This post is a follow up to last week’s one on Dell Technologies PowerStore, which was featured – among other products from the vendor’s portfolio – as part of Tech Field Day’s special two days event. Here we will focus on VxRail and I will briefly go through a few things that raised my interest, once again with a specific emphasis on its integration with the VMware stack. As usual, if you want to learn more, my recommendation is to head to the TFD event page where you can find plenty of videos, including demos, where the solution and its capabilities are presented in detail.
A “brief” overview of VxRail
VxRail is not exactly a new product, it has been around for a few years now and it has been developed by Dell as the platform of choice to run VMware VSAN clusters. As a matter of fact, it is a co-engineered solution that has its roots in the unsuccessful EVO Rail branded systems that many vendors marketed at the dawn of the HCI era. While most of them failed, Dell EMC went all in and launched its own platform, VxRail, on the solid foundations of the Dell-VMware partnership. VxRail’s strength was that it was a hardware platform purpose built to run vSphere and VSAN: when deployed and powered up, VSAN clusters came up quickly as a standardized, fully HCL compliant, best-practices adhering, future proof and ready to be consumed commoditized HCI workhorses. VxRail Manager was natively integrated within vCenter and VxRail Life Cycle Manager (LCM) enabled one click upgrades of the entire Hardware and Software stacks, from BIOS to HBAs firmwares, to vSphere and VSAN patches and updates, all validated against the compatibility matrix, fully supported and easily available in “VxRail bundles”. VMware admins could then shift their focus from spending considerable amounts of time cross-checking HCL matrixes and looking after delicate systems always prone to breakage to tasks like operations automation or applications development that brought higher value to the business. This created a massive shift in the approach to VMware-centric operations: it freed up energies and resources in operations teams and provided them with the elementary building blocks needed to establish the private cloud, moving the focus from the (now almost invisible) infrastructure to the more valuable workloads.
After this product establishment phase, things quickly became more complex: VCF came along with NSX and the SDDC Manager, then vRealize Automation was put on top of VCF to enable private cloud management capabilities, and eventually vSphere 7 and Tanzu were released opening the Kubernetes Pandora Box. VxRail wasn’t born with all of this in its sights, but it reacted and adapted quickly and while at the beginning it was just “another VSAN ready platform” it is now probably one of the best (if not the very best) solution to run the on-premises side of the multi-cloud unicorn.
If you had the patience to go through this very long intro, you’ll see shortly why I picked a few interesting points from the many presented at Tech Field Day Extra: LCM and vLCM integration, VCF and Tanzu on VxRail and finally VxRail Dynamic nodes.
Life-Cycle Management made (even) easier
With vSphere 7, VMware introduced their own vLCM to ensure consistency and HCL compliance of ESXi nodes, regardless of the server hardware vendor and having to consider all the possible exotic combinations of components. vLCM is a solution that augments and eventually replaces Update Manager by creating and enforcing a desired state and by pushing to the hosts an image that includes the latest ESXi build, HW vendor additional tools and drivers and finally the correct firmwares. You can see how similar this is to Dell’s own VxRail LCM: these two solutions aim to the same end result but while Dell’s LCM is tailor-made for VxRail, VMware’s vLCM has to work with any possible combination of HW (as long as all the parts are on the HCL, of course). The great news is that with vSphere 7 and VxRail 7.0 these two upgrade orchestrators can work together, fully integrated: vLCM can leverage LCM to package all the bits required to update the VxRail layer, applies any needed 3rd party drivers and firmwares (for instance, when non-VxRail HBAs are installed) and even updates the needed NSX-T and Tanzu specific components on the hosts. It’s immediately visible how these two tools can work well in synergy to ensure consistency, stability and compliance of vSphere clusters.
VxRail runs Tanzu at scale
Touching briefly on VCF support, that has been around since 2019; the latest combination of VxRail, VCF and Tanzu make VxRail an ideal platform to deploy Tanzu Workload Domains at scale. The value brought by the integrations between VCF’s SDDC Manager and VxRail Manager once again demonstrate how solid were the initial platform design pillars: VxRail is adaptable enough to support types of workloads and operational models that were not even there when it was still on the drawing board. VxRail is future-proof to say the least of it. If you want to see and hear more about the VxRail/VCF/Tanzu integration, there is a joint Dell/VMware presentation and demo available on the Tech Field Day website.
Going beyond VSAN with Dynamic Nodes
So far VxRail has proven to be more than just a platform to run VSAN and indeed… Dell Technologies just launched VxRail Dynamic Nodes which are essentially VxRail nodes with no local storage and therefore no VSAN! Primary storage is remotely accessible over FC as there is only compute capacity on the nodes. This sounds confusing: how can it be possible that a solution that has been designed to be the platform of choice for VSAN workloads is stripped of its HCI core capabilities? According to Dell Technologies, this is a request that comes straight from their customer base: there are many customers with considerably sized implementations of traditional VxRail nodes that want flexibility. They want to keep managing their VxRail clusters with that operational model I described above but they also want to be able to run nodes where storage is provided by storage arrays that have a different set of features (and capabilities) compared with VSAN. While VSAN is a very versatile storage solution, we cannot hide the fact that some workloads can work better on other SAN platforms because of richer feature sets or sheer performance requirements (PowerMax anyone?). These customers want to retain a consistent operational experience while having the flexibility of running their workloads on the platform that works best for them: VxRail Dynamic Nodes, decoupling compute from storage while preserving all VxRail’s operational management benefits fit the requirement precisely.
To conclude, I would like to re-iterate something I mentioned in my previous post: the real takeaway of the TFDx special event is the realization of the flexibility that the latest Dell Technologies portfolio brings to their customer. In addition to what PowerStore T, PowerStore X, PowerMax and the “regular” VxRail bring to the table in terms of infrastructure adaptability, the newly introduced VxRail Dynamic Nodes bring even more options to architects, who can count on a solid common philosophy while they can cherry-pick the piece of hardware that works best for their applications.