HPE’s new ProLiant servers target cost-conscious enterprise data centers and third-party cloud service providers eager to deliver cloud-native workloads.
The new Arm-based HPE ProLiant RL300 Gen11 is a single-socket system that integrates cloud-native Ampere Altra and Ampere Altra Max processors. The new chip packs 128 cores and is designed to take advantage of the Arm processor’s power-saving capabilities, the company said at its HPE Discover conference this week.
“This system gives us an advantage in running certain workloads, especially cloud-native workloads, while delivering performance with lower power consumption, which translates into better cost efficiency,” said Neil MacDonald, executive vice president and general manager of the compute business group at HPE, during a press briefing.
Ampère’s Arm chips specifically target cloud workloads and are also used in Oracle Cloud Infrastructure.
HPE’s new system is designed for IT stores with existing cloud-native workloads, MacDonald said.
“It’s not a platform for the mainstream business with classic off-the-shelf enterprise software; that’s not our target,” he said. “Increasingly, however, companies are adopting more in-house development that needs to take advantage of the technologies associated with cloud-native development.”
The only way enterprise data centers can grow over the next few years is to introduce more cost-effective systems that don’t sacrifice performance, said Renee James, president and CEO of Ampere Computing.
“You won’t see users saying ‘more data centers’ because they can’t afford the power source costs,” she said. “They can only achieve their ESG goals if they can improve the power-to-performance ratio.”
While Ampere chips are optimized for cloud-native workloads, users of these workloads can also use the new HPE system as general-purpose chips capable of running Windows and Linux-based applications, AI-based applications, modern databases and gaming software, James added. .
Dana GardnerSenior Analyst, Interarbor Solutions LLC
Some analysts say that integrating cloud-native support into silicon might prove to be a prudent move at this stage. They believe such an approach will find an audience among those who face not only increased energy costs in large data centers, but also the ongoing costs associated with cloud computing.
“Cloud computing costs are still expensive and difficult to control, and we’re about to enter tough economic times,” said Dana Gardner, principal analyst at Interarbor Solutions, LLC. “You’re seeing serious moves from companies like HPE and Dell with servers designed to not only fight cloud costs, but total cost of ownership.”
The new ProLiant system represents a commitment to Arm-based servers, which many HPE competitors are doing.
“HPE recognizes that this can’t be the company that doesn’t have some level of commitment to Arm,” said Dan Newman, principal analyst at Futurum Research and CEO of Broadsuite Media Group. “If they don’t, they run the risk of alienating a subset of its customer base.”
The RL300 will ship with the company’s Integrated Lights Out (iLO) server management software that allows users to configure, monitor and update servers from any remote location.
Other system features include support for up to 16 DIMMs with the ability to store 4 terabytes per system, three PCIe Gen4 expansion slots, and direct-attach support for NVMe storage with up to 10 NVMe SSDs and two NVMe M.2 SSD options.
The new server, the first in a line of systems that will follow, focuses on cloud-based services including media streaming and financial services on IaaS, PaaS and SaaS platforms.
The RL300 Gen11 server is expected to be available as a service through HPE GreenLake or through HPE’s partner network by the end of the third quarter of this year.
Pricing for the system was not available at the time of publication.
As Editor-in-Chief of TechTarget’s News Group, Ed Scannell is responsible for writing and editing breaking news, news analysis, and articles focused on technology issues and trends affecting IT professionals. business computing.