Intel’s two new server families are now available: The Intel Server System M50CYP family for mainstream workloads, and the Intel Server System D50TNP family for high-performance computing (HPC) and AI applications.
Introduced last month, thse server families both support the latest 3rd gen Intel Xeon Scalable processors. These server CPUs deliver up to 40% higher performance than the previous generation, according to Intel. They also offer built-in AI with Intel Deep Learning Boost, and keep systems safe with hardware-based security features.
Both server families also support Intel Optane Persistent Memory (PMem) 200 Series. It delivers up to 512GB of persistent memory, extending capacity for memory-intensive workloads.
This latest generation of Intel Optane PMem technology also outperforms the first generation by delivering an average of 25% more memory bandwidth, according to Intel. The same Intel Optane PMem 200 Series has also been optimized to run with the latest 3rd Gen Intel Xeon Scalable processors.
Hardware-based security is another feature shared by both new server families. This includes:
> Intel Crypto Acceleration: Speed data encryption with enhanced processing of AES, SHA and RSA/DH protocols.
> Intel Software Guard Extensions: Protect sensitive applications and data, and also protects against threats even if software layers or firmware are compromised.
> Intel Total Memory Encryption: Protect data and virtual machines with full physical memory encryption.
> Intel Platform Firmware Resilience: Protect firmware by first detecting corruptions, then restoring firmware to a known-good state.
Here’s a more detailed look at each of these two new Intel server families.
Mainstream server, better-than-mainstream innovation
The Intel Server System M50CYP family is designed to run as a data center’s primary server platform. These servers are also designed for high performance, a wide range of workloads, and easy deployment.
All that’s important because mainstream workloads have grown increasingly diverse. These workloads now commonly include hyperconverged infrastructure (HCI), storage, collaboration, database and business intelligence.
For this reason, today’s servers need to be adaptable, high-performing, secure and flexible. It’s with these needs in mind that the new Intel Server System M50CYP family was created.
What can your customers do with the Intel Server System M50CYP family’s advanced server technology? Quite a lot. These servers are designed for two broad categories of use cases:
> Cloud and analytics: IaaS, PaaS, web server, e-commerce, data analytics and AI
> Enterprise: HCI, database, BI, storage and collaboration
Other use cases include small and midsize business (SMB) and edge computing, as well as workloads such as email, remote desktop, web server and infrastructure.
Each Intel Server System M50CYP family member can take up to 2 of the latest 3rd gen Intel Xeon Scalable processors. The server family also offers up to 8 memory channels per CPU, up from the previous max of 6.
For storage, the Intel Server System M50CYP family supports up to 24 x 2.5-inch SSDs or up to 12 x 3.5-inch HDDs. The system supports 2 mid-plane switches, enabling up to 24 NVMe drives without giving up a PCIe slot.
There’s also a new storage controller that offers higher performance and throughput, lower latency, new security features and “tri-mode” operation. The latter means that a single controller can connect to SAS, SATA and NVMe drives.
Chassis for the new servers are available in both 1U and 2U form factors. The 1U handles SSD drives only. The 2U chassis comes in 2 versions: 2.5 inches of SSD-only, and 3.5 inches supporting HDD drives.
Finally, the Intel Server System M50CYP family has been validated and certified for leading cloud enterprise software. These tools include Nutanix Enterprise Cloud, VMware vSAN and Microsoft Azure Stack HCI. All are now available as Intel Data Center Blocks.
A powerful, flexible platform for HPC and AI
The Intel Server System D50TNP family delivers a complete hardware solution for HPC and AI. Target use cases include workloads in life sciences, manufacturing, energy, finance, weather/climate and materials science.
Under the hood you’ll find a single half-width board that supports multiple module types, full support for the latest 3rd gen Intel Xeon Scalable processors, up to 128 lanes of PCIe 4.0 connectivity, NVMe-based storage, and optional liquid cooling for compute modules.
Depending on your customer’s workload, the Intel Server System D50TNP family is available in any of 4 modules:
> High-performance compute module: Packs up to 8 CPUs in a single 2U chassis, and keep things from overheating with a choice of air or liquid cooling.
> Management module: Increases I/O and storage, plus up to 270W CPU air-cooled support.
> High-speed storage module: Adds 16 E1.L “ruler” SSDs with direct PCIe connections for the maximum speed and storage per module.
> Accelerator module: Supports up to 2 high-performance CPUs and 4 300W PCIe accelerator cards within a single 2U chassis.
The new server modules are supported by the new Intel Server Chassis FC2000 family, first introduced with Intel Server System S9200WK product. It’s designed for improved power, cooling and flexibility.
This 2U chassis sports a wide (8.5 in.) node tray that supports the D50TNP family’s new module configurations. Its I/O design provides module access from the front, power supply and fan access from the rear, and the maximum cooling airflow for the highest performance.
Are you a member of Intel Partner Alliance? Activate your membership or join now.