Traditional school computer classrooms are often difficult to manage and have relatively low operational efficiency. They can also incur significant costs due to factors such as maintenance and software usage. Therefore, Taipei University of Technology chose to implement NVIDIA's virtual display solution. The virtualization model makes computer classroom system management easier and allows for real-time adjustment of virtualization configurations to meet different teaching needs. Virtualization configuration also helps control software and hardware procurement costs and reduces management and maintenance manpower and time. It even provides appropriate hardware computing performance configurations for departmental research needs, which is helpful for recent big data analysis, high-resolution 3D image design applications, and the construction of artificial intelligence systems.
In conjunction with the construction of Taiwan's first enterprise-scale cloud computing room on campus in 2015, Taipei University of Technology has now virtualized most of its information systems. Through cloud-based management, all computer classrooms and equipment in departments and research units can operate in a virtualized mode, thereby achieving the goals of centralized management and energy and carbon savings.
Challenges: In the past, computer classrooms often faced difficulties in management and huge upgrade costs.
Wang Yongzhong, director of the Computer and Network Center at Taipei University of Technology, said that in the past, computer classroom hardware equipment in schools often required time and manpower to be reinstalled after teaching, and it was necessary to ensure at all times that no illegal software or inappropriate content was installed. Regular software and hardware replacement and updates often accumulated considerable expenses. Hardware maintenance between departments also often brought a lot of cost burdens, and there was even the possibility of errors in research data due to different system versions or hardware performance differences.
In addition, due to the differences in the needs of various departments, the hardware configuration is different, and the requirements for software system usage are also different. At the same time, if the handover of the construction and maintenance personnel is incomplete, it will also affect the difficulty of subsequent system maintenance, and even make it difficult to clarify the details of the management resources.
After introducing the virtualization operation mode, not only does it only need to be centrally managed through the computer and network center, but the terminal computers only need to be configured with simple hardware devices (such as processors, graphics cards, etc.), or they can be directly configured with virtual hardware and software through the virtualization platform without installing a hard drive or operating system. After the teaching course is completed, the original state of the original device can be immediately restored, and students and researchers can even continue on-campus homework or research content.
Solution: Improve and enhance usage efficiency and user satisfaction through virtualization solutions
In addition to building Taiwan's first enterprise-scale cloud data center on campus, which was completed in 2015, Taipei University of Technology currently has introduced the NVIDIA GRID virtualization platform solution based on Tesla M10 GPUs, which can support up to 220 sets of virtual computer devices and be used for professional 3D drawing software such as AutoCAD, MAYA, and Solidworks, as well as software services such as Photoshop that can process high-resolution image content. It can also improve research operation efficiency and improve the problem of school computers being easily infected by computer viruses in the past.
Wang Yongzhong explained that because the system is fully cloud-based, all incoming access is monitored in real time. Furthermore, firewall settings ensure that legitimate connections are allowed. Operating through a virtual machine effectively prevents internal computer system infections and minimizes security concerns. Compared to traditional standalone operation, operating through NVIDIA virtual display performance is virtually indistinguishable, with performance very similar. This further reduces hardware and software procurement costs and even helps achieve energy and carbon reduction goals. Nearly 90% of students expressed satisfaction with the virtualized computer classroom.
Impact: Driving new teaching models and more research and development opportunities
After introducing the new solution, in addition to providing a more convenient maintenance and management model, it can also save more in terms of software and hardware installation costs, power loss, etc., and it can also make the original school teaching content more diversified. Through cloud settings, different teaching content can be arranged according to the current needs of teachers, or various computing requirements can be built for different research. It can also allow students to continue computer learning courses at off-campus locations, or open computing resources for external rental, allowing computer and network centers to generate additional revenue at the same time, thereby continuously providing better cloud virtualized computing resources.
Overall, the introduction of NVIDIA solutions has resulted in a high level of satisfaction among faculty and students. The team anticipates continued procurement of Tesla M60 GPU solutions next year, as well as expanding the number of servers for use by more departments. Furthermore, the virtualization system will enable departments to invest in more research projects, and research funding will be used to assist the Computer and Network Center in building more virtualization system resources. This will have a mutually beneficial effect on overall research and application development.
NVIDIA Virtual Desktop Infrastructure will improve computer classroom teaching applications and management difficulties, and help departmental research projects use virtualized computing resources for data analysis and the development of artificial intelligence technologies for deep learning applications. Wang Yongzhong stated that the department will continue to expand the number of servers and purchase higher-end graphics cards, including Tesla M60, to build a computing environment that better meets research needs and achieve mutually beneficial results.
