Requirements Analysis

ShanghaiTech University was founded in 2013. It is a young university jointly built by the Shanghai Municipal People’s Government and the Chinese Academy of Sciences, and is one of the core supporting institutions of the Zhangjiang Comprehensive National Science Center. At full capacity it is expected to have 4,000 undergraduates, 6,000 graduate students, 2,000 faculty, and 200 administrative staff. Because the share of administrative staff is far below that of traditional universities, the university needs an intelligent information-system project to help build a lean, efficient governance system. The goal of “lean administration” must be achieved through informatization: intensive construction and object-oriented functional design can build an intelligent university at high cost-performance; an energy-management system can save energy effectively; and behavior-management construction can strongly support rational resource allocation and leadership decision-making. Building leading intelligent informatization service infrastructure on mainstream IT, the university aims to avoid typical problems of traditional informatization such as long system-construction cycles, high IT investment, low equipment utilization, high total cost of ownership, many deployment steps, and long time-to-launch, thereby providing significant support for ShanghaiTech’s reform of higher-education systems and mechanisms and for building an innovative university.

The project draws on Sugon’s core strengths in cloud computing, high-performance computing, and foundation hardware and facilities. Sugon helped ShanghaiTech complete intelligent information-system construction facing the university, students, teachers, administration, university leaders, and partners. Starting from the university’s educational needs and centered on students and research results, it promotes deep integration of IT with teaching, research, and partnership resources, establishes a residential-college (tutorial) education model, builds a personalized, object-oriented informatized learning environment, and constructs an industry-university-research ecosystem service platform to support governance, talent development, and science and education innovation, continuously optimize resources, and inject momentum into a modern smart university.

To meet ShanghaiTech’s informatization goals, the project uses a phased strategy. This phase focuses on an intelligent data center, on which more than 30 business applications will go live over the next three years to meet day-to-day needs.

Solution

ShanghaiTech has two campuses: the Zhangjiang campus in Pudong and the Yueyang Road campus in Puxi. To date, most buildings on the Zhangjiang campus are in basic handover. In September 2016 the campus welcomed its first intake of new students. This project’s goal is to build the information-system data center on the Zhangjiang campus.

The system covers cloud hosts, cloud management, cloud desktops, cloud network disks, unstructured storage, structured storage, and a high-performance computing cluster. The overall architecture is as follows:

1. Cloud hosts and cloud-management platform

Given the problems of traditional data centers, ShanghaiTech’s data center is built with overall planning and phased implementation. Based on cloud-platform virtualization, IT infrastructure is gradually deployed in the cloud, virtualizing compute, storage, and network, sharing resources, allocating them flexibly, consolidating and coordinating business servers, and applying centralized, policy-based management to match ShanghaiTech’s rapid business growth while lowering total IT cost of ownership. Data are stored and managed in a unified way. The backbone uses 10-Gigabit fiber direct connections (40GE uplink to the core) so that business systems interconnect and information silos are resolved at the source. The new Cloudview system builds the virtualization infrastructure and cloud-management platform, providing centralized resource views, monitoring and alerting, and tiered configuration.

2. Cloud desktops

For office and teaching desktops, desktop virtualization shifts the traditional PC model to a cloud-desktop model: the personal computer desktop is separated from the physical machine through cloud computing and delivered as a “cloud + endpoint” desktop service. Compute and storage needed by the personal desktop are concentrated on central servers, replacing local compute, storage, and network at the client. Central-server compute, storage, and network are also shared and scalable; desktop resources are allocated and delivered on demand, raising utilization and lowering total cost of ownership. The design is zoned by scenario: an office zone, a general teaching zone, and a 3D-application teaching zone. For storage, the office zone uses external FC-SAN storage, while the teaching zones use a Virtual SAN (vSAN) distributed storage design to meet day-to-day business and innovation needs, with later expansion in mind: when vSAN space is insufficient, external cloud storage can provide more capacity.

3. Cloud network disks

The cloud network-disk system is built with Sugon Mobile Cloud Disk, helping users with work-file sync and backup, document-library management, file viewing and retrieval, file sharing and permission control, and other file-processing efficiency needs for faculty and researchers, while raising work-file transfer security and preventing data loss and leakage. It also helps students create personal learning spaces, share quality resources, and store personal resources. On top of server, network, and storage resources, it builds an easy-to-manage, easy-to-monitor, scalable cloud-disk system that meets the five-year plan for 15,000 users.

4. Unstructured storage

Cloud storage is built with Sugon ParaStor300 parallel distributed cloud storage to meet the customer’s 3.8 PB unstructured storage need. Sugon ParaStor300 is a high-end storage system designed for massive data. It uses a fully redundant architecture, with load balancing, redundancy, and self-healing at every stage to prevent storage faults from stopping service. ParaStor300 separates index and data paths so that each data controller can independently deliver bandwidth. Data controllers can be configured with the required number of disk drives and network interfaces, so each can provide high-bandwidth service. As index and data nodes increase, overall storage performance grows linearly with capacity. ParaStor300 also provides a browser-based management tool so administrators can complete deployment, monitoring, management, and maintenance through simple web pages, greatly lowering total cost of ownership.

5. Structured storage

This project uses vSAN distributed storage for virtual-desktop VM image storage, while structured data for the cloud-host cluster and business systems need an efficient unified storage system to meet growing business needs. The Sugon DS900 high-end disk array has powerful storage software that effectively raises system and data reliability. DS900 dual-controller active-active or four-controller active-active technology gives the customer more high-availability options.

6. High-performance computing cluster

The HPC system draws on Sugon’s leading position in high-performance computing and is built on the TC4600E blade system designed for HPC, Gridview cluster management and job scheduling, and high-performance four-socket servers to meet the customer’s HPC needs.

Customer Benefits

For ShanghaiTech’s intelligent information-system project, Sugon’s solution and product teams, based on the customer’s status and needs, fully used mature strengths in cloud computing, high-performance computing, and parallel storage. Starting from the university’s educational needs and centered on students and research results, the overall design promotes deep integration of IT with teaching, research, and partnership resources, establishes a residential-college (tutorial) education model, builds a personalized, object-oriented informatized learning environment, and constructs an industry-university-research ecosystem service platform to support governance, talent development, and science and education innovation, continuously optimize resources, and use modern informatization to move from traditional IT to an intensive IT model.