Making Money in Hyper-Converged Infrastructure

This week, Scale Computing announced record revenue growth and a transition into profitability. As a company we’ve tracked for a long time, what can we observe about the market and why Scale has achieved something none of the other HCI vendors have yet delivered?

Background

HCI is, of course, Hyper-Converged Infrastructure or the convergence of computing, storage and virtualisation into a solution that eliminates the need for dedicated storage. Back in 2009, Nutanix (one of the leading HCI vendors by revenue) brought their HCI solutions to the market under the banner “No SAN”. We first saw the company at a Tech Field Day event in 2011.

As a concept, HCI is great for businesses that don’t want to manage storage resources directly, meaning the installation and deployment of large-scale storage networking and storage appliances. In large enterprises, the idea of SANs isn’t at all daunting, despite the myths about the cost of running such types of infrastructure. For small or medium businesses, the HCI model looks more attractive, especially if the majority of the applications use virtual machines and traditional databases.

SDDC

During the last ten years, we have seen a transition in enterprise computing that has primarily been driven by the concept of Cloud, either public or private. Without going into a debate about the definition of Cloud, there are some characteristics we can all agree on.

Scalability – Scale up/down on demand in a granular fashion
Automation – consume resources on demand
Utility billing – pay by services used, not the infrastructure cost
Resource pooling – consolidation of resources with multi-tenant capability

We can also add resiliency, solid security models and cost efficiency to this list, but these are features we should expect in any IT infrastructure.

Cloud characteristics align well with HCI, enabling IT organisations to build on-premises “software-defined data centres” that increase agility for the end user, while reducing operating costs. An additional aspect of HCI is the granular nature in which servers/nodes can be added to a deployment and provide incremental growth in capacity.

Scale Computing

The solutions offerings from Scale Computing meet all the above requirements. We’ve been lucky to have run a Scale Computing cluster in the Architecting IT Lab since 2015, so have first-hand experience with the technology.

As a solution for small/medium business, a Scale Computing cluster represents ease of use and value for money. The platform is built on KVM, with additional features such as an O/S native storage data plane called SCRIBE. Provisioning can be driven through automation via API and CLI. All of the hardware aspects of a cluster are hidden from the user. If you want a platform that easily enables the resilient deployment of virtual machines, HC3 just works.

Applicability

To be fair, there are some gaps in the HC3 offering. The platform is nowhere near as evolved as VMware’s vSphere and associated ecosystem tools. You won’t find application deployment catalogues or support for containers (at least, not yet). These restrictions may be an issue for large enterprise customers.

However, Scale has found a sweet-spot for their solutions with edge computing and VDI. In October 2019, we discussed the challenges of edge computing on a Storage Unpacked podcast with Scale CTO Phil White. In this scenario, “edge” can mean small remote branch or office locations, including retail stores that need local compute resources delivered cost-effectively and efficiently. Scale has delivered thousands of clusters to customers using this model. The podcast is embedded here.

At the start of the pandemic, I had a conversation with Johan Pellicaan (VP and MD, EMEA at Scale Computing) on how the company was tackling home working with their VDI offerings. The HE150 (small-scale cluster) and HC3250DF were appliances directly aimed at solving VDI for companies that wanted quick remote access without huge infrastructure expense.

Focus

By focusing on small businesses and edge computing requirements, Scale Computing has built a profitable company. We’ve not seen Scale grow like the hyper-expansion of Nutanix, but that hasn’t mattered, as the focus for the company appears to have been constant growth to profitability. This goes against the grain of the typical Silicon Valley model, but then again, the company is headquartered in Indianapolis.

Future Plans

Where does the company go next? The most obvious approach for Scale is to build out further offerings that create an entire ecosystem for application deployment and operations. With this level of maturity, classic Innovator’s Dilemma rules state the focus should be upwards to larger enterprises.

However, a big push into the enterprise will be expensive and challenging in the midst of a continuing pandemic. In addition, as Nutanix has learned, many enterprises are reticent to abandon VMware in favour of open-source KVM solutions. So new vendors have to tread carefully.

My recommendation would be to fill some of the gaps in the existing products, namely:

Data Protection – build out native features within SCRIBE that enable backup vendors to integrate their solutions directly into HC3. Partner with at least one or two to build out certified solutions.
Data Mobility – resolve some of the issues of VM mobility by expanding the current SMB support of image importing and exporting. Both NFS and S3 are obvious starting points.
Visibility – expand on tools to manage and monitor multiple clusters. Consider SaaS as a solution for those customers with no desire to run internal management platforms. These capabilities could also be extended to automating the management of VMs themselves (e.g. applying patches and upgrades), including limited deployment catalogues. SaaS offers an option for building recurring revenue.
Support containerisation – look at options for supporting containers and Kubernetes. This option is quite challenging and likely to be driven by demand. I can’t imagine that edge customers are asking for this level of support, so this might be a “back burner” project.

The Architect’s View

It’s rare to come across a company taking time to build a profitable business without a race to IPO. Scale Computing has achieved success without classic fanfare and I look forward to seeing how the story of their next decade evolves.