snapshots of a cluster and that can in case of a problem automatically restore its contents in a virtual cluster on AWS.

“The relevance of the snapshot system is that, unlike a backup system, you don’t have to extract data to put it into useable form again,” said Gilmartin. “With snapshots, you access just the blocks that contain the copies of your data in blocks that are directly useable. It’s instantaneous. You just have to restart restored VMs, and Ransomware Recovery handles it.”

In practice, to access Ransomware Recovery an enterprise has to have a VMware Cloud on the online portal, then subscribe to the VMware Disaster Recovery service, then buy the Ransomware Recovery option.

Originally, only production VMs on-site or run via VMware Cloud in AWS can be protected and restored like this. Currently, this functionality can be applied to VMs that run in the GCP VMware Cloud service. VMware didn’t say when this functionality will work with VMs in the Azure service. In all cases, however, restores are to AWS.

From here to the end of the year, Ransomware Recovery will benefit from a new file system that will completely partition restored VMs from the network infected by cyber attack.

“In future, we will adapt our file system so that restores are possible to clouds other than AWS,” said Gilmartin. “Our other technology direction is to make on-demand restores from backups more productive, such as during VPN roll-outs to allow different categories of user to work.”

NSX+ to unify the network across different sites

The aim of the online network engine NSX+ is to automatically apply network rules from a source cluster to all of its replicas, whether deployed on a distant site or in a cloud, on VMs or using containers.

“We evolved our virtual network offer, NSX, towards NSX+ to simplify our customer journey towards multicloud,” said Kit Colbert, technical director at VMware.

NSX+ will eventually be equipped with a supervision console, NSX+ Intelligence, which is intended to map vSphere cluster deployments across multiple sites in a visual manner, highlight faults, bottlenecks and other potential problems. NSX+ Intelligence will offer means to resolve issues.  

For now, NSX+ aims to unify routing rules, rights of access and ID attribution for VMs and containers. Colbert suggested that VMware is planning a further evolution of NSX+ that will take charge of mesh networks on Kubernetes clusters and manage the API presented by the containers.