Design and Simulation Study of a Scalable and Resilient Network Architecture for a Hybrid Cloud; Its Billing System
Chapter One
PURPOSE OF THIS RESEARCH WORK
The purpose of this research work is to propose an evaluation mechanism that can be used to evaluate the effect of the billing factors in a scalable and resilient hybrid cloud environment. The evaluation mechanism is used in simulations to ascertain its efficiency in evaluating the billing system, –used to bill the cloud customers. The simulations are done using the extended cloud simulator “Cloudsim [24]” on a single computer node which reduces the burden of looking for a physical data center for the simulation. The main goal of the simulations is to evaluate the effect of the billing factors and that of virtualization in a hybrid cloud.
CHAPTER TWO
LITERATURE REVIEW
INTRODUCTION
This chapter elaborates on the concepts of cloud computing that are relevant to the realization of this research work. It equally presents a review of the works of authors who worked in the related areas.
CLOUD COMPUTING CONCEPTS
VIRTUALIZATION
For cloud computing to achieve its purpose of scalability, virtualization is used as the underlying technology. Virtualization [4] is the emulation (duplication) of hardware within a software platform. This technology allows a single computer to perform the functions of many computers. Virtualization as defined above is known as full virtualization which allows the simulation of the entire machine. However, there are other different types of virtualizations which do not simulate the entire machine. The most common of which is the virtual memory which makes data scattered across RAM and Hard drive to appear as being stored contiguously and in order. Virtualization allows the simulation of hardware via software with the help of virtualization software that is discussed later in this section.
As seen in the diagrams above, virtualization helps to:
- Increase the use of hardware resources (as the work of two physical servers is done by one server with maximum utilization),
- Reduce the management and resource cost (as only one server will be bought instead oftwo),
- Improve business flexibility,
- Improve security and reduce run
The concept of virtualization came into existence as far back as 1960s [1] where IBM implemented it to help split large mainframe machines into separate “virtual machines”. This helped to maximize available mainframe computers efficiency. Virtualization technologies have proven to be of great importance of late. Some of the popular technologies include:
VMware
VMware is one of the most widely known virtualization companies. Some of its virtualization applications include:
- VMware WorkStation which allows users to create multiple x86-based virtual machines on a single physical machine. The majority of these guest operating systems such as Windows, Linux and MAC OSX can be installed onto these virtual machines. This is referred as hosted virtualization
- VMware ESX Server: This does not require a host OS to be installed as it is installed directly ontoa server‟s This is referred as bare metal virtualization solution or hypervisor.
APPLICATION VIRTUALIZATION
Application virtualization is the encapsulation of applications from the operation that are installed on. This helps to prevent application malfunction in a case where more than one application use the same drivers. Application virtualization in this case makes separate copies of shared resources and this makes applications to run in complete isolation from one another even though installed on the same operating system.
STORAGE VIRTUALIZATION
This involves the virtualization of physical storage devices. This technique allows many different users or applications to access storage, regardless of its location. The virtualized storage appears local to host although it may be distributed across many different locations. An important benefit derived from storage virtualization is the ability of the administrators to mask particular hard drives or storage volumes from particular machines and this improves security and the ability for the administrator to increase a storage volume in short time especially when the server appears to be running out of space.
VIRTUALIZATION SOFTWARE
This is a kind of software that deals with the emulation of hardware and splitting up of both physical and software-based resources. Some of the most popular used virtualization software includes:
- Microsoft Virtual PC: This is Microsoft‟s hosted virtualization which supports mainly windows operating systems and to a lesser extent some Linux-based operating
- Sun xVM Virtual Box: This is sun Microsystems‟s hosted virtualization software which runs on x86 hardware platforms. It can run multiple virtual machines on windows, Solaris, Linux and Mac
- VMware ESX Server: This is VMware‟s bare metal virtualization solution. It is the most widely used solution and can support a wide range of operating
- Virtual Box: This is a hosted virtualization application created by Innotek that can emulate standard x86-based hardware and run a number of different guest operating systems like windows, Solaris,Linux.
CHAPTER THREE
HYBRID CLOUD NETWORK ARCHITECTURAL DESIGN
INTRODUCTION
This chapter discuses the network architectural model of a hybrid cloud with much attention on the mathematical model of the scalable data center. It also discusses the types of routing protocols used at different levels of the design and finally ends with a proposed solution to the security and confidentiality aspects of data in a hybrid cloud.
HYBRID CLOUD NETWORK DESIGN
The hybrid cloud is constructed by interconnecting the components of a private cloud to those of a public cloud. The constructed architecture then provides services to the immediate users of the organization as well as the public. The components for the architecture include: Data Center Network, Data Center Interconnect Network, Cloud in a box and Network Service Nodes. For the purpose of this thesis, the network architecture of the components will be individually discussed, finally, presenting the “big picture”.
DATA CENTER NETWORK DESIGN
Data Center Network Architecture is an architecture that interconnects cloud resources like servers and storage devices in a cloud data center. The data centers in a cloud environment are used for the provisioning of scalable cloud services. This scalability is achieved with the help of the construction of Data Center Network (DCN) that may connect as many as thousands of servers together which then provide scalable services to the end users. The data center is made up of the core layer, the aggregation layer and the access layer [16]. Below is a view of a common Data Center Network Architecture which uses 10 GigaE Switches at aggregation layer and GigaE switches at access layer.
CHAPTER FOUR
SIMULATIONS
INTRODUCTION
The chapter gives the step-by-step procedures, with an evaluation of how pricing (billing) and scalability (virtualization) are achieved in the designed hybrid Cloud Computing environment. The chapter starts by presenting the factors that affect pricing followed by the effects of virtualization in the hybrid cloud environment and ends with the results obtained from the simulations.
CHAPTER FIVE:
CONCLUSION AND FUTURE WORK
CONCLUSION
The recent works of Rodrigo et al in [14] and [18] on cloudsim did not incorporate pricing policies. To fill this gab, we have extended the cloudsim tool kit by implementing a pricing mechanism that effectively manages the bills in the cloud environment. The efficiency of this new implementation has been tested in this thesis to evaluate the effect of the pricing factors in the hybrid cloud. This work will therefore serve as a tool for future research on cloud management thereby improving on the number of prospective cloud users.
FUTURE WORKS
In our design of the network architecture for a hybrid cloud, we concentrated on the network architecture of the immediate vicinity of the cloud provider, paying little attention on what happens en route when the packets are transmitted between the provider and the users. This security issue is one of the main setbacks of cloud computing. This is because most customers (companies) of cloud computing deal with sensitive and confidential data that if leaked to a third party, the company may suffer a severe loss. So as future work, we plan to work on the best approach to improve on the confidentiality and security of data when hosted in the cloud.
REFERENCES
- Borko Furht. Cloud Computing Fundamentals. Hand Book of Cloud Computing,Springer Science + Business Media, LLC 2010
- Victor Delgado. Exploring the limits of Cloud Computing. Thesis, Stockholm2010
- Hai Jin, Shadi Ibrahim, Tim Bell, Wei Gao, Dachuam Huang, Song Wu. Cloud Types and Services. Hand Book of Cloud Computing, Springer Science + Business Media, LLC2010
- The Art of Service: Cloud Computing Specialist Certification Kit; http://www.theartofservice.com/. October 2011
- Jinzy Zhu. Cloud Computing Technologies and Applications. Hand Book of Cloud Computing, Springer Science + Business Media, LLC2010
- Cisco Data Center Infrastructure 2.5 Design Guide.http://www.cisco.com/univercd/cc/td/doc/solution/dcidg21.pdf. October 2011
- Anthony M. Middleton. Data-Intensive Technologies for Cloud Computing. Hand Bookof Cloud Computing, Springer Science + Business Media, LLC 2010
- Cisco Data Center Interconnect Design and Implementation Guide. www.cisco.com/en/…/data_center_interconnect_design_guide.pdf. September 2011
- David Villegas, Ivan Rodero, Liana Fong, Norman Bobroff, Yandbin Liu, Manish P. and Sadjack S. The Role of Grid Computing Technologies in the Cloud Computing. Hand Book of Cloud Computing, Springer Science + Business Media, LLC2010
- Buyya R, Rajiv R, Rodrigo N. Calheiros. Modeling and Simulation of Scalable CloudComputing Environments and the Cloudsim Toolkit: Challenges and Opportunities, Proceedings of 7th High Performance Computing and Simulation Conference (HPCS 09), IEEE Computer Society, June 2009
