IBM Unveils New All-Hard-Disk Data Storage System

IBM just launched its latest data storage system, the Elastic Storage System 5000, for customers who require hard disk-based systems. The ESS5000 comes nine months after the company unveiled its ESS3000 all-flash storage systems.

The ESS5000 includes IBM Spectrum Scale software and is faster and boasts more density than the company’s previous all-disk model. A top ESS5000 feature is that customers can now run it with all-flash systems using the same software.

IBM’s Eric Herzog

Eric Herzog, chief marketing officer for IBM Storage, told Channel Futures that the company built ESS5000 for data lake use. The new arrays are great for big data, AI and analytics workloads, he said.

The ESS5000 ships with the same software as the ESS3000, which allows customers to run the systems together, he said.

“Some people need the performance of all-flash and some people need the capacity of hard drives,” said Herzog. “We mixed the all-flash variant and the all-hard-drive variant in the same name space. Customers can use the all-flash and hard drive systems together.”

A huge benefit of this capability is that it helps customers keep data together, he said. It also provides faster performance and better storage densities for opex and capex reduction, he added.

“Data silos are bad for AI and other demanding workloads because we’re talking about a lot of data,” said Herzog. “It’s easier to have a single name space, keeping it all in one repository, and avoiding data silos.”

The ESS5000 is faster and less expensive than the model it replaces. It has a density of 13.5 PB in 36 rack units. The previous array had a density of 8PB in 55 rack units. That means the ESS5000 requires less rack space, floor space, cooling and management, said Herzog.

The ESS5000 features bandwidth capabilities up to 55GBps, compared to 36GBps for the previous model.

IBM also announced a new Spectrum Scale software release for the fourth quarter. The company built Spectrum Scale Data Acceleration for AI specifically to help with demanding AI workloads.

The upcoming software will allow file systems to see objects as well as files at the same time, said Herzog. That will eliminate another reason that data gets siloed in data storage, he said.

“Some customers only use object storage, while some use file storage,” said Herzog. “This will allow the file system to actually see the objects. This allows consolidation of both for AI, big data or analytics applications from file stores or object stores.”

Channel partners and customers asked for the features in the ESS5000, he said.

“Partners and users said last time they needed an all-flash unit,” so IBM built it, said Herzog. “Here they said they needed an improved hard drive array. So we made the ESS5000 and improved it over the last model.”

Partners Optimistic About the ESS5000

Tom Bingham, sales director of U.K.-based MSP and channel partner CSI Limited, said the systems will offer additional customer capabilities.

“The tighter coupling with object storage and cloud workloads better meets our clients’ agendas as they seek to …

… move data to the most appropriate storage tier, seamless to the user and easier for the operations teams to manage. This extends CSI’s capabilities in delivering a perpetual edge to our clients and reiterates IBM’s value for price and performance.”

The product improvements will also help CSI help customers find new opportunities with hybrid and multicloud, said Bingham.

Another partner, John Zawistowski, a global systems solutions executive with reseller Sycomp, said IBM’s improvements are notable enhancements.

Syncomp’s John Zawistowski

“The ESS5000 brings additional storage capacity while maintaining the minimal rack unit space possible. Enhancements to throughput are also realized with IBM Power9 servers utilizing PCIe4 providing additional network port availability.”

The upcoming AI accelerator data storage software is also a positive move, he said.

“Adding AI Accelerator capabilities to IBM Spectrum Scale enables best-in-class data access for workloads that previously wrote data to an object store layer outside of Scale,” he said.

The moves show that IBM is truly listening to the channel and its clients, said Zawistowski.

“We have been asking for faster data platforms at reasonable costs to our end clients,” he added.

Portfolio Diversification

Dave Vellante is an analyst with Wikibon Research. He likes how IBM’s latest products focus on data-oriented workloads.

“This will help channel partners further diversify their portfolios and drive margin,” he said. “Everyone talks about being data-driven, but the reality is most organizations still aren’t. The pandemic has created a situation that if you’re not digital, you can’t transact business. Having new systems specifically designed for big data and AI apps that bring more automation is valuable. It gives channel partners another workload to target.”

Peter Burris, an independent storage analyst, said IBM is moving to meet customer storage needs wherever those requirements are.

“Storage technology no longer can force customers to move data to the device or cloud service,” he said. “Instead, a leading storage portfolio must be able to support data services where the data is created and used. IBM’s dramatically simplified storage portfolio does a good job of that.”

For partners, that means not having to convince customers about how to organize their data storage, he said. Instead, partners can apply IBM’s product and service portfolio to meet customer data needs.

Analyst Steve McDowell of Moor Insights & Strategy said the ESS5000 offers high-value enterprise engagements for IBM’s channel partners.

“It’s a win all the way around,” said McDowell.

The improvements to IBM’s Spectrum Scale software suite are particularly important, said McDowell.

“We’ve seen IBM dramatically ratchet up its software capabilities for object and unstructured data over the past two years,” he said. “This is no accident. Unstructured data is at the heart of modern analytics and AI. The ability to wrangle that data is a core challenge facing enterprise IT.”

The IBM Spectrum Scale updates are another example.

“The upcoming Data Acceleration for AI helps balance data between storage tiers,” said McDowell. “That’s one of the biggest challenges of hybrid cloud — keeping data where you need, when you need it.”