Large Scale Network-Centric Distributed Systems

15

A Cloud Broker Architecture for MultiCloud Environments

Jose Luis Lucas-Simarro, Iñigo San Aniceto, Rafael Moreno-Vozmediano, Ruben S. Montero, and Ignacio M. Llorente

Contents

15.1 Introduction

15.2 State of the Art on Cloud Brokering

15.3 Challenges of Cloud Brokering

15.4 Proposal of a Broker Architecture for Multicloud Environments

15.4.1 Broker Components

15.4.1.1 Database

15.4.1.2 Scheduler

15.4.1.3 VM Manager

15.4.1.4 Cloud Manager

15.4.2 Service Description

15.4.2.1 Service Components and Lifetime

15.4.2.2 Scheduling Parameters

15.4.2.3 Cloud Instance Usage and Instance Performance

15.5 Scheduling Policies for Efficient Cloud Brokering

15.5.1 Static vs. Dynamic Scheduling

15.5.2 Optimization Criteria

15.5.3 User Restrictions

15.6 Results

15.6.1 Cost Optimization in a Single Cloud

15.6.2 Cost Optimization in a Multicloud Environment

15.6.2.1 Infrastructure for a Virtual Classroom

15.6.2.2 Infrastructure for a Web Server

15.7 Conclusion

Acknowledgments

References

15.1 Introduction

In the last few years the cloud computing market [1] has experienced important growth. With the market expansion, the complexity is also increasing as users must deal with different cloud providers and their particular features (e.g., virtual machine (VM) types, pricing schemes, and cloud interfaces).

The cloud computing market provides users with different cloud providers, such as Amazon EC2 [2], CloudSigma [3], ElasticHosts [4], Flexiant [5], GoGrid [6], and Rackspace [7], which offer users ...