Cloud Abstraction Libraries: Implementation and Comparison

Udit Agarwal, Clemson University


A valid concern for organizations interested in cloud computing is vendor lock-in. How do I move assets if a cloud provider disappears or if I find a better option? A potential solution to this problem is the creation of cloud interoperability standards tosupport basic use cases of user authentication, workload migration, data migration, and workload management. Cloud interoperability would ease the migration of workloads and data from one provider to another. A common tactic for enabling interoperability is the use of open standards, and cloud abstraction APIs. This thesis explores the role of a cloud abstraction layer for cloud computing. The abstraction layer simplifies the manipulation of virtual machines in clouds for easy and controlled development and deployment of cloud services. Our work is to qualitatively and quantitatively evaluate Dasein cloud and jClouds against Amazon EC2 and Google Compute Engine. We believe that cloud standardization efforts should focus on finding common representations of user identity, workload (virtual-machine images), cloud-storage APIs, and cloud management APIs.