Large Scale Network-Centric Distributed Systems

19

Addressing Open Issues on Performance Evaluation in Cloud Computing

Beniamino Di Martino, Massimo Ficco, Massimiliano Rak, and Salvatore Venticinque

Contents

19.1 Introduction

19.2 Benchmarking Approaches

19.2.1 HPC-Like Benchmarking

19.2.2 Benchmark Standards

19.2.3 Cloud-Oriented Benchmarks

19.2.4 Benchmark as a Service Approach

19.2.5 Considerations

19.3 Monitoring in Cloud Computing

19.3.1 What Should Be Monitored

19.3.1.1 Infrastructure as a Service (IaaS)

19.3.1.2 Platform as a Service (PaaS)

19.3.1.3 Service as a Service (SaaS)

19.3.2 How to Monitor

19.3.2.1 Supporting Tools

19.4 Attack Countermeasures in Cloud Computing

19.5 Conclusion

References

19.1 Introduction

Cloud computing is an emerging paradigm that allows customers to easily obtain services according to a pay-per-use business model. The current Cloud providers include Quality of Service (QoS) guarantees in their Service Level Agreement (SLA) proposals [1]. In general, they offer guarantees in terms of service availability and performance during a time period of months or a year. The provisioning contracts regulate the cost that customers have to pay for provided services and resources. Cloud service provider must pay a penalty if the customer's requirements are not satisfied. For example, the Cloud provider is liable to pay a penalty for service requests that are rejected due to unavailability of resources, or if the average service response time exceeds a fixed threshold. In order to support this model, ...