Cloud performance and placement in cloud brokering 

Get Complete Project Material File(s) Now! »

Studies related to Cloud providers performance evaluation

An exhaustive study about the academic approaches of commercial Cloud services evaluation has been carried out by the Australian National University [LZO+13]. A Systematic Literature Review (SLR) was the methodology employed to collect the relevant data to investigate the Cloud services evaluation. As a result, 82 relevant Cloud service evaluation studies were identified. The key findings of this study represent a state-of-practice when evaluating Cloud services and are as follows:
• 50% of the relevant studies investigated applying Cloud computing to scientific issues, while only 16% of the studies focused on the evaluation of business applications in the Cloud.
• 21 Cloud services over 9 Cloud providers were identified. 70% of the relevant studies evaluated Cloud services provided by Amazon Web Services (AWS).
• Three main aspects and their properties for Cloud services evaluation have been investigated: performance, economics and security, performance being the most studied aspect (78 studies).
• There is no consensus regarding the definition and the usage context of metrics. Some metrics with the same name were used for different purposes, some metrics with different names were essentially the same. The study identified more than 500 metrics including duplications.
• There is a lack of effective metrics vis-à-vis elasticity and security aspects in Cloud computing. Therefore, it is hard to quantify these apects.

Placement in Cloud brokering

The placement or resource allocation in Cloud brokering refers to the mechanisms to distribute infrastructure resources across multiple Cloud providers based on end-users’ needs and constraints. The optimization goal in placement is to select a single or a set of Cloud providers to optimally deploy a service based on an optimization criteria, for example cost optimization or performance optimization. Placement mechanisms can be classified into non-functional requirements-based placement and application-aware placement. The non-functional requirements placement corresponds to the allocation of Cloud infrastructure based on the match of both Cloud provider resources and end-user requirements. The application-aware placement is based on the constraints that guarantee a Quality of Service (QoS) of the application running on top of the infrastructure.

Memory benchmarks performance

RAM and cache memory bandwidth have been measured with the Stream and Cachebench benchmarks, respectively (Figure 3.7). Stream is a simple synthetic benchmark program that measures sustainable memory bandwidth and the corresponding computation rate for simple vector kernels. In our performance evaluation, Stream was set up to measure the memory bandwidth through the copy and add operations. The copy operation consists of fetching two values from memory and updating the value of one of these fetched values with the other. The add operation fetches three values from memory and updates one of the fetched values with the sum of the other two fetched values. Cachebench is a benchmark designed to evaluate the performance of the cache memory present on a system. In this performance evaluation, CacheBench was set up to measure the cache memory bandwidth through read and write operations. In general, CacheBench results show the writing speed is around 60%-80% faster than the reading speed (Figure 3.7e).

READ  Routine-Biased Technological Change and HoursWorked over the Business Cycle 

Table of contents :

List of Figures
List of Tables
1 Introduction 
1.1 Motivation, objectives and thesis outline
1.2 Contributions of this thesis
1.3 Publications
I Value-added services in Cloud brokering 
2 State of the art: Cloud performance and placement in cloud brokering 
2.1 Introduction
2.2 Cloud performance evaluation
2.2.1 Motivations and challenges
2.2.2 Studies related to Cloud providers performance evaluation
2.2.3 Cloud Virtual Machine (VM) characterization
2.3 Placement in Cloud brokering
2.3.1 Non-functional requirements-based placement
2.3.2 Application aware placement
2.4 Conclusion
3 Towards a figure of merit of Cloud performance 
3.1 Introduction
3.2 Performance evaluation
3.2.1 Evaluation methodology
3.2.2 Experimental setup
3.2.3 Provisioning time
3.2.4 Computation benchmarks performance
3.2.5 Memory benchmarks performance
3.2.6 Storage benchmarks performance
3.2.7 Variability
3.3 Figure of merit of VM Cloud performance
3.3.1 Mean and radar plot as figures of merit
3.3.2 Simple figure of merit
3.3.3 Figure of merit based on Analytic Hierarchy Process
3.4 Case study: CPU-intensive application
3.5 Summary
4 An exact approach for optimizing placement 
4.1 Introduction
4.2 Goal programming
4.3 An exact approach for the Placement problem
4.3.1 Parameters
4.3.2 Variables
4.3.3 Goal
4.3.4 Constraints
4.4 Case study: Online trading platform
II A new pricing model in Cloud brokering 
5 The Pay-as-you-book pricing model 
5.1 Introduction
5.2 Pricing models in Cloud computing
5.3 Advance Reservations
5.3.1 Advance Reservation specified by Cloud providers
5.3.2 Advance Reservation specified by end-users
5.4 Pay-as-you-book
5.4.1 Initial scheduling of Advance Reservations
5.4.2 Pricing and rewarding end-users
5.4.3 Resource allocation policies
5.5 Case Study: A Virtual Cloud Provider maximizing revenues through the Pay-as-you-book pricing model
5.5.1 Experimental setup
5.5.2 Results and analysis
5.6 Summary
6 Conclusion and future works 
A Cloud performance evaluation 
A.1 Related issues to the performance evaluation
A.2 VM configurations
A.3 Benchmark duration
A.4 Performance-price correlation with a simple figure of merit of Cloud performance.
A.4.1 Correlation among VM sizes from different Cloud providers
A.4.2 Correlation among different VM sizes from a single Cloud provider
B Résumé en français 
B.1 Introduction
B.2 Mesures de performances des fournisseurs de Cloud
B.2.1 Enjeux
B.2.2 Études relatives à l’évaluation de la performance des services de Cloud
B.2.3 Caractérisation des machines virtuelles
B.2.4 Mesure de performance Cloud
B.3 Le placement dans les Clouds brokés
B.3.1 Placement basé sur des exigences non-fonctionnelles
B.3.2 Placement basé sur des exigences de l’application
B.3.3 Approche exacte au problème de placement en Cloud brokering
B.4 Les politiques de prix et les réservations faites à l’avance
B.4.1 Les politiques de prix en Cloud computing
B.4.2 Les réservations faites à l’avance
B.4.3 La politique de prix pay-as-you-book
B.5 Conclusions et travaux futurs
Bibliography

GET THE COMPLETE PROJECT

Related Posts