I recently read an interesting article about the efforts of the good people at Scale Computing to integrate their product into Google Cloud. This uniting of on premises and cloud is significant.
HCI has been wanting to move in this direction for some time. The players involved just weren’t able to make the necessary investments in R&D until now. That’s changing, but a lot of work still needs to be done to make this hyperconvergence a reality. Naturally, the desired outcome is to build a cloud OS that can span on premises and public with a common framework and toolset. I liken it to a “data center fabric.”
I’ve been using some form of the following slide since 2015, when I was working at Cisco on building a rack scale solution and investigating the possible software-defined players that could make this solution happen.
The analogy I use here is that back in 2007, I paid about $2,700 for an all-in-one laser color printer that allowed me to get rid of a flatbed scanner, a copier, a printer, and a fax machine. The first wave of HCI was just like that: a consolidation of appliances that made up the common data center virtualization stack.
When that all-in-one printer broke this year, I needed to replace it. Looking online I found that most of the same features could be had for a fraction of the price I paid 10 years ago. The hardware itself was commodity and almost irrelevant. The real value was in the ink.
I didn’t pay for the printer; I paid for the ink. And I will continue to do so because it is the limited resource that I need.
The limited resource for most organizations looking to take advantage of the hybrid cloud is a cohesive software stack with orchestration, automation, and management, combined and packaged together for the common enterprise customer — a data center fabric. That’s what I would put my money on for long-term success.
That said, HCI in its current form is still stuck between the hardware (the delivery vehicle) and the software stack (the ink). Packaging those items together provides the foundation for the next phase. It will be like the virtualization market was in 2004-2005: lots of dev/test.
Eventually HCI will mature and become commonplace. VMware and Nutanix are both pushing into this space as forcefully as they can, and the smart money says that others will follow suit if they have the architecture and means to do so.
It is my opinion that NetApp is uniquely positioned to leap past the first-wave HCI solutions we see rolling out (which rely on partnership with other hypervisor vendors), and into the cloud OS/data center fabric that will be the foundation of hybrid cloud that is approachable for the enterprise customer.
NetApp’s recent acquisition of Greenqloud is a good first step in building a cohesive management layer that can span on-premises resources and serve as a multistack broker for a hybrid cloud approach. The final piece of the puzzle is the NetApp Data Fabric suite of technologies and products that allow global data management and mobility. These technologies afford the customer choice in terms of data management and placement and removes the risk of locking data to one consumption model.
Oh, and NetApp HCI is available now. Go check it out!