Shih Chien University Drives Cloud-Based Training on AWS

By migrating to the AWS Cloud, Shih Chien University (USC) expects to save 20 percent on labor costs, lower utility bills by 25 percent, improve data security, and enhance its reputation as a leading-edge technical institution. USC instructs 15,000 students as well as IT industry professionals from its campuses in Taipei and Kaohsiung. Its Taipei campus uses Amazon S3 for data lake storage, Amazon Redshift as a data warehouse, and AWS Identity and Access Management to help protect its data. 

Situated within 5 kilometers of the Neihu Technology Park and Nankang Software Park in Taiwan, Shih Chien University (USC) Taipei has a reputation for leading technology development. In addition to instructing students, USC hosts five courses each year designed for employees in the tech industry to upgrade their skills.

“Taiwan’s future is in the electronics and computer industry, and USC is a training ground for technology,” explains Meng-Huang Lee, dean of library and information services at USC. “We offer a practical, hands-on approach to technology that students can immediately apply in a business setting.”

USC has a history of working with Amazon Web Services (AWS). The university introduced a cloud computing curriculum through AWS Academy, and in 2017 it decided to migrate its on-premises campus information infrastructure to the AWS Cloud. Now the university has seven classrooms on its Taipei campus dedicated to cloud-based instruction, and each classroom has 60 computers that link directly to Amazon Elastic Compute Cloud (Amazon EC2) instances.

As of March 2020, USC had migrated about 20 percent of its workloads. The university’s goal is to move all of its data centers by the end of 2022. Some of the workloads currently being migrated include an online chatbot and an AI-based facial recognition system developed by USC engineers to automate access to university libraries.

The university uses Amazon Simple Storage Service (Amazon S3) as a data lake for storing de-identified information on the 8,500 students and faculty on the Taipei campus, and for archiving medium for large files and learning videos. It relies on Amazon DynamoDB, a fully managed, multiregional, NoSQL database service, for its facial recognition system. Amazon Redshift powers USC’s data warehouse to run queries against its data lake.

System engineers are pleased with the scale and performance they’ve experienced running on Amazon Redshift, particularly given the low cost of the service. “On-premises data warehouses can be very costly, and we were hesitant to pay in advance for something that may or may not be useful in our case,” Lee says. “Building a data warehouse on the AWS Cloud is a big benefit, as we save a lot using the pay-as-you-go model. In comparison to on-premises servers, the performance is much more stable, and it’s guaranteed even when we need to add more data for processing.” Lee’s team is confident in its ability to easily scale its data and analysis platform when university enrollment increases.

The pay-as-you-go model is especially useful in the university setting, where obtaining funds for capital investments can be a complex, lengthy process. “It was taking at least two months to acquire resources for a new project,” Lee says. “Because we’ve converted our infrastructure to an operating rather than a capital cost, we’ve been able to save time and simplify processes.”

In addition to AI workloads, USC has migrated several departmental websites and learning sites used by students to the AWS Cloud, running Amazon CloudFront for content delivery. At first, some stakeholders were reluctant about putting data on the cloud for security reasons. To solve this, USC uses AWS Identity and Access Management (IAM), which controls access to online resources. The university also relies on the implementation of security best practices and encryption controls inherent in services such as Amazon S3.

Since its migration, USC’s stakeholders have come to recognize the AWS Cloud as a highly secure means for storing and processing data. Leveraging cloud capabilities also puts the university at the forefront of collegiate competitors. “Cloud computing has really become a buzzword in Taiwan these days,” Lee explains. “Our stakeholders recognize that embracing the AWS Cloud will definitely give our university a competitive edge over other institutions.”

Lee’s IT team is likewise passionate about the potential of the AWS Cloud. Among the 12-person team, two have already been certified as AWS Certified Solutions Architect –Associates, a certification that can be attained after one year of experience on the AWS Cloud. Another engineer is an AWS Certified Cloud Practitioner, an entry-level certification that’s granted after six months of working with AWS Cloud services. The other team members have attended AWS Cloud Foundations training. “Our staff noticed that many new companies in the neighboring technology parks are already using the AWS Cloud, so upgrading their skill set improves their marketability as IT professionals,” Lee adds.

From a management perspective, university leaders are pleased with the projected cost savings once migration is complete. The team estimates a 20 percent savings in labor costs from eliminating server maintenance and a 25 percent savings on utilities by 2021. Furthermore, the cost of setting up and running a new instance on the AWS Cloud is 65 percent lower than an on-premises server.

Lee was granted a larger budget in 2020 to continue building new cloud-based infrastructure. The budget allows for more innovation and experimentation by taking advantage of new services on the AWS Cloud. Lee is particularly excited about implementing Amazon Kinesis, which will improve streaming video data from the campus’s Internet of Things (IoT) system. USC prides itself on being a green campus and reducing energy consumption wherever possible. The current IoT system, which was designed to run on on-premises servers, monitors and automatically adjusts the temperature in every room. The IT team will start migrating the IoT system by October 2020 and are looking to enhance the system’s capability by integrating a real-time surveillance function.

The IT team is also planning to use AWS Lambda for running IoT code in a serverless environment. “In terms of development time for IoT projects, we expect to build applications two to three times faster,” Lee says. “AWS serverless functions enable speedy development compared to a server-based architecture. Migrating to the AWS Cloud will greatly improve our entire application development process.”

To learn more, visit aws.amazon.com/cloud-migration.