SGER: A Scalable, Flexible, Secure Educational Computing Lab

University Of Colorado At Boulder, Boulder CO

Investigators

Abstract

Educational computer laboratories have evolved from rooms full of terminals connected to a large timesharing system, to rooms full of desktop computers connected to a file server. In the former case, the main computer provides both cycles and file storage; in the latter, the desktop computers provide the cycles and some of the file storage, while the file server provides more stable file storage. This approach does not scale well in universities where space and other capital improvements are nearly insurmountable barriers (and even if funded, take years to implement). The quality of educational labs for computer science students deteriorates due to these space constraints, economy of scale requirements, and to the rate at which laboratory computers become obsolete. A survey of our undergraduate students shows that they typically use their own computer for university tasks rather than using the lab computers. This is because they have newer equipment and a familiar computing environment. Unfortunately, it is rare for a student's computer to have regular file archiving; if the student's machine crashes, the consequences are dire. This project describes a new model for constructing an educational laboratory - one that provides a scalable, flexible, and safe computing infrastructure in which students can use their own machines in concert with a lab infrastructure. The infrastructure provides software-specific extensions to the students' computers, e.g., a particular version of a compiler or package, as well as a very large amount of file storage space that is systematically archived to support failure recovery. The model encourages the development of tailored environments to enable an instructor to create a virtual lab with a unique operating system (e.g., Windows .NET for web service programming instead of Linux), language environments, etc. Intellectual Merit: This study investigates the application of leading edge storage system research to an important practical problem. Although various aspects of the technology currently exist, others will be created along with the synthesis of all aspects to create a new approach to distributed computing that leverages large storage systems. Broader Impact: Universities are in a continual cycle of lab computer acquisition, deployment, and replacement. This approach shifts the "rapidly aging" aspects of computers to end users. These users can decide whether or not they desire newer, faster equipment to use with the infrastructure, or if they prefer to avoid acquisition to save money. The university saves space by providing a cafe -like atmosphere where students can use their equipment, rather than providing rooms full of rapidly-aging computers. More importantly, the infrastructure provides a robust and configurable infrastructure where it is possible to ensure that students have access to exactly the right tools (and software versions) for a particular course. Students benefit by having their personal computer become a component in the educational computing infrastructure (with appropriate authentication and authorization), and by having a large, safe, archived file system at their disposal.