In July 2021, ZTE Corporation was rewarded Global Product Innovation Award by Frost & Sullivan, for its continuous innovation and leading-edge technologies position in the server industry. The award recognizes the company’s versatility in server technologies and great contribution to the overall development of the server industry. It also recognizes excellent features and benefits of ZTE x86 servers the company brings to customer, including ultra-reliable performance, scalability, cost savings, global reach and system applicable to a variety of application use cases.
In June 2021, China Mobile announces ZTE as one of the prime partners for the Engineering Computing Server Procurement of its Centralized Network Cloud Resource Pool Phase III, ZTE won the bid with a share of 60%.
ZTE has acquired the largest share of “general server” in China Mobile’s and ZTE feels proud to be selected as a partner for the third time. Furthermore, ZTE has already won two projects of China Mobile in 2019-2020 and 2020 for PC Server Centralized Procurement.
Industry networks, telecom networks and internet networks must be “Always-on”, hence reliability is essentially mandatory. However, telecom and internet providers have built their networks considering more test indicators for server performance, security, and reliability. Despite the strict requirements for carrier-grade servers, ZTE’s excellent product design, function and performance passes effortlessly the three-phase of centralized procurement test in China Mobile.
With its robust R&D support, strong overall performance and customer centric approach, ZTE earns the 2021 Sullivan’s award and new projects year on year.
How has ZTE become one of the communications equipment vendor and who can serve and develop a complete range of servers and storage products?
A multi-horse race – The competitive x86 server market
With the advent of the cloud era, x86 server has begun to lead dedicated servers such as mainframe servers and minicomputers. Moreover, cloud computing has new requirements related to software, hardware specification of servers and storage devices, which includes generalization, ultra-large scalability, virtualization, reliability, scalability, service by area, and intelligent maintenance, as well as low costs.
In early 2005, ZTE was the first vendor in China who has started server R&D and established the server & storage product line for related research. From the earliest OEM model to the self-research era since 2012, ZTE has launched server and storage products to the market, and swiftly achieved revenue growth above the industry average level. During past years from 2012 to 2020, the revenue of ZTE's server and storage products increased by more than 60 times.
Till date, ZTE has launched dozens of self-developed server and storage products, which are meeting various application requirements of cloud computing, big data, and high-performance computing, and have been widely used by the Internet providers, telecoms industry, finance industry and many other fields.
In the financial industry that has extremely high requirements for server security and reliability, ZTE has put intensive efforts to improve the security and reliability of server products by adapting and optimizing software and hardware, thus such products are empowered with strong competitiveness.
In addition, ZTE server products play an important role in smart city projects, epidemic prevention and control platform, smart transportation, and other fields.
Breakthrough - ZTE new-generation server for diverse demands
In April 2021, ZTE officially launched a new server product based on Intel's Xeon scalable processor, including ZXCLOUD R5300 G4X general server and R5500 G4X large storage server.
The 2-socket R5300 G4X server got the laurel title in two globally renowned SPEC CPU 2017 benchmark tests, and set a new world record in floating-point computing and integer computing performance test, verifying again its leading performance level. (Note: The data source is based on the test data of the G4X server released on the SPEC official website on June 8, 2021)
In terms of performance, the system efficiency of ZTE's new-generation G4X server has been improved efficiently. Compared with the previous generation of G4 server, it has made breakthroughs in aspects of general computing, memory capability and IO capability. The 2-socket server supports 80 cores at most, the integer computing performance is improved by 50%, and the floating-point performance is improved by 52%. Thus, it can provide super-strong general computing capability. Moreover, it provides 16 channels and 32 memory slots. The memory rate is increased to 3200MT/s, the memory broadband is increased by 47%, and the memory capacity is increased by 33%. The PCIe bandwidth is increased by 2.6 times.
More to the point, the G4X server is mounted with the third-generation Intel Xeon scalable processor which works as the data-center-level CPU integrated with the artificial intelligence (AI) acceleration capability. This processor is a new product developed by Intel after decades of innovation accumulation. The G4X server not only optimizes various workload types, such as cloud computing, AI, high-performance computing, network, and IoT, but also provides end-to-end data science tools and supports a huge intelligent solution ecosystem, helping to build a flexible and efficient intelligence platform.
Taking the application of AI as an example, the third-generation Intel Xeon scalable processor optimizes more than 20 mainstream AI algorithm frameworks and integrates the AI acceleration capability, meeting the requirements of complex AI scenarios such as image processing and video analysis with high performance, strong security, and operation control.
In addition, the G4X server has 4-8 built-in intelligent acceleration engines with heterogeneous computing, which can flexibly dispatch different heterogeneous computing resources and achieve the best combination of computing power. It can meet the requirements of various scenarios such as AI, NFV and industrial Internet, and give full play to the computing potential.
In terms of expansion capability, the G4X server, based on modular hardware design and unified software platform, is more flexible and can be configured on demand to meet differentiated requirements in various scenarios.
In the virtualization scenario, the G4X server supports a maximum of 80 cores, which can effectively increase the number of virtual hosts that can be deployed on a single physical server. At the same time, the G4X server has 32 DDR4 memory slots, meeting the requirement of a VM for consuming a large amount of memory.
For the big data processing scenario, the G4X server implements fast analysis and processing of high-traffic data. With a maximum of 41 2.5” disk slots configured on the basis of strong general computing power, the G4X server can provide large-capacity storage. At the same time, up to 28 NVMe SSDs provide high-speed I/O interfaces to solve the bottleneck of slow hard disk access in traditional solutions. That effectively reduces latency while increasing data processing speed.
For the AI scenario, the G4X server can provide high-level heterogeneous computing by configuring up to eight T4 GPUs or four A100 GPUs, to accelerate algorithm implementation.
In terms of security and reliability, through built-in software protection extension (Intel® SGX) and encryption acceleration algorithm (Intel® Crypto Acceleration), the G4X server protects data and application privacy in real time, effectively reducing the impact of encryption on performance. Moreover, the G4X server uses such technologies as external trusted module and trustworthy execution to ensure the security of server operation.
In terms of hardware, the G4X server effectively improves the equipment reliability by optimizing the layout of the main board, power module, intelligent heat dissipation management and intelligent fan speed control. The availability of the whole server system can reach up to 99.999%.
In terms of energy saving, on the one hand, the G4X server adopts the solutions of direct city power supply and HVDC to improve the overall energy efficiency to 96%. On the other hand, compared with the previous generation of platinum power supply, the titanium power supply the G4X server is 2% more efficient. At the same time, according to different configurations and scenarios, the G4X server provides a variety of power supply options. Users can select the power supply solution with the best cost-performance ratio as needed, so as to fully meet the energy saving requirements in differentiated scenarios.
ZTE server accelerates towards the future
With the further global digital transformation, the server and storage market will continue to develop at a high speed. It is estimated that the global server market size will reach US $25.7 billion in 2024 and the storage market size will reach US $4.8 billion.
With the rapid development of cloud computing, AI, big data, and 5G technologies, a variety of scenarios have been created, so servers have to address new requirements for powerful computing performance, massive storage, high bandwidth, and low latency.
10000 + rigorous tests have polished out ZTE's high standard server products. In the transformation of new technologies, on the one hand, ZTE continuously promotes server innovation and creates customized server and storage products that closely meet enterprise development requirements with the artisan spirit. On the other hand, with an open mind, ZTE works with upstream and downstream partners to concentrate on optimizing product capabilities, improving product competitiveness, and achieving a win-win situation.
ZTE will continue its technological and ecological advantages, and continue to provide more secure and reliable products with better performance from the perspective of customers. In this way, ZTE will take steady steps to accelerate its development, and provide firm support and development momentum for digital transformation of enterprises.