Many enterprises today have prioritized cloud computing and virtualization due to their nature in running new applications in virtual environments. According to the RightScale 2018 State of the Cloud Report, 81% of enterprises have a multi-cloud strategy that covers private, public, and hybrid clouds. In recent years, the cloud computing markets have experienced explosive growth, meaning that enterprises that run mass cloud services are searching for a server deployment method that is cost-effective, scalable, and flexible.
Unlike common desktop computers, servers for cloud computing and virtualization must be able to provide two key features. The first is high availability, which is essential in mission-critical systems required to support 24/7 service continuity, such as bank transactions or flight booking, where even a minor failure can have major consequences. The other key feature is high scalability. In response to ever-growing service requirements, a server system needs to expand critical capabilities, such as the computing capacity, memory, storage, and I/O bandwidth.
Great strides have been made in almost every key component of a server. In line with an increasing quantity of transistors housed in a processor, the processor performance and execution speed improve every year. In particular, the past few years have seen the rapid development of the computing architecture of servers, as well as rising numbers of processor cores used in computing. For example, a single 2nd-gen Xeon® Platinum processor supports up to 28 cores for parallel computing. In addition, breakthroughs in storage technology, especially NAND flash technology, have contributed to larger capacity and faster data read/write of a single drive. Now servers run on drives of larger memory capacity and provide larger-than-ever storage capacity. In terms of I/O, innovations in the Ethernet, IB, and OPA technologies have helped achieved a 100 Gbit/s transmission rate, which is turbocharged by the parallel transmission of multiple I/O interfaces.
However, conventional DDR memory, which has been plagued by limited capacity and high prices, seems unable to catch up with the industry trend. High-performance and large-capacity memory are indispensable for the unprecedentedly high computing requirements, but any increase in memory density in a single server will see costs skyrocket. Consequently, the low-capacity and costly memory technology prevents more VMs or containers from being deployed in a certain space, which has put a ceiling on cloud computing workloads. The traditional method for expanding computing capabilities is scale-out, where more servers are deployed for larger memory capacity, but this is an imperfect method because of the subsequent high costs. As a result, the industry is looking for an affordable model to expand memory capacity.
These problems are answered by the Intel® Optane™ persistent memory (PMem), a next-generation non-volatile dual in-line memory module (DIMM). This technology boosts the storage media to 1,000 times that of the NAND Flash, and a storage density 10 times higher than DRAM, helping turbocharge the storage capacity of the Intel® Xeon® Cascade Lake Scalable processors.
Figure 1 Intel® Optane™ persistent memory (PMem)
By providing larger VM memory and more VMs in a single server, the Optane™ PMem boosts the server scalability while reducing the number of servers required. In addition, the PMem improves the computing density of rack servers, allowing more servers to be deployed in a single cabinet. This reduces power consumption, network infrastructure, cooling, and maintenance costs, and helps customers cut the TCO.
Huawei was one of the first to jump on this cutting-edge technology, and subsequently launched the next-generation FusionServer Pro intelligent servers running on the 2nd-gen Intel® Xeon® Scalable processors and PMem. Huawei FusionServer Pro intelligent servers equipped with the latest Optane™ PMem support three memory capacity specifications: 128 GB, 256 GB, and 512 GB per DIMM. Each processor (configured with 256 GB DRAM and 512 GB PMem) can provide memory capacity up to 4.5 TB, which marks a remarkable boost in processor memory capacity.
FusionServer Pro 1288H V5, 2288H V5, and 2488H V5 are three premium products for cloud computing and virtualization. With exceptional performance, these three models are supercharging a wide array of industries, winning high recognition from customers around the world.
FusionServer Pro 1288H V5 (1U 2S) FusionServer Pro 2288H V5 (2U 2S)
FusionServer Pro 1288H V5 and 2288H V5 support two full-power 2nd-gen Intel® Xeon® Scalable processors, 24 DIMM slots, and various local storage resources, such as SAS, SATA, and NVMe SSDs. The combination of 12 Optane™ PMems and 12 DDR4 memory modules provides up to 9 TB memory capacity. In terms of virtualization, the number of VMs and containers housed by the two models is two to three times higher than servers equipped with only DDR4 memory modules. The 1288H V5 and 2288H V5 run on the perfect combination of PMem and DDR4, offering high computing density and intelligent energy saving in an easy-to-use manner. The products improve the space utilization of data centers with high-density deployment, making them must-haves for any enterprise working in cloud computing and virtualization scenarios.
FusionServer Pro 2488H V5 (2U 4S)
In a compact 2U space, the FusionServer Pro 2488H V5 innovatively supports four full-power 2nd-gen Intel® Xeon® Scalable processors with up to 112 cores, 48 DIMM slots, and diverse local storage options, including SAS, SATA, and NVMe SSDs. It combines 24 Optane™ PMems and 24 DDR4 memory modules for up to 18 TB memory capacity, greatly improving the number of VMs and containers accommodated in a single server. For in-memory database virtualization, the FusionServer Pro 2488H V5 model can also provide larger memory capacity for the databases supported by VMs. In addition, the four processors of the 2488H V5 server running on the full mesh architecture deliver 14.9% higher performance than the ring architecture in HammerDB laboratory tests. The FusionServer Pro 2488H V5 server also ranks first in the SPECvirt_sc2013 test, which evaluates the virtualization computing performance of 4-socket servers. The results create a standard to compare the performance of computing-intensive applications, such as clusters and HPC. The extraordinary performance of the 2488H V5 server in the preceding tests helps it earn the recognition of customer, making it applicable to computing-intensive scenarios such as virtualization, cloud computing, and in-memory databases.
Huawei FusionServer Pro servers effectively reduce OPEX and ROI by innovatively adopting Fault Diagnosis & Management (FDM), Dynamic Energy Management Technology (DEMT), and Boot Speedup Storage Technology (BSST).
- FDM locates faults with an accuracy rate of 93% and predicts a fault 7 to 30 days before it occurs to reduce OPEX.
- DEMT reduces energy consumption by as much as 15% without compromising services.
- Products powered by the BSST are an ideal option for virtualization scenarios. Without occupying the resources of data drive slots, the BSST allows two M.2 SSDs to run on the PCIe card to install independent OSs. The two M.2 SSDs support hardware RAID and hot plug, an industry-unique feature that enhances OS image protection and facilitates O&M.
The continual improvement of Huawei servers in quality, reliability, performance, and energy efficiency have fueled their success over the past two decades. Innovative design and strict quality control continue to ensure the excellent quality of Huawei servers. Over the past 18 years of strategic partnership, Huawei has joined hands with Intel to improve server technology, creating a growing demand for Huawei FusionServer Pro servers. The shipment volume of Huawei servers exceeded 4.31 million units from 2012 to 2019. Huawei servers will continue to reach new levels with state-of-the-art technologies and ultimate services.
Over 10,000 global customers, including 211 companies on the Fortune 500, are using Huawei server products and solutions to supercharge services. Huawei is supercharging industries in public utilities, Internet, telecom, energy, finance, transportation, healthcare, education, media, and manufacturing. Huawei server products and solutions optimize TCO and help meet the market demand for higher computing power to fuel cloud computing, big data, HPC, and AI applications.
Learn more about FusionServer Pro.
Copyright © 2020 IDG Communications, Inc.