IBM's rack mount Z16 mainframe targets edge computing
IBM's new rack mount Z16 mainframe gives edge locations the ability to process workloads locally, taking the burden off systems running workloads in larger data centers.
Aiming to give users more flexibility in deploying mainframes, IBM delivered a new rack-mounted Z16 system containing the same capabilities of its standalone Z16, including AI inferencing and quantum-safe cryptography.
The single-frame and rack-mounted designs of the new Z16, available with LinuxOne Rockhopper 4 options, makes the system more practical for space-constrained data centers and edge locations where users look to process data locally, IBM said.
"This system's design was driven by users telling us they needed smaller form factors to make their edge computing capabilities more vibrant," said Marcel Mitran, IBM's CTO of cloud platform and IBM ZSystems. "Increasingly we are seeing the decentralization of IT infrastructure so users can bring compute power closer to where their data is, especially for hybrid cloud environments."
Users transferring workloads from their headquarters' data centers to edge locations also require stronger security, higher availability and more resilience -- something Mitran said the new Z16 can also offer, particularly for the healthcare and financial industries. For instance, the new system, like the standalone version delivered last September, has 99.9999999%, which is the equivalent of three seconds of downtime over the course of a year.
"There are still a lot of applications running on mainframes. But for users moving from data centers to hybrid environments, like remote offices, portability is a big deal," said Jack Gold, an analyst at J. Gold Associates, LLC. "The market is changing, and IBM has to change with it. It's all about compute power and flexibility these days."
IBM also focused on sustainability issues. The new system has partition-level power monitoring. If users choose to consolidate Linux-based workloads onto a single LinuxONE Rockhopper 4 instead of running them on x86-based servers under similar conditions, they can reduce energy consumption by 75% and floor space by 67%, according to IBM. The system's design also makes it easier to integrate it into hot or cold aisle thermal management data center configurations using common data center power and cooling.
The new unit fits into IBM's standard 19-in. rack and can be used with the company's existing Power server units. The new footprint opens up the possibility for an IBM mainframe to work in distributed environments with servers, storage units and SANs made by other vendors, with all of them housed in the same rack.
"What's interesting with this announcement is IBM didn't target the system at a specific market like most of its mainframes. It is positioned more as a general-purpose system," said Frank Dzubeck, president of Communications Network Architects. "This gives (IBM) more opportunity to sell it more as a next-generation mainframe to a broad array of industries."
Other IBM system Z mainframes have been available as single frame systems, including Z14 and Z15.One area of concern could be how IBM prices the new system. Given its technical specifications, it is sandwiched between the full-blown Z16 on the high end and IBM's Power servers on the low end, which are also available in rack mount versions.
IBM plans to provide pricing options for varying uses, Mitran said.
"We can make it available on the IBM Cloud where users can consume just one core," Mitran said. "Or we can offer the LinuxOne Express version where we can offer a pre-configured system that better matches price with features. For larger systems, the pricing becomes more fit for purpose."
Jack GoldAnalyst, J. Gold Associates, LLC.
Containing IBM's Telum processor, the system has on-chip AI inferencing along with a new version of the operating system, Z/OS 3.1. The new version of the operating system is now fully enabled to support AI-based applications, Mitran said. Whether the system is deployed in a rack mount, single frame, multi-frame configuration, users can train and deploy AI models close to where their data resides making it easier for users to optimize AI-based applications, IBM said.
Asked if IBM had plans to explore the possibility of incorporating ChatGPT -- the current hottest topic in the tech world -- into its longer-term strategies, Mitran said he was impressed with the technology but has no specific plans for it.
"ChatGPT is a once-in-a-lifetime sort of disruptive technology," he said. "We have been doing a lot of work with our Foundation Models, which, in some sense, is a superset of the [Chat] GPT models. You can think of GPT models as a specific flavor of our Foundation Models. Our [mainframe] systems could certainly run ChatGPT very well."
As Editor At Large with TechTarget's News Group, Ed Scannell is responsible for writing and reporting breaking news, news analysis and features focused on technology issues and trends affecting corporate IT professionals.