Optimize VMware with HCI & Pure Storage


Something interesting has been occurring over the past 12 months, as the year has progressed, we’ve seen more and more customers deploy hyper-converged infrastructures (HCI) alongside all-flash arrays (AFAs) in their vSphere environments. Customers are leveraging the strength of VMware’s SDDC technologies to align deployment success criteria to the type of infrastructure best capable of exceeding the success threshold.

Bottom line: Customers are optimizing vSphere environments with a combination of HCI and the Pure Storage FlashArray.

Pure Storage has strong partnerships with both the VMware vSAN and Cisco Hyperflex teams. We absolutely compete with the rest of the HCI and storage market.

VMware Makes Infrastructure Ubiquitous

VMware SDDC delivers interoperability and operational consistency across dissimilar architectures be they HCI, AFAs, or the public cloud. This interoperability spans vSphere functions like data and VM mobility (VMotion), orchestration and automation (vRealize) and disaster recovery (SRM), as well as the physical infrastructure variants including dissimilar storage fabrics (i.e. Fibre Channel, iSCSI, FCoE) and storage formats (i.e. VMDK, VVol, VMFS, NFS, vSAN)

I wrote about this topic last year for VMware, specifically on vSAN and FlashArray VVols interop. The piece is an excellent primer for this post if you’re interested.

To truly maximize the benefits of infrastructure ubiquity, one needs to focus on applying this flexibility to optimize success measured in deployment results. Put in other terms – align deployment success criteria to the capabilities of the underlying infrastructure.

Deployment Types & Success Criteria

There are three common types of vSphere deployment types, each of which has a distinctly different success criterion. These are…

  • Server and Desktop Consolidations:
    This is the most common form of private cloud deployment. Consolidations are commonly comprised of tier-2 applications and thus seek to optimize cost, simplify management at scale, and focus on infrastructure availability.
  • Tier-1 Applications:
    This type of deployment focuses on the performance and availability of the business-critical application.
  • Remote and Branch Offices:
    These types of deployment occur outside of the core data center, tend to be modest in scale (3 to 8 servers are common), require to be physically maintained by non-IT staff, while remotely managed by IT. 

I believe the shared success criteria to be fundamentally sound; however, you may need to modify for your particular needs. 

Aligning Infrastructure to Success Criteria

If we apply attributes that are broadly observed across the FlashArray customer base, it is trivial to align infrastructure platforms to deployment success criteria.

  • Server and Desktop Consolidations are Best Served on FlashArray:
    This conclusion is based on the storage efficiency (the combination data reduction and capacity overheads like RAID), the reduction of compute required with a disaggregated architecture versus HCI, and platform availability.
  • Tier-1 Applications are Best Served on FlashArray:
    This conclusion is based on platform performance capabilities measured in latency and throughput with consideration for application availability. The latter includes data protection levels, the impact of platform upgrades, and overall platform availability.
  • Remote and Branch Offices are Best Served with HCI:
    This conclusion is based on the platform’s ease of initial hardware setup and affordability when supporting smaller footprints.

I don’t expect all readers to agree with my conclusion in this chart and frankly, that’s OK. I suspect Pure Storage FlashArray customers and channel partners are more inclined to agree with the contents compared to those who’ve yet to experience Pure Storage. 

What’s Next: VMware Cloud Foundation Featuring vSAN and FlashArray

VMware’s modern platforms VMware Cloud Foundation (VCF) and VMware Cloud on AWS (VMC) share a standard architectural design principle, one based on management and workload domains which enable better lifecycle management and automation. VCF & VMC initially only supported vSAN for both the management and workload domains. Recent versions of VCF have added support for 3rd party storage for workload domains. The extended support helps customers leverage existing infrastructure investments (i.e. Fibre Channel fabrics or AFAs) and the unique capabilities from their infrastructure platforms.

Note: VMC currently only supports vSAN and NFS for workload domains. We’re working with VMware to expand support for 3rd party block storage in the future.

Pure’s VMware Solution Architects are heavily invested in VCF and VMC. This means we are actively working with the vSAN team at VMware. If you join us in Austin in September for Accelerate, the Pure Storage user conference, you’ll be among the first to learn about the initial fruits of the team’s efforts.

Wrapping Up This Post

One size fits all infrastructure is a fallacy for most organizations; there’s simply too diverse a set of success criteria for a single platform to deliver on. Hopefully, I’ve been able to share my appreciation of VMware’s SDDC capabilities to make infrastructure ubiquitous and the ability to deploy dissimilar architectures like HCI and AFAs in order to best meet the success criteria of a deployment. Yes, there’s product positioning included, but that’s the price of admission for this post.

If you’re attending VMworld 2019 next week, stop by our booth to have a technical discussion with one of our VMware experts and take in a few demos. You can also register for a private briefing. If you’re not attending, reach out to a Pure Storage partner and let’s discuss your organization’s goals, complete the TCO tool, and set up a POC.

Skeptical questions are welcomed in the comments section.

Leave a Reply