Software-defined server upstart TidalScale Inc. is brimming with confidence today after securing a $24 million round of funding.

The Series B round was led by South Korean semiconductor firm SK Hynix Inc. with participation from previous investors, including Bain Capital, Hummer Winblad, Infosys Ltd. and Sapphire Ventures.

TidalScale has largely flown under the radar since its inception, but the interest from its backers stems from its software-defined servers that are said to be compatible with all kinds of applications and operating systems.

Software-defined servers are commodity hardware that run specialized software to interact with infrastructure such as the processor, memory, storage, input/output and other communication ports. The software unites the resources of multiple commodity servers into a single unit, so these can be used by resource-heavy workloads.

TidalScale’s SDS platform is specifically geared for customers with existing storage investments that need to access massive amounts of memory for workloads such as large-scale simulations and computational genomics, it said. The platform works by aggregating commodity hardware into a single “software-defined supercomputer” that’s able to provide the heavy resources needed for these tasks.

The platform delivers something called “inverse virtualization” by creating a single pool of dynamic random-access memory. It also virtualizes the central processing units, memory and I/O resources, making them available to the workload as required. Additionally, it packs a kind of distributed hypervisor called a “HyperKernel” that helps to optimize the flow of resources and share CPUs, memory, storage and I/O across multiple servers.

“A software-defined server allows you to create a virtual server of essentially any size by combining one or more physical servers connected via a standard Ethernet network,” Gary Smerdon (pictured), chief executive officer at TidalScale, told SiliconANGLE. “Unlike a traditional virtualization hypervisor, which allows you to run multiple system images on a single server, TidalScale’s inverse hypervisor lets you run a single system image across multiple servers so you have just the right amount of resources to handle your workload.”

Smerdon said TidalScale’s software-defined servers provide a faster, easier, more flexible and affordable alternative to traditional approaches to scaling, which include overprovisioning or rewriting code so it uses less server resources. The problem with overprovisioning, he said, is it means making investments in hardware that may not be fully utilized for several years. Meanwhile, he said, rewriting code is both difficult and costly.

“TidalScale offers a different way to solve these problems by using an organization’s existing systems to create virtual servers that are dynamically sized to fit the workload,” Smerdon said.

The next step for TidalScale, he said, is to propel his company from a pioneer to become the market leader in software-defined servers. That will involve amplifying the company’s sales and marketing efforts on application-specific use cases, plus new partnerships in the offing that will extend its reach into more enterprise environments.

“We’ve hired experts in Oracle Database and SAP HANA to help us demonstrate our value to those markets,” Smerdon said. “We’ll be expanding that focus into other use cases, of course, as we build out our team and broaden the scope of our marketing and sales efforts.”

Smerdon appeared on SiliconANGLE’s mobile livestreaming studio theCUBE during the 2017 Open Compute Summit, where he provided more insights on TidalScale’s approach to software-defined infrastructure:

Image: SiliconANGLE