Cost-aware orchestration of applications over heterogeneous clouds

Area of Science:

• Cloud Computing
• Distributed Systems
• Software Engineering

Background:

• Vendor lock-in is a significant challenge in distributed and cloud computing.
• Cloud application orchestration, particularly with standards like TOSCA, has advanced significantly.
• The practical adoption of orchestration is hindered by a lack of measurable benefits for consumers.

Purpose of the Study:

• To present a novel approach for cloud orchestration that integrates cost and policy considerations.
• To enable cost-aware application orchestration across heterogeneous cloud environments.
• To provide a compelling reason for cloud computing consumers to utilize orchestration by demonstrating measurable benefits.

Main Methods:

• Developed an approach combining an orchestration model with a cost and policy model.
• Incorporated consideration of the operating cost of applications on each cloud provider.
• Implemented forward projection of operating costs to ensure adherence to cost constraints.

Main Results:

• Demonstrated the capability to provide scaling features for applications.
• Showcased effective orchestration of application components distributed across heterogeneous cloud platforms.
• Validated the approach's ability to manage costs while orchestrating applications.

Conclusions:

• The proposed cost-aware orchestration approach enhances the utility of cloud computing services.
• Integrating cost and policy models with orchestration addresses practical adoption barriers.
• This method facilitates efficient and cost-effective application deployment across diverse cloud infrastructures.