Advanced Computing in the Age of AI | Wednesday, May 22, 2024

Nebulon, Cloud Storage Startup, Emerges from Stealth 

via Shutterstock

A growing list of cloud storage startups continue to emerge as enterprises struggle to get a handle on big data while seeking to move computing closer to where data resides.

The latest entrant is Nebulon Inc., a Bay Area cloud-defined storage startup that emerged from stealth mode this week with an API-based approach designed to cut storage costs for hybrid IT.

Nebulon was launched in 2018 by early employees at 3PAR, the utility storage specialist acquired by the former Hewlett Packard in 2010. The startup’s web site lists former 3PAR CEO David Scott as Nebulon’s executive chairman. Siamak Nazari is Nebulon’s CEO, along with CTO Sean Etaati and COO Craig Nunes.

The startup is attempting to thread the needle between expensive storage arrays and server-based solid-state drives that fall short handling critical applications and hyperconverged infrastructure (HCI) requiring additional server software. The latter restricts virtual machine and application “density,” the startup asserts, along with limiting operating system and hypervisor choices.

Nebulon further argues that neither approach alone scales in terms of IT automation and self-service infrastructure, which it assumes most IT manager and app owners want.

The startup’s software-defined storage framework runs applications on servers but is promoted as consuming no server CPU, memory or networking resources. Instead, resource allocation is determined and managed in the cloud via what the startup dubs “services processing units” (SPU) linked via Nebulon’s multi-cloud control plane.

The SPU is a PCI Express storage engine installed within a customer’s application server, where it is attached to internal server SSDs. “Think of the Nebulon SPU like an array controller that has been refactored into a PCIe-based device in the style of a GPU,” CEO Nazari said in a blog post.

The approach is designed to reduce storage costs by running on commodity drives and servers. No addition server software is required that would otherwise consume CPU and memory resources to run applications, the startup said. “API-driven automation [makes] enterprise-class storage a simple attribute of the data center fabric with self-service infrastructure provisioning and storage operations as-a-service for application owners,” said Nazari

The goal is making enterprise storage another “part of the datacenter fabric,” he added.

The storage framework attempts to cover the waterfront in terms enterprise applications, including cloud-native distributed NoSQL databases to container applications running on Kubernetes and OpenShift. Also covered are server virtualization and single instance and clustered SQL databases running on bare metal.

Among the advantages of the cloud-defined storage approach is the promised ability to offload CPU costs associated with data preparation

The startup’s early datacenter infrastructure partners include Hewlett Packard Enterprise (NYSE: HPE) and Supermicro (NASDAQ: SMCI). HPE said it would combine Nebulon’s cloud-defined storage framework with its ProLiant servers as a means of automating storage requirements.

It will also run on Supermicro’s line of Ultra servers.

Nebulon, which has raised more than $18 million in two funding rounds, emerges as industry analysts predict growing demand for software-defined storage and backbone HCI components. As enterprise storage demand soars, the hyperconverged approach has made it easier and cheaper to add capacity without the need for highly-trained IT administrators.

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).