Huawei High Performance server helps Zhejiang Education Examination Institute

I. Customer background

Zhejiang Provincial Institute of Education and examination, mainly undertakes the province's national education examination and the enrollment work laws and regulations, the principle policy, formulates and organizes the province to carry out the education examination, the enrollment work concrete measure and the method; College enrollment and other work.

Second, the customer faces the challenge

With the expansion of the business scale of Zhejiang Educational Examination College, it is necessary to build a large number of business systems and undertake a large number of data calculation work, and the traditional decentralized rack servers and storage can not meet the business requirements in terms of computational density, reliability, manageability, green energy saving or performance. Therefore, the construction of a set of blade server + high-performance Rack-type server system is imminent. Blade Server + high-performance rack server, not only can solve high-density computing, management difficulties, a single point of failure, computer room energy consumption, data can not be shared, local expansion bottlenecks and other issues, but also to enable the performance of the business system has been greatly improved, the security of business data more effective protection.

Iii. Huawei Solutions

In order to achieve high security, high reliability, high performance, easy remote centralized operation and smooth expansion of the target, the use of High-performance servers to achieve Virtual desktop, application publishing, server virtualization and other requirements. The project plan is mainly considered as follows:

Resource pool design: Depending on the requirements of this project, the server is supported to install virtualization software and pool the server. It is easy to manage and monitor the virtual machine running on the server after pooling. Virtual machines can implement custom policy migration, manual thermal migration, and fault heat migration in the cluster. The design of resource pool has the characteristics of high reliability and smooth expansion.

Compute resource Pool

Compute a resource pool to provide users with CPU, memory computing resources. Virtualization software can be installed on the server to create multiple virtual machines on a single server and provide flexible virtual desktops.

Iv. Customer Value

With the development of technology and application, the server cost, performance, maintainability, reliability and other aspects of higher requirements, backward performance, high management costs, poor reliability and very low utilization of the user business has become a stumbling block. Tecal RH5885 V2 Rack Server (hereinafter referred to as RH5885 V2) is a high-performance enterprise server with independent intellectual property rights developed by Huawei based on Intel's latest processor technology combined with years of server experience.

RH5885 V2 can provide more flexibility and scalability than previous servers, provide more robust performance and reliability, and support a variety of applications such as database, virtualization, and memory computing by selecting different processing capabilities, memory capacity, and IO capabilities.

RH5885 V2 is a highly standard 4U rack server that supports a variety of specifications and can be easily upgraded to 4-way servers on a standard 2-processor basis, or two RH5885 V2 combined to 1 8-way systems. Memory supports 2 DIMMs to 128 DIMMs, according to a certain combination of rules, in an economically reasonable way to meet the needs of memory, and for your subsequent business development left enough to upgrade space, effectively protect the return on investment.

Tecal e9000h positioning for high-end computing platform, supporting high-end core applications, differentiated design for different needs and in the Hardware Computing platform RAS (reliability/Availability/serviceability), energy-saving emission reduction, life cycle, market access, intelligent control and service and other aspects of the overall upgrade; To provide the quality and service capabilities of small and medium-sized machines, to provide continuous competitiveness for enterprise software business, in the enterprise's general business, Internet services, to provide the industry with the common low-cost server competitiveness, but also for computing and data/media fusion business with High-bandwidth, low-latency Exchange, support computing and media integration.

Technical advantages

Performance Excellence

E9000 can support a full range of E5 Intel x86 CPUs and can match the evolution of future generations of CPUs.

Supports rich computing node specifications: 2s/4s and half width/Quanquan slot computing node flexible configuration.

The compute node supports a rich network interface and an accelerator card to meet a variety of high-performance requirements through the Mezz buckle card. Supports GPU,SSD PCIe extensions.

The whole machine supports half width 700w, full width 1400KW heat dissipation and power supply design.

10ge,40ge,100ge,ib Fde,edr Evolution.

The hard drive is arranged in an active drawer to support a single hard drive online.

Write Cache supports power off protection.

Dynamic Architecture Design

Modular design of computing, storage, switching, cooling and power supply.

Provides a 2p,4p,8p blade server dynamic Extension schema.

Medium, Small, machine cabinet series of evolution, support business smooth expansion.

Rich Exchange Module Ge,10ge,fc,fcoe,ib, flexible configuration according to business requirements.

Green and reliable

Non-State computing, physical machine migration, support for critical application high availability.

The use of High-performance Platinum AC power supply, power conversion efficiency of up to 95%, support dynamic energy-saving management power supply dormancy.

Optimized system duct design, thermal efficiency of the industry first.

Functional modules full redundancy, fully load balanced design, support the failure of seamless switching.

Passive backplane design to avoid single point of failure.