Heterogeneity, High Performance Computing, Self-Organization and the Cloud

ISBN-13: 9783030093877 / Angielski / Miękka / 2018 / 165 str.

Theo Lynn; John P. Morrison; David Kenny

Spis treści

1        Chapter 1: Addressing the Complexity of HPC in the Cloud: Emergence, Self-Organisation, Self-Management and the Separation of Concerns

1.1         Introduction. 14

1.2         Cloud Computing. 14

1.3         High Performance Computing. 16

1.4         HPC and the Cloud. 18

1.5         Heterogeneous Computing. 19

1.6         Addressing Complexity in the Cloud through Self-* Design Principles. 21

1.7         Application Scenarios. 25

1.7.1     Oil and Gas Exploration. 25

1.7.2     Ray Tracing. 26

1.7.3     Genomics. 26

1.8         Conclusion. 28

1.9         Chapter 1 Related CloudLightning Readings. 29

2        Chapter 2: Cloud Architectures and Management Approaches. 35

2.1         Introduction. 36

2.2         Cloud Architecture. 36

2.2.1     Infrastructure Organisation. 37

2.2.2     The Cloud Management Layer. 39

2.2.3     The Service Delivery Layer. 40

2.3         Transitioning to Heterogeneous Clouds. 41

2.3.1     Resource Management. 42

2.3.2     Resource Abstraction. 43

2.4         The CloudLightning Approach. 43

2.4.1     Infrastructure Organisation. 43

2.4.2     Hardware Organisation. 44

2.4.3     The Cloud Management Layer. 44

2.4.4     Service Delivery Model. 48

2.4.5     Advanced Architecture Support. 49

2.5         Conclusion. 51

2.6         Chapter 2 Related CloudLightning Readings. 52

3        Chapter 3: Self-organising, Self-managing Frameworks and Strategies  55

3.1         Introduction. 56

3.2         Key Concepts. 56

3.3         Augmenting the CloudLightning Architecture. 57

3.4         Self-organisation and Self-management in CloudLightning Architecture. 59

3.4.1     Directed Evolution. 59

3.4.2     Self-management mechanisms. 61

3.4.3     Self-organisation mechanisms. 64

3.5         CloudLightning SOSM strategies. 65

3.5.1     Self-management strategies. 65

3.5.2     Self-organisation strategies. 67

3.6         Conclusion. 70

3.7         Chapter 3 Related CloudLightning Readings. 71

4        Chapter 4: Application Blueprints and Service Description.. 72

4.1         Introduction. 73

4.2         Representative Application Lifecycle and Resource Management Frameworks. 73

4.3         Cloud Lightning Stakeholders and Associated Concerns. 74

4.4         The CloudLightning approach based on separation of concerns. 74

4.4.1     CloudLightning Requirements. 74

4.4.2     Separation of Concerns. 76

4.5         The CloudLightning Gateway Architecture. 77

4.5.1     Gateway Service Architecture. 78

4.5.2     Service Decomposition. 79

4.5.3     Interaction with the SOSM System.. 79

4.6         The CloudLightning Blueprint Extensions. 81

4.6.1     CloudLightning Brooklyn Extensions. 81

4.6.2     CloudLightning Abstract Blueprint. 82

4.6.3     CloudLightning Blueprint. 83

4.7         Example of Application Creation and Deployment. 83

4.8         Conclusion. 85

4.9         Chapter 4 Related CloudLightning Readings. 86

5        Chapter 5: Simulating Heterogeneous Clouds at Scale.. 88

5.1         Introduction. 89

5.2         Cloud Simulation Frameworks. 90

5.3         CloudLightning Simulator. 90

5.3.1     Architecture and basic characteristics of the parallel CloudLightning simulation framework  91

5.3.2     SOSM engine. 92

5.4         Experimental Results. 97

5.5         Conclusion. 100

5.6         Chapter 5 Related CloudLightning Readings. 101

6        Chapter 6: Concluding Remarks. 104

O autorze

Theo Lynn is Professor of Digital Business and the Associate Dean (Industry Engagement & Innovation) at DCU Business School, Ireland.

John P. Morrison is the founder and director of the Centre for Unified Computing, University College Cork, Ireland.

David Kenny is the project manager of the CloudLightning project at University College Cork, Ireland.

Blurb

This open access book addresses the most recent developments in cloud computing such as HPC in the Cloud, heterogeneous cloud, self-organising and self-management, and discusses the business implications of cloud computing adoption. Establishing the need for a new architecture for cloud computing, it discusses a novel cloud management and delivery architecture based on the principles of self-organisation and self-management. This focus shifts the deployment and optimisation effort from the consumer to the software stack running on the cloud infrastructure. It also outlines validation challenges and introduces a novel generalised extensible simulation framework to illustrate the effectiveness, performance and scalability of self-organising and self-managing delivery models on hyperscale cloud infrastructures. It concludes with a number of potential use cases for self-organising, self-managing clouds and the impact on those businesses.