How to Get Cloud-Defined Storage Without Buying New Storage Arrays

Nebulon doesn’t think businesses need to buy big storage arrays to increase their critical data storage capacity.

Todd R. Weiss

June 23, 2020

5 Min Read
Cloud storage icon

Nebulon has a new way to do mission-critical data storage using its just-launched cloud-defined storage services.

Managed by Nebulon’s cloud, the company is providing on-premises, server-based data storage using server hardware that customers already run. A secure, cloud-based control plane and special Nebulon PCIe cards that make it happen are keys to the infrastructure. The PCIe cards are services processing units (SPUs) that install inside customer application servers. There they attach to internal server SSDs, much like a RAID card.

The system then provides data services such as compression, encryption, deduplication, erasure coding, snapshots and data mirroring from Nebulon’s cloud.

This offloads storage processing, memory and other resources to the PCIe cards, which reduces loads on the servers, thus improving performance. Nebulon works with any server operating system or hypervisor and requires no software on the server.

As a result, customers get enterprise-class storage without having to buy additional, costly new server hardware, Nebulon COO Craig Nunes told Channel Futures. The PCIe cards are in customer servers on premises, giving them direct control for users with compliance concerns. The customer data stays on premises and not in the cloud.


Nebulon’s Craig Nunes

“What’s in the cloud is the Nebulon control plane,” said Nunes.

It’s similar to a Nest thermostat in your house, where it acts like a services processing unit through the cloud. The Nebulon cloud-defined storage system uses its PCIe cards as services processing units for storage.

Less Hardware

“Some CIOs told us they were tired of having to buy more server arrays,” said Nunes. “They asked why they needed something separate from the many servers they’d already bought.”

Unlike alternatives such as hyperconverged infrastructure (HCI), Nebulon’s design consumes no CPU memory or other server resources, he said.

“They are PCIe-based cards we developed and provide to server vendors,” which customers can get when they order servers. All needed CPU, and memory and networking are built into the cards.

“What we’re doing is, your server is your server,” said Nunes. “You don’t have to share it with the data services like HCI. When you update your software on HCI, it will stop your data storage operations for a while. But using Nebulon, in this environment you can take your server down and the data remains available. You don’t have to buy 30% more servers to do what you were doing with your arrays.”

Nebulon is going to market through large server OEMs, including HPE and Supermicro, which already work with channel partners. Customers can work with their existing partners or with partners who deal with the appropriate server vendors.

Partners who sell Nebulon will benefit because they can keep cloud products in their conversations with customers, said Nunes.

“And it still allows them to fulfill customer needs on-premises,” he added. “It’s also an opportunity to talk about innovation with customers.”

Resellers will have opportunities to consult with clients about digital transformation and lower-cost, cloud-defined storage, he said.

The idea for moving the power needs from the servers to PCIe cards is not new, said Nunes.

“This has been going on for a long time in graphics processing. It is happening, but just not in storage until now,” he said.

The new services will be available to customers sometime in …

… September.

A group of ex-3PAR executives founded Nebulon, which came out of stealth mode on June 23.

An MSP’s Take

Marc Pratt, sales manager for U.K. and South Africa-based MSP and reseller SYSDBA, said he sees promise in Nebulon.


SYSDBA’s Marc Pratt

“It’s an alternative for customers who have seen the promise of HCI but have been disappointed with it,” said Pratt. “Being able to manage storage from the cloud creates new opportunities for partners. This creates services that monitor and maintain storage for customers in ways that have not been possible with traditional storage.”

Time is running out to apply for the MSP 501, a list of the biggest and most successful MSPs in the world. The deadline is June 30, 2020, so get your application in now.

The Nebulon technology provides some enterprise SAN features, without the architectural penalties in products from competitors, said Pratt.

“With Nebulon you get the cloud storage solution, but have the compute thrown in for good measure,” he said. “It’s a compelling value proposition. Startups that address real-world issues are exciting. We think we can solve real customer problems and carve out a very special niche.”

An Interesting Approach, Say Analysts

Camberley Bates, an analyst with Evaluator Group, said Nebulon is talking an interesting approach to help customers with cloud-defined storage.


Evaluator Group’s Camberley Bates

“Key to their go-to-market will be through their server OEMs,” she said. “I do not expect customers to crack open a server to add the services processing unit PCIe cards.”

For customers, lower data storage costs and better performance are the values in Nebulon’s approach.

“I believe we will see this adopted into the newer NoSQL database environments and used for data analytics apps. There are others, but think this is the first entry,” said Bates.

Analyst Stuart Miniman of Wikibon Research said bringing the cloud operating model to the data center is not new.

“For a long time, private clouds haven’t met the operational simplicity and flexibility of public clouds they are measured against. Solutions like Nebulon provide services and management of cloud storage in the flexible deployment of a server-based solution.”

This provides an easy transition for the channel, starting with HPE ProLiant and Supermicro servers, he said. Nebulon can provide a comprehensive on-premises storage solution without the complexity of a legacy storage array.

“It’s another step to enable storage with less dependence on expensive hardware platforms that are challenging to deploy and maintain,” Miniman explained.

Read more about:


About the Author(s)

Todd R. Weiss

Todd R. Weiss is an award-winning technology journalist who covers open source and Linux, cloud service providers, cloud computing, virtualization, containers and microservices, mobile devices, security, enterprise applications, enterprise IT, software development and QA, IoT and more. He has worked previously as a staff writer for Computerworld and, covering a wide variety of IT beats. He spends his spare time working on a book about an unheralded member of the 1957 Milwaukee Braves, watching classic Humphrey Bogart movies and collecting toy taxis from around the world.

Free Newsletters for the Channel
Register for Your Free Newsletter Now

You May Also Like