IBM’s Power server line gained a single number update that involves four overarching feature enhancements that highlight the broader market focus on powering enterprise AI opportunities.

The first main update is for zero “planned” downtime for system maintenance. This feature relies on autonomous patching and automated workload movement that allows maintenance to happen without taking “critical” applications offline.

That autonomous patching taps into IBM’s Concert generative AI (genAI) automation platform for insight and AI-powered services from IBM’s Technology Lifecycle Services (TLS) infrastructure support provider to help with system optimization.

Power11 also uses IBM’s Power Cyber Vault platform, which follows the National Institute of Standards and Technology (NIST) cybersecurity framework, to provide protection against cyberattacks, and uses NIST-approved built-in quantum-safe cryptography to protect from “harvest-now, decrypt-later” cyberattacks.

These two features are also backed by some heady resiliency claims. Bargav Balakrishnan, VP of product management at IBM Infrastructure, explained during a press briefing that Power11 is “designed to be the most resilient server in history,” with six-nines of planned uptime and Power Cyber Vault providing less than one-minute guaranteed ransomware threat detection.

The Power11 AI angle also includes support from embedded on-chip acceleration for inferencing designed to scale through IBM’s Spyre accelerator chips when those become available later this year, the ability to tap Red Hat’s OpenShift AI for support across hybrid environments, and the IBM Watsonx Code Assistant operating system help manage use of IBM’s RPG programming language.

IBM also touted platform efficiency gains of twice the performance per watt compared with x86 servers and an Energy Efficient Mode that provides 28-percent better server efficiency compared to when the server is running in Maximum Performance Mode.

Enterprise and beyond

Tom McPherson, GM for IBM Power, explained during the press briefing that the offering is targeted as an enterprise server, but can support more robust deployment models.

“We actually have coverage from scale-up systems all the way through the scale-out and actually edge computing systems,” McPherson said. “And, in fact, a lot of the growth we've been seeing over the past few years has been in the in the scale-out part of our portfolio.”

This includes customers tapping hybrid-cloud models and deployment structures to support data sovereignty requirements. “It’s really letting clients have that flexibility and using essentially more of our portfolio out to the edge … to put AI acceleration at the edge, for example, that's one of the use cases we see with clients,” McPherson added.

The Power11 launch, which is scheduled for July 25, will also for the first time include availability of high-end, mid-range, and entry level servers as well as the vendor’s Power Virtual Server in IBM Cloud offering. That last platform is an IBM-hosted option that can act across hybrid-cloud deployments.

“Clients are asking for this, to have both [on-premises] and then onboard power virtual server. So, it’s more of an ‘and’ and not a migration to Power Virtual Server,” McPherson explained. “We see a lot of clients leveraging the hybrid capabilities and having Power Virtual Server deployed and also have an [on-premises] system.”

The launch also forms the third leg of what McPherson said was IBM’s “integrated value stack for AI and hybrid-cloud infrastructure.”

This includes IBM’s z17 mainframe launch from earlier this year. That unveiling had a specific AI support focus, with a second-generation on-chip AI accelerator built on the IBM Telum II processor, which the company said allows the z17 to process 50-percent more AI inference operations per day than its predecessor.

It also comes on the heels of IBM’s LinuxOne 5 launch, which is its Linux computing platform that runs on the Telum II processor and supports containerized workloads.

“We are deeply, deeply aligned and integrated with our IBM AI strategy,” Balakrishnan said of how these systems bolster the vendor’s stack. “To me, it's more aligned than ever, and it's really anchored on those three things that our customers are asking us to solve for.”