Cloud Computing with e-Science Applications

Olivier Terzo, Lorenzo Mossucca

• Publisher: CRC Press
• Year: 2017
• Print ISBN: 9780367828233
• eBook ISBN: 9781351831543
• Edition: 1
• Topic: Computer Science
• Subtopic: Cloud Computing
• Index: Computer Science

The amount of data in everyday life has been exploding. This data increase has been especially significant in scientific fields, where substantial amounts of data must be captured, communicated, aggregated, stored, and analyzed. Cloud Computing with e-Science Applications explains how cloud computing can improve data management in data-heavy fields such as bioinformatics, earth science, and computer science.

The book begins with an overview of cloud models supplied by the National Institute of Standards and Technology (NIST), and then:

• Discusses the challenges imposed by big data on scientific data infrastructures, including security and trust issues
• Covers vulnerabilities such as data theft or loss, privacy concerns, infected applications, threats in virtualization, and cross-virtual machine attack
• Describes the implementation of workflows in clouds, proposing an architecture composed of two layers—platform and application
• Details infrastructure-as-a-service (IaaS), platform-as-a-service (PaaS), and software-as-a-service (SaaS) solutions based on public, private, and hybrid cloud computing models
• Demonstrates how cloud computing aids in resource control, vertical and horizontal scalability, interoperability, and adaptive scheduling

Featuring significant contributions from research centers, universities, and industries worldwide, Cloud Computing with e-Science Applications presents innovative cloud migration methodologies applicable to a variety of fields where large data sets are produced. The book provides the scientific community with an essential reference for moving applications to the cloud.

1
Evaluation Criteria to Run Scientific
Applications in the Cloud

Eduardo Roloff, Alexandre da, Silva Carissimi,
Philippe Olivier and Alexandre Navaux

Contents

1. Summary
2. 1.1 Introduction
3. 1.2 Cloud Service Models
   1. 1.2.1 Software as a Service
   2. 1.2.2 Platform as a Service
   3. 1.2.3 Infrastructure as a Service
4. 1.3 Cloud Implementation Models
   1. 1.3.1 Private Cloud
   2. 1.3.2 Community Cloud
   3. 1.3.3 Public Cloud
   4. 1.3.4 Hybrid Cloud
   5. 1.3.5 Summary of the Implementation Models
5. 1.4 Considerations about Public Providers
   1. 1.4.1 Data Confidentiality
   2. 1.4.2 Administrative Concerns
   3. 1.4.3 Performance
6. 1.5 Evaluation Criteria
7. 1.6 Analysis of Cloud Providers
   1. 1.6.1 Amazon Web Services
   2. 1.6.2 Rackspace
   3. 1.6.3 Microsoft Windows Azure
   4. 1.6.4 Google App Engine
8. 1.7 Cost Efficiency Evaluation
   1. 1.7.1 Cost Efficiency Factor
   2. 1.7.2 Break-Even Point
9. 1.8 Evaluation of Providers: A Practical Example
10. 1.9 Conclusions
11. References

Summary

In this chapter, we will present a brief explanation of the services and implementation of models of cloud computing in order to promote a discussion of the strong and weak points of each. Our aim is to select the best combination of the models as a platform for executing e-science applications.

Additionally, the evaluation criteria will be introduced so as to guide the user in making the correct choice from the available options. After that, the main public cloud providers, and their chief characteristics, are discussed.

One of the most important aspects of choosing a public cloud provider is the cost of its services, but its performance also needs to be taken into account. For this reason, we have introduced the cost efficiency evaluation to support the user in assessing both price and performance when choosing a provider. Finally, we provide a concrete example of applying the cost efficiency evaluation using a real-life situation and including our conclusions.

1.1 Introduction

To create a service to execute scientific applications in the cloud, the user needs to choose an adequate cloud environment [1, 2]. The cloud computing model has several possible combinations between the service and implementation models, and these combinations need to be analyzed. The public cloud providers offer an alternative to avoid the up-front costs of buying machines, but it is necessary to evaluate them using certain criteria to verify if they meet the needs of the users. This chapter provides a discussion about these aspects to help the user in the process of building an e-Science service in the cloud.

1.2 Cloud Service Models

According to the National Institute of Standards and Technology (NIST) definition [3], there are three cloud service models, represented in Figure 1.1. They present several characteristics that need to be known by the user. All three models have strong and weak points that influence the adequacy for use to create an e-Science service.

The characteristics of the service models are presented and discussed in this section.

Figure 1.1 Service models.

1.2.1 Software as a Service

The software-as-a-service (SaaS) model is commonly used to deliver e-science services to users. This kind of portal is used to run standard scientific applications, and no customization is allowed. Normally, a provider ports an application to its cloud environment and then provides access for the users to use the applications on a regular pay-per-use model. The user of this model is the end user, such as a biologist, and there is usually no need to modify the application.

One example of a provider porting a scientific application and then providing the service to the community is the Azure BLAST [2] project. In this project, Microsoft ports the Basic Local Alignment Search Tool (BLAST) of the National Center for Biotechnology Information (NCBI) to Windows Azure. BLAST is a suite of programs used by bioinformatics laboratories to analyze genomics data. Another case of this use are the Cyclone Applications, which consist of twenty applications offered as a service by Silicon Graphics Incorporated (SGI). SGI provides a broad range of applications that cover several research topics, but there is no possibility to customize and adapt them.

The big problem with SaaS as the environment to build e-science services is the absence of the ability for customization. Research groups are constantly improving their applications, adding new features, or improving their performance, and they need an environment to deliver the modifications. In addition, there are several applications that are used for only a few research groups, and this kind of application does not attract the interest of the cloud providers to port them. In this case, this model can be used to deliver an e-science service but not as an environment to build it.

1.2.2 Platform as a Service

The platform-as-a-service (PaaS) model presents more flexibility than the SaaS model. Using this model, it is possible to develop a new, fully customized application and then execute it in the provider’s cloud environment. It is also possible to modify an existing application to be compatible with the provider’s model of execution; in the majority of cases, this is a realistic scenario for scientific applications [4]. The majority of the services provided in this model consist of an environment to execute web-based applications. This kind of application processes a large number of simultaneous requests from different users. The regular architecture of these applications is composed of a web page, which interacts with the user; a processing layer, which implements the business model; and a database, used for data persistence. Each user request is treated uniquely in the system and has no relationship with other requests. Due to this, it is impossible to create a system to perform distributed computing. However, the processing layer of this model can be used if the service does not have a huge demand for processing power.

In the PaaS model, the provider defines the programming languages and the operating system that can be used; this is a limitation for general-purpose scientific application development.

1.2.3 Infrastructure as a Service

The infrastructure-as-a-service (IaaS) model is the most flexible service model of cloud computing. The model delivers raw computational resources to the user, normally in the form of virtual machines (VMs). It is possible to choose the size of the VM, defining the number of cores and the amount of memory. The user can even choose the operating system and install any desired software in the VM. The user can allocate any desired quantity of VMs and build a complete parallel system. With this flexibility, it is possible to use IaaS for applications that need a large amount of resources by the configuration of a cluster in the cloud.

1.3 Cloud Implementation Models

The service models, presented in the previous section, can be delivered using four different implementation models: private cloud, community cloud, public cloud, and hybrid cloud. Each one has strong and weak points. The four models can be used to build an e-science service, and they are analyzed to present their main characteristics to help the user decide which one to choose.

1.3.1 Private Cloud

A private cloud is basically the same as owning and maintaining a traditional cluster, where the user has total control over the infrastructure and can configure the machines according to need. One big issue in a private scenario is the absence of instant scalability, as the capacity of execution is limited to the physical hardware available. Moreover, the user needs to have access to facilities to maintain the machines and is responsible for the energy consumption of the system. Another disadvantage is the hardware maintenance; for example, if a machine has physical problems, the user is responsible for fixing or replacing it. A case for which the private cloud is recommended is if the application uses confidential or restricted data; in this scenario, the access control to the data is guaranteed by the user’s policies. The weakness of this model is the absence of elasticity and the need for up-front costs. Building a private cloud for scientific applications can be considered the same as buying a cluster system.

1.3.2 Community Cloud

In a community cloud, the users are members of one organization, and this organization has a set of resources that are connected to resources in other organizations. A user from one of the organizations can use the resources of all other organizations. The advantage of this model is the provision of access to a large set of resources without charging because the remote resources belong to other or...

Table of contents

• Cover Page
• Half title
• Title Page
• Contents
• Preface
• Acknowledgments
• About the Editors
• List of Contributors
• 1 Evaluation Criteria to Run Scientific Applications in the Cloud
• 2 Cloud-Based Infrastructure for Data-Intensive e-Science Applications: Requirements and Architecture
• 3 Securing Cloud Data
• 4 Adaptive Execution of Scientific Workflow Applications on Clouds
• 5 Migrating e-Science Applications to the Cloud: Methodology and Evaluation
• 6 Closing the Gap between Cloud Providers and Scientific Users
• 7 Assembling Cloud-Based Geographic Information Systems
• 8 HCloud, a Healthcare-Oriented Cloud System with Improved Efficiency in Biomedical Data Processing
• 9 RPig: Concise Programming Framework by Integrating R with Pig for Big Data Analytics
• 10 AutoDock Gateway for Molecular Docking Simulations in Cloud Systems
• 11 SaaS Clouds Supporting Biology and Medicine
• 12 Energy-Aware Policies in Ubiquitous Computing Facilities
• Index