HPE Adapts New Server Hardware to Changing Cloud Models
Server hardware has taken over software-defined virtual machines handling data center workloads, but HPE is emphasizing the importance of hardware in these virtual operating models.
HPE created waves when it launched the next-generation ProLiant Gen11 servers with a flagship server based on Arm processors, which sent a strong signal that it was diversifying away from its historical reliance on x86 d ‘Intel and AMD.
The RL300 ProLiant Gen11 Arm Server was targeted at cloud-native applications, which benefit from high core counts and low-power scalability for Linux applications such as Apache and NGINX in containerized environments.
On Tuesday, the company reintroduced x86 processors to the ProLiant Gen11 portfolio, with new servers aimed at high-performance compute, storage, machine learning and virtualized environments. The servers, which will ship next week, will be based on AMD’s 4th generation Epyc processors, also known as Genoa.
The new x86 servers are aimed at workloads that require higher performance chips to execute more instructions per cycle, and facilitate multiple accelerators such as GPUs and FPGAs for high performance computing.
Before the advent of cloud computing, x86 servers were capable of handling applications such as databases, ERPs, and analytics tools deployed on-premises. Arm made its mark when developers moved the use of common tools like the LAMP (Linux/Apache/MySQL/Python and Perl) stack for web-based software to the cloud. Intel and AMD have adapted their x86 chips for cloud implementations of high-performance science and artificial intelligence applications, which rely on accelerators.
The new x86 ProLiant servers will ship on the same day that AMD officially announces its 4th generation Epyc chips on November 10. The servers feature faster throughput technologies with support for DDR5 memory and PCIe Gen5 interconnects. The servers also support CXL (Compute Express Link) version 1.1, a technology that facilitates lightning-fast communication between GPUs, memory, CPUs, and other components.
New offerings include the ProLiant DL385 Gen11 server, a two-socket 2U enclosure that supports up to 96 CPU cores, 6TB of DDR5 memory, and up to 48 SSDs. The enclosure is designed for machine learning, which typically requires setups that support accelerators such as GPUs. The system has eight PCIe 5.0 slots for GPUs.
The other ProLiant are variants that support up to 96 x86 cores and include support for PCIe 5.0, CXL 1.1 and DDR5 memory. The machines include DL365 Gen11, which is a two-socket 1U server, which supports up to 96 CPU cores, 6TB of DDR5 memory and 2 PCIe 5.0 x16 slots.
The DL345 Gen 11, which is a 2U single-socket server, and the DL325 Gen11, which is a 1U single-socket server, support up to 3TB of DDR5 memory. The DL345 has 6 PCIe 5.0 x86 slots and the DL325 has two.
The servers are designed for HPE’s hybrid cloud environments called GreenLake, which the company markets as a utility-like service where customers can pay as they go for compute needs. HPE has built a software environment within the model that supports applications such as vSphere from VMware, which dominates the virtualization markets, and Linux implementations from Red Hat and SUSE.
The new systems support the Apache Cassandra application for managing databases, which has been tested to deliver better performance per watt and cost savings over Arm processors. Part numbers are from Ampere, which supplies the Arm silicon for HPE’s ProLiant RL300 Gen11 server.
Enterprises can combine their public cloud and on-premises infrastructure into a single compute system with GreenLake. In some cases, HPE decides on the best hardware and software combination for customers. In such cases, the hardware is virtualized and invisible to customers.
But there are customers who specify hardware and processor requirements, said Krista Satterthwaite, senior vice president and general manager of consumer computing at HPE.
“We have customers who are looking for performance. They want the best possible performance because their workloads demand it. These customers are going to be using the latest and greatest infrastructure,” Satterthwaite said.
For example, a customer augments their video surveillance with in-store servers, and they purchase more space than they need to support accelerators such as GPUs.
“They’re like, ‘hey, forget what we need right now. We are building this for the future. When you have these generational changes, and they don’t happen every year, many close customers seize the opportunity to build for the future,” Satterthwaite said.
Purchasing models for HPE ProLiant servers have changed in recent years, with an emphasis on specialization based on workloads where the Arm-based server offering fits.
“While before, they would buy a DL380 and run everything on it. We see very different behaviors,” Satterthwaite said.
The x86 servers will serve the rest of GreenLake’s general purpose computing offerings, she said.
“We view the business as primarily what we have in our offerings for the rest of the ProLiant family,” Satterthwaite said.