How ZeroStack plans to deliver 'self-driving' clouds

We hear a lot about autonomous vehicles and the benefits they can bring the world. ZeroStack wants to use that metaphor for IT infrastructure.

gear robotic hand touching human hand

ZeroStack is an IT management vendor. More specifically, it provides tooling that lets organizations turn bare-metal (i.e., physical) servers into cloud infrastructure. Basically ZeroStack -- like cloud platforms from OpenStack, Microsoft and others -- wants to make physical infrastructure work like the public cloud.

ZeroStack offers the operating tools and the self-service platform that helps customers make decisions about capacity planning, troubleshooting and optimized placement of applications. On top of that, ZeroStack offers a Bitnami-like integrated app store that allows organizations to offer their end users a "one-click" installation of popular applications onto their cloud platform. Finally, ZeroStack is integrated with public clouds to offer migration between on-premises and public cloud infrastructure.

Which all makes sense, right? Well, yes, but today ZeroStack is claiming a new launch, that of the "First Artificial Intelligence Suite for Self-Driving Clouds." Which makes sense from a semantic point of view (as in, I know what the individual words mean), but doesn't make much sense from an IT perspective. Does ZeroStack automagically automate the racking and stacking of physical servers? Does it miraculously make applications "cloud-ready"? What does "self-driving" mean?

Well, according to ZeroStack, it has a phased three-part strategy to disrupt the economics of cloud. From the briefing materials:

  • First, a suite of capabilities that back up the self-driving claim with three levels of "cognitive" capacity planning, a state machine to perform "hands-free" upgrades, and "intelligent" optimization, managing virtual machine (VM) sizing. These capabilities guide the user to more optimized use of infrastructure.
  • Next, ZeroStack will release a cloud optimization suite meant to help broker which cloud the workload should be on for best cost/performance metrics.
  • The third component is automated performance troubleshooting. The capability will help identify the root cause of application performance, such as a VM with high transmit bandwidth, indicating a potential security issue.

So it seems this is a strong A.I.-driven automation play. The ZeroStack platform leverages self-healing software and algorithms developed from a quoted 1 million datagrams with a promise to make deploying, running and managing on-premises cloud as "hands-off" as using a public cloud. The platform collects telemetry data and leverages machine learning to help customers make decisions about capacity planning, troubleshooting and optimized placement of applications.

"The future of the data center is A.I. because fewer and fewer companies want to manage any infrastructure. As a result, the responsibility to manage increasing complexity is shifting from the customer to the vendor," said Jim Metzler, principal analyst at Ashton, Metzler & Associates. "By incorporating A.I. technology into their software, ZeroStack is at the forefront of these tidal changes in IT."

Someone is a little bit ahead of themselves, I'd suggest.


It's interesting. "A.I." is one of the increasingly over-used terms in the technology industry. The first stage of ZeroStack's vision, the "right-sizing," automated upgrades and optimization tools, all make sense. But I'm not convinced that they're actually A.I. What they seem to be, to me, is a series of rules-based tools that offer admittedly useful -- but probably not revolutionary -- infrastructure management.

As for the second and third parts of the company's vision: Cloud brokerage and cloud spend choice optimization is a field that many vendors are currently involved in. And as for application performance troubleshooting, this is a space that many other vendors are involved in -- look to NewRelic, Boundary and Dynatrace for examples.

That's not to say that ZeroStack's offering isn't useful and interesting. I certainly is. But to suggest that this is a self-driving cloud, and truly A.I., is probably gilding the lily a bit.

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.

Copyright © 2017 IDG Communications, Inc.

IT Salary Survey: The results are in