Abstract

Cloud computing infrastructures enable enterprises to acquire equipment and pay associated costs only when their business demands this, which is very attractive. However, there are notable challenges to wide use of clouds for enterprise services, such as security and privacy of data stored in clouds and reliability of cloud services.
Our research focuses on two subproblems in this space. First, we believe that the future of cloud computing lies in cloud specialization and differentiation. Future businesses will thus be allocating resources from different clouds, some public, some private, for a single application. Our research investigates how to perform automated, optimal allocation of these resources starting from (1) a description of distributed application components’ needs, and (2) a description of available cloud resources. Second, our research investigates how to offer multi-layer reliability monitoring in a manner that is application-transparent, platform-transparent, and real-time. This paper presents a language, called DADL (distributed application description language), that can be used to express (1) architecture, behavior and needs of a distributed application that may be deployed on clouds; and (2) specifics of available cloud resources. In our future work, this language will become input to our automated allocation and reliability monitoring systems. DADL is implemented as an extension of SmartFrog [8]–a framework for configuring, deploying and managing distributed software systems.

Date

February 7, 2026

Authors

Jelena Mirkovic, Ted Faber, Paul Hsieh, Ganesan Malaiyandisamy, Rashi Malaviya

Journal

USC/ISI Technical Report# ISI-TR-664