Authentication with user driven web application for accessing remote resources

Abstract

With the emerging of data-driven analysis and data science, the needs of high-performance computing resources and cyberinfrastructure (CI) have been widely spread across almost all domain fields in academia. While CI providers have continued success with the infrastructure-as-a-service model (IaaS), there are increasing demands to offer more service models for diverse users and use cases. We have proposed a web application framework, which empowers users running their own analytic workflows as a web service that is accessible by other users authenticated with existing credentials. Through this framework, ad-hoc analysis routines can be described and wrapped in a workflow model that is shareable and re-useable. While the framework offers a new way for any user with valid credential to access and utilize remote resources, there are cases where more flexible authentication methods can further lower the access barrier of HPC resources. In this paper, we detail credential management of this framework and report our preliminary works on enabling temporary access to the remote resources including using dynamic generated credentials and social logins.