November 20, 2014 | Written by: Rossella De Gaetano
Share this post:
Co-authored by Francesco Pedulla
Have you ever thought about how cloud computing could affect teaching at a university? As cloud solutions developers, we found ourselves discussing this topic at a recent social media workshop. In this post, we thought we’d share our ideas and experience. We’ll also show how we’ve implemented cloud solutions to solve common problems that universities face.
Our experience and implementation
We have been working together on a project that involved a case study for a public university. We needed to automate classroom setups for exams that required the use of a computer lab. This would help the exam process go more smoothly and reduce wait times for teachers and students. But how would it be implemented?
Cloud was the obvious choice.
There were a few key points to consider in the design:
- The possibility to reuse the existing lab hardware
- The possibility for the professors themselves to instantiate/remove their class on demand
- The ability for the students to bring their own device
The actual implementation was based on IBM Cloud Orchestrator as a cloud management tool for two main reasons: it uses OpenStack as a hypervisor manager (a few governments now require looking at open source software first), and it provides an incredibly easy way to organize and customize the offering catalog.
So, the lab technician became the person responsible for creating the virtual template with the software stack required for the class and creating the offerings to be consumed by the professors (using available services like “create a virtual class,” “delete the virtual class” and “customize existing Linux image”). Since we decided to use Cloud Orchestrator, the lab technician could actually customize images with the required operating system by adding the required software at deployment time through script packages.
The challenge was to use the existing hardware in the classrooms in some way. It was too old to use as a reliable host for the hypervisor, and we did not want to enter into the bare metal provisioning game since it did not provide the required speed, flexibility and resource optimization that was best for us. Also, in time we would have met issues in adopting new operating systems because of the age of the hardware.
By relying on very strong network connectivity, the existing hardware could be used just as a virtual desktop infrastructure (VDI) terminal running a lightweight Linux distribution, or it could even keep using an old Windows version to host a client of the remote desktop protocol (we used Virtual Network Computing). In this way, we achieved 100 percent lab hardware reuse.
The virtual classroom concept supports the course teaching and exam activities for the university.
Once the time for a specific exam is established, the professor can immediately schedule the creation of the class to ensure resource reservation. They can also wait until a few minutes before the beginning of the exam, click a button and have the class started.
With the same ease, the professor can either immediately schedule the class for removal after the end of the exam (or course) or click a button and remove it once the exam (or course) is over.
The actual content of the exam or course is automatically pushed into the deployed images by Cloud Orchestrator, and the results of the exam are automatically collected back on a central repository with a notification sent to the professor.
Authentication into the specific virtual machine can be done using the specific university user registry.
With this implementation, we could set up a classroom in minutes rather than a couple of days.
One of the main issues with the traditional lab was that students used it to navigate the web, frequently leading to cases of systems infected by viruses. With the approach we used, access to anything other than needed course materials can be denied (this is especially important for exams), and even if an instance gets infected it can simply be removed and generated in a few minutes.
But there were also some unexpected results of this implementation that went beyond what we originally had in mind.
For example, students often requested to continue using the virtual machine instance after the lesson to move forward with the exercises. This suggests an opportunity to let students use the “virtual” lab in a different way and from any location (such as from their own assets): no more waiting, no need to be on the university campus to run the exercise and no time restriction on the system availability.
But that’s another story, and we think it deserves its own dedicated post. Do you think cloud can help universities in other ways? Let us know in the comments below.