Novel private cloud architecture: A three tier approach to deploy private cloud using virtual machine manager

Abstract

In Present days, one of the most appealing technologies is cloud computing in the current time. Now a day, different organizations are offering various cloud services concerning their platforms. These organizations are profitable in cost-efficiency to the end-users in various factors that include process, storage, and applications instance. Most of the end-users choose the services from the cloud service provider, whereas some medium-scale organizations are also depending on the service providers even though they have enough resources. This causes the problem due to a lack of knowledge on design and deployment issues. In this paper, we proposed a private cloud that offers cloud services with limited resources so that small and medium scale organizations can be benefited. In this work, we proposed a novel three tier architecture that contains Domain Controller (DC), Virtual Machine Manager (VMM), and Member Servers (MSs) at a different level of functioning and also have their responsibilities. In this Novel three tier architecture was designed in a hierarchal fashion in which the upper layer accesses the lower layer users. The top layer, which deals with the domain controller whose responsibility, is to control the domain-specific issues like adding the member in the network, granting permissions, creating network groups, etc. The middle layer has various administrative users like Virtual Machine Manager (VMM), Storage Area Network (SAN), Disaster Recovery (D/R), and Operational Manager (SCOM), which deals with their responsibilities in cloud services. In the bottom layer, the physical servers exist and are configured by DC and get access by the middle layer users. This architecture was tested with various applications with different instances concerning storage and process utilities.

Keywords

Domain controller virtual machine manager storage area network disaster recovery operational manager member servers

1. Introduction

The market first started talking about cloud computing, the public cloud received the most attention. The release of the concept by significant providers such as Google and Amazon sparked a crisis. IT departments studying cloud architecture adoption nowadays are emphasizing private cloud-based solutions in scaling the organization. Most of the industries have resources but in the deployment of private cloud network administrators don’t know. Moreover, organizations will scale if they have proper guidelines in hosting the private cloud.

In Cloud computing, there are different types of clouds namely public, private and community, and hybrid clouds. The public cloud allows for easy access to systems and services by the entire public, although it is less secure and more robust. Software as a Service (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (IaaS) is among the cloud services offered by the public cloud providers [1]. Moreover, there are third-party tools, which provide cloud services in profitable cost earnings. In a private cloud, it is accessible within an organization moreover this increases efficient resource utilization and security. In Community cloud allows systems and services to be accessible by a group of organizations but compare to a private cloud it is less secure. However, Hybrid cloud, on the other hand, is a combination of private and public cloud, with key operations carried out in the private cloud and non-critical activities carried out in the public cloud. In Cloud technology the major aspect is virtualization; through this technology cloud platforms are possible.

Virtualization technology, that operates on different server operating systems which are completely isolated from each other. In single physical resource (such as a server, an operating system, an application, a network, or a storage device) appear as multiple logical resources. In the real world, Virtualization exists on various utilities such as Server Virtualization, Desktop Virtualization, Application Virtualization, Memory Virtualization, Storage Virtualization, Data Virtualization, Network Virtualization, etc. The major benefits are consolidation, redundancy, legacy hardware, migration, and centralized management [2].

In most of the server local partition possible through virtualization, this technology industry implies with server operating systems. Each server operating system has its hypervisor to offer virtualizations technology. There are different types of hypervisors like Type_1 and Type_2. Type-1 hypervisors directly run on hardware whereas Type-2 hypervisors run on a host operating system. In Cloud deployment servers are configured by server operating systems, each can have its hypervisors to provide virtualization. For example, Microsoft, Citrix, VMware hypervisors are hyper-V, Xenserver, and vSphere respectively. The following table [1] shows in detail vendors, hypervisor, host, and VM configuration.

Table 1
Hypervisor with server operating systems

Vendor		Microsoft	Citrix	VMware
Version		Hyper-V	XenServer	vSphere
Edition		DataCenter	Commercial Edition	Enterprise Plus
Host config	Max Consolidation ratio	1024 VMs/host, 2048 vCPUs/host	500 Windows VMs/host	512 VM/2048 vCPU
	Max CPU Host	320 logical CPUs	160 logical CPUs	160 logical CPUs
	Max Memory Host	4TB	1TB	2TB
VM config	Max vCPU per VM	Up to 64 vCPU	Up to 16vCPU	Up to 64 vCPU
	Max RAM per VM	1TB	128GB	1TB

The Contribution of the work, we are providing a solution to deploy a private cloud by making use of their limited resources. Moreover, the architecture is broken down into three layers each layer has its functionalities, roles, and responsibilities which detailed explanation given in the proposed methodology and the architecture shown in figure [1]. This work will helps in the IT industry in scaling of the industry. The key factor, large scale industries have their cloud infrastructure to offer tools and services but in medium scale and small scale industries even though having their limited resources, they are not building their Cloud infrastructure. In this proposed model, we deployed the private cloud and tested few application. The application performance has been tested with different user’s instances, which was discussed in performance and evolution section.

2. Related work

The traditional architectures have risks and issues associated with accessing policies of the system behaviors. In the latest technology growth of the software, inventors cloud has given characteristics on demand self-service, broad network access, resource pooling, and rapid elastic. There are many challenges namely data integrity, confidentiality, privacy, security issues, unauthorized threats, etc. Puri et al. [1]. In Cloud virtual environment aims maximum resource utilization and also provides network virtualization system for security turned more efficient. Many cloud vendors offer network virtualization which includes hypervisor technology Ashalatha et al. [2]. The framework of performance evaluation by the software-defined network(SDN) slice of physical resources, which provides a centralized architecture and distributed virtualization Blank et al. [3].

Numerous Approaches are proposed for different applications in cloud architectures. These architectures addressed different performance levels for different domains of applications. An efficient secured framework for mobiles uses cloud architectures that, improve various parameters like CPU utilization, power consumptions, execution time, and memory usage. The framework provides offload decisions in mobility devices Elgendy et al. [4]. Some approaches migrate only a process from the mobile device to the cloned virtual machine (VM) on the cloud. The assistive technology (AT) gives a human interaction to the lively activation of cloud interaction been rapidly increased by application users Mulfari et al. [5].

In marketplaces, Cloud service brokerage architectures extend the services layers between consumers and providers in future cloud technology and research. This framework identifies different challenges and technical requirements for service brokerage Fowley et al. [6]. The vulnerability of Cloud Computing Systems (CCSs) is significant to government and industry. Infrastructure as a service (IaaS) security evaluation approach that enhances cloud trust. Cloud-Trust – a measure that estimates high-level security metrics to evaluate a CCS or cloud service provider’s level of secrecy and integrity (CSP) Gonzales et al. [7].

A trusted computer system architecture provides the user with a friendly interaction to access the machines in such a way that integration between trusted computing vs Cloud computing Ibrahim and Hemayed [8]. The emerging CC generates a huge amount of data which helps for new advancement at deep learning algorithms. The adaption of new approaches gives future scope to artificial intelligence. Nevertheless, These approaches create an opportunity for problem-solving in a new dimension Jauro et al. [9]. There exist some application related cloud architecture for music composition’ Alvaro and Barros et al. [10] deals with Diatonic composer and autonomous vehicles Mokhtarian et al. [11], which interconnects the vehicles.

In this related survey, small, medium, and micro-sized Enterprises SMMEs Modisane and Jokonya[12] there is novel existence of layer cloud architecture to deploying a private cloud. Moreover, the proposed architecture tested with various applications interns their performance issues in various aspects like instance sizes, memory, process utilization.

3. Proposes work

3.1 Private cloud architectural design

In this work, a local private cloud was deployed, which uses a three-tier infrastructure as well as provides application service to the end-users or customers. The infrastructure consists of three-layer namely as a domain controller (DM), Virtual Machine Manager (VMM), Storage area Network (SAN), Disaster Recovery (D/R), and Member Servers (MSs). The role of domain control is to provide active directory domain service (ADDS) to the member servers. A member server role is storing the Virtual Machines (VMs). VMM’s role is to manage the resources as well as provide the services based on the requirements of the customer. Moreover, SAN is responsible for a high-speed network that interconnects the devices to store backup data on multiple servers. Each server can access shared storage as if it were a drive directly attached to the server. In D/R, It is a combination of approaches and services that are planned to back up mechanisms to the applications. In the following section, detailed functional behaviors of layered private cloud architecture are discussed as shown Figure [1]. The layered architecture facilitates the functional behavior of different administrative loads handling in the cloud.

Figure 1.

Three tier architecture.

3.1.1 Domain Controller

A Domain Controller is a cloud administrator manager that provides functions feature to authenticate user characteristics and authorize IT resources to be accessed. These resources generally include systems, applications, files storage systems, networking servers, networking facilities, storage services, application accessing, locally accessible components that are administrated within the organization. This collection of resources is referred to as a domain, and a domain controller ensures that only authorized users have access to domain-bound resources. DCs are used in conjunction with a directory services database, also known as an identity provider (IdP), which stores user information such as username, and passwords. This enables the DC to leverage the identity provider as the source of truth for verifying user identities [5].

Figure 2.

Domain controller.

In this session, we explain the procedure to deploy the private cloud in terms of a layered approach. The private Cloud architecture majorly contains three layers namely domain controller, VM manager, and member server. Moreover, each layer has its roles and responsibility. In the following section, we will explain the layer modules in cloud application deployment [6].

Domain Controller Configuration

Step 1:

Administrator checking details

Adding Roles to Sever

Choosing Domain Controller

Report DNS Issue to the administrator

Install and Configure DNS

Step 3:

Adding Active Directory Domain Service role (Based on the Server Operating System)

Step 4:

Create a new user account

Step 5:

To map to the Kerberos service, create a user account

3.1.2 Virtual machine manager

Today IT operations are being expanded resources in managing and supervisory controls in various challenging issues. Those are provided in a degree of difficulties and proliferation of tools, which controls enterprise-class solutions. Microsoft providing provisioning infrastructure namely SCVMM (Microsoft System Center Virtual Machine Manager) is a virtual machine manager that provides an enterprise-class solution in various server machines under the supervision of manages, virtualizing, configuring, and transforming in datacenters [5]. Moreover, these supervisors are different based on vendors that are shown in table [2]. In the main objective of VMMs are to provide a host machine in VMs and each VM can be accessed by various group users. The responsibility of VMMs is to provide access to user groups so that users can access VMs. Each VM may contain applications and services deployed by the user management.

Figure 3.

Virtual machine manager.

VMMs help in unified management, which provides experience to servers, storage, and security in software-based datacenters. The VMM Software tools are optimized locally to partition the host machine and management solution to business vendors. Moreover, the management solution to business vendors manages the full life cycle of VMs at a scale up to 1000 hosts to 25000VMs [6]. Fabric can quickly deploy and service multi-tier applications and manage both on-premises and in the cloud with a single tool. With advanced management for Windows servers, users can deploy hyper-V hosts and scale-out file servers on bare metal devices. It makes provisioning and migration of virtual machines, SAN devices, and network devices in data centers easier.

3.1.3 VM creation

There are different ways to create a VM in the cloud environment. In this section, we are explaining in detail methods to deploy VMs. The first method uses the VM template here template means already existing stored procedure that is been used. The second method is installation media which uses a server pool based on virtualization or paravirtualization (provides a path for installing files or applications). The third method is bootable VMs which are used based on the requirements of users [13].

Method 1: (VM Template)

Step 1:
Create and start an instance
a.
Set up a region zone
b.
Setup permissions
c.
Allows network traffic

Step 2:
Select Server Pool
Step 3:
Select VM template
Step 4:
Enter the Configure information for VM
Step 5:
Confirm

Method 2: (VM Installation Media)

Step 1:

Create and start an instance

Set up a region zone

Setup permissions

Allows network traffic

Step 2:

Select Server Pool

Step 3:

Select VM Method

Fully Virtualized

Select the required ISO File

ii)

Enter the Configure information for VM

Para virtualized

i).

Provide a path for installation files

ii).

Enter the Configure information for VM

Step 4:

Enter the Configure information for VM

Step 5:

Confirm

Method 3: (VM Template)
Step 1: Create and start an instance a. Set up a region zone b. Setup permissions c. Allows network traffic Step 2: Select Server Pool Step 3: Create a network bootable Virtual Machine Step 4: Enter the Configure information for VM Step 5: Confirm

3.1.4 Storage area network (SAN)

A storage area network (SAN) is easy, high-pace information transfer network that provides access to the very last storage consolidation, information storage devices. Storage area Networks are generally made from hubs, switches, garage components, and storage gadgets that are interconnected with an expansion of technologies, protocols, and topologies.

There are different units of SAN i.e Server Zone, Connectivity Zone, and Cables. To connect Hubs, switches mostly fiber or copper cables have been used. Nevertheless, to transfer data in fast transmission high-speed fiber switches are used. The advantage of this SAN provides Storage virtualization, high-speed disk technology, Centralized Backup, Dynamic failover protection, and Data Security [5, 8].

3.1.5 Disaster recovery (D/R)

Cloud disaster recovery (cloud DR) is a set of techniques and services for backing up data, apps, and other resources to a public cloud or a dedicated service provider. When a disaster strikes, the enterprise’s data, apps, and other resources may be restored to a local data center or a cloud provider, allowing it to continue normal operations. The advantages of D/R are as follows:

Pay As you Go: This is based on organization demand. The major advantage of cloud service is paying as you, which enables the organization to pay monthly in recurrent pay. This Cloud Service provider charges based on the resources that are been added/ removed, the payment will change.

Flexibility and Scalability: Traditional D/R methods, whether done locally or remotely, frequently improved flexibility and scalability limits. The testing, design, and maintenance of the infrastructure are the problems in disaster recovery.

Easy testing and Fast Recovery: VM image files make it simple to evaluate cloud workloads. To maximize data transfer, it has high bandwidth and quick disc i/o choices. It has restricted connection, compliance constraints, and optimal recovery in traditional D/R.

3.1.6 Member server

A member server is a backbone server that provides services and applications in a domain. belongs to a domain but is not the domain controller. It can function as a file server, database server, application server, firewall, remote access server, and certificate server. The domain controller is responsible for authenticating security requests such as logins and permission checking.

Table 2
Member Server roles and responsibility

Member server	Role	Responsibility
File Server	Provides file systems to the connected clients	Central Storage and Management of data files So that other member servers can also access the files
Database Server	Provides permissions/privileges that can be assigned to one or more users	Database backup/recovery, security authentication, and performance monitoring
Application Server	Provides both tools for developing web applications and a server environment in which to run them.	Application services to clients/ end users
Firewall Server	To control traffic between internal and external servers	Application owner, System administrator, information security, and account manager
Remote access Server	Provides a dashboard for administrators to manage, configure, and monitor network access.	Network accessing, backup maintenance
Certificate Server	Providing access tokens to the users	Authorized token generation for the users

Figure 4.

Storage network architecture.

4. Performance metrics and evaluation

In this section, we explain architecture performance to the process, memory, and application instance. A detailed determination of the application instance to VMs in dynamic runtime has also been considered based on the number of user accessing the application VMs [10].

As Figure [1] is shown, the architecture consists of three layers and the performance matrices of the private cloud that depend on the matrices namely memory utilization, process utilization, power consumption, and application size [4].

Memory utilization: It is the average utilization derived from the percentage of available memory in use at a given moment.

$\displaystyle\textit{Cmemory}=\sum\limits_{i=1}^{n}M_{i}\ast\left({1-x_{i}}% \right)+\sum_{i=1}^{n}\textit{M\_Sec}_{i}\ast x_{i}$

Table 3
Applications used in experimental

Application	Description
Java Application (JDK)	Using JDK implemented sample java programs
Oracle Tool	Using Oracle Tool implemented Database applications
Rational Rose	Using Rational Rose software implemented Object-oriented design models
Turbo C	Using turbo C implemented Sample C Programs

Figure 5.

VMs performance.

Process utilization: The amount of time to be taken by executing a process. In the execution of any process, RAM is used to save data and the CPU will process the data.

$\displaystyle\textit{Ccpu}=\sum_{i=1}^{n}\textit{CPU}_{i}\ast\left({1-x_{i}}% \right)+\sum_{i=1}^{n}\textit{CPU\_Sec}_{i}\ast x_{i}$

Power Consumption: It refers to the electrical energy supplied to operate some operation per unit time.

$\displaystyle\textit{Cpower}=\sum_{i=1}^{n}P_{i}\ast\left({1-x_{i}}\right)+% \sum_{i=1}^{n}P_{i}\ast x_{i}$

Application Performance: It refers to the amount of memory, process utilization, power consumption, and number of users accessing the particular application at any moment in time

$\displaystyle\textit{CApp}=\sum_{i=1}^{n}\textit{User}_{i}\ast(\textit{Cmemory% }\left(i\right)\ast\textit{Ccpu}\left(i\right)\ast\textit{Cpower}\left(i\right))$

4.1 Evaluation and analysis

In Evaluation, It was tested by four different types of applications, as shown in Table [4]. The proposed. After deploying the cloud initially four VMs are created and hosted each listed application respectively in VM1 to VM4. This architecture tested with various parameters includes CPU utilization, memory usage, and the number of users accessing the host machines. This evaluation improves the application efficiency with VMs [14].

The experimental results measure four parameters for running the application methods locally on a deployed architecture and when offloading to the cloud by using the framework. These parameters include processing Time, CPU utilization, power consumption, memory usage, and the number of users. The evaluation demonstrates how the proposed design can improve the performance of certain applications [4].

4.2 Experimental setup

The private cloud framework is composed of various VMs, which are tested by the experimental setup. Each VMs consists of a single application with different instances of memory sizes shown in Table 3.

Table 4
Virtual Machine with different instances

Application	Virtual machine	Application instance	Memory size	Performance
Oracle tool	Oracle-VM1	80 MB	2 GB	98
			4 GB	65
			6 GB	40
			8 GB	20
Java Application (JDK)	JDK-VM2	54 MB	2 GB	88
			4 GB	60
			6 GB	30
			8 GB	15
Rational rose	RR-VM3	48 MB	2 GB	80
			4 GB	50
			6 GB	20
			8 GB	15
Turbo C	TC-VM4	30 MB	2 GB	60
			4 GB	35
			6 GB	20
			8 GB	10

Figure 6.

Process Utilization with different memories.

Figure 7.

VMs performance bases on users.

4.3 Experimental and evaluation results

“In the proposed framework, the linear model is used to determine the offloading decision and selects a correct decision for the four applications when solving for the parameters of the methods. However, the framework is modified to allow applications in a different virtual machine with respect to different memory instances” [4].

The applications are deployed in different Virtual machines namely JDK-VM1, Oracle-VM2, RR-VM3, TC-VM4 respectively. It contains a maximum memory size of 8 GB and storage of 128 GB size. The VMs loaded with different instances of users with 2 GB, 4 GB, 6 GB and 8 GB respectively.

In figure [5] shows that the number of users increasing than CPU utilization also rapidly grown. The performance is varied to application instance sizes oracle software had taken 80 MB of memory when it is deployed in the cloud whereas JDK, RR, and TC deployment instances 54 MB, 48 MB, and 30 MB respectively. The following figure has shown that VM performance for different memory sizes.

In this private cloud deployment, the VMs are tested to memory, process, and several users. In this work, we tested the VMs with a maximum of 100 users. Moreover, In deployment, we had created 100 user accounts under different groups and each group consists of a maximum of 20 users. These groups are loaded with different instances of times. Figures 6 and 7 are based on the VM’s performance to users as well as process utilization with different memory sizes.

5. Conclusion

In this paper, novel private cloud architecture is proposed to improve the efficiency of VMs with various applications. This framework can be offered application methods that are tested with different users, instances sizes, and memory sizes. This study showed how the private cloud is working, and in each layer who are responsible agents that act their responsibilities. Finally, the offline execution of demanding techniques of applications comes to a close. Furthermore, we must allow parallelization for the method’s execution in the cloud to potentially reduce execution time.

References

Puri

Tiwary

Shukla

. A review on cloud computing. In: 2019 9th International Conference on Cloud Computing, Data Science & Engineering (Confluence), IEEE, 2019 Jan 10, pp. 63-68

Ashalatha

Agarkhed

Patil

. Network virtualization system for security in cloud computing. In: 2017 11th International Conference on Intelligent Systems and Control (ISCO), IEEE, 2017 Jan 5, pp. 346-350.

Blenk

Basta

Reisslein

Kellerer

. Survey on network virtualization hypervisors for software defined networking. IEEE Communications Surveys & Tutorials. 2015 Oct 9; 18(1): 655-85.

Elgendy

Zhang

Liu

Hsu

. An efficient and secured framework for mobile cloud computing. IEEE Transactions on Cloud Computing. 2018 Jun 18; 9(1): 79-87.

Mulfari

Celesti

Villari

. A computer system architecture providing a user-friendly man machine interface for accessing assistive technology in cloud computing. Journal of Systems and Software. 2015 Feb 1; 100: 129-38.

Fowley

Pahl

Jamshidi

Fang

Liu

. A classification and comparison framework for cloud service brokerage architectures. IEEE Transactions on Cloud Computing. 2016 Mar 2; 6(2): 358-71.

Gonzales

Kaplan

Saltzman

Winkelman

Woods

. Cloud-trust – A security assessment model for infrastructure as a service (IaaS) clouds. IEEE Transactions on Cloud Computing. 2015 Mar 30; 5(3): 523-36.

Ibrahim

Hemayed

. Trusted cloud computing architectures for infrastructure as a service: Survey and systematic literature review. Computers & Security. 2019 May 1; 82: 196-226.

Jauro

Chiroma

Gital

Almutairi

Shafi’i

Abawajy

. Deep learning architectures in emerging cloud computing architectures: Recent development, challenges and next research trend. Applied Soft Computing. 2020 Nov 1; 96: 106582.

10.

Alvaro

Barros

. A new cloud computing architecture for music composition. Journal of Network and Computer Applications. 2013 Jan 1; 36(1): 429-43.

11.

Mokhtarian

Kampmann

Lueer

Kowalewski

Alrifaee

. A cloud architecture for networked and autonomous vehicles. IFAC-PapersOnLine. 2021 Jan 1; 54(2): 233-9.

12.

Modisane

Jokonya

. Evaluating the benefits of cloud computing in small, medium and micro-sized enterprises (SMMEs). Procedia Computer Science. 2021 Jan 1; 181: 784-92.

13.

Yang

Wei

Zhu

Tan

. Virtual private cloud based power-dispatching automation system – Architecture and application. IEEE Transactions on Industrial Informatics. 2018 Jun 29; 15(3): 1756-66.

14.

Chao

Qing

Kui

Yuwen

Hailong

. Industrial Design and Development Software System Architecture Based on Model-Based Systems Engineering and Cloud Computing. Annual Reviews in Control. 2021 Jan 1; 51: 401-23.

Novel private cloud architecture: A three tier approach to deploy private cloud using virtual machine manager

Abstract

Keywords

1. Introduction

Table 1 Hypervisor with server operating systems

3. Proposes work

3.1 Private cloud architectural design

Step 1: Create and start an instance a. Set up a region zone b. Setup permissions c. Allows network traffic Step 2: Select Server Pool Step 3: Select VM template Step 4: Enter the Configure information for VM Step 5: Confirm

3.1.5 Disaster recovery (D/R)

3.1.6 Member server

Table 2 Member Server roles and responsibility

Table 3 Applications used in experimental

4.2 Experimental setup

Table 4 Virtual Machine with different instances

5. Conclusion

References

Table 1
Hypervisor with server operating systems

Step 1:
Create and start an instance
a.
Set up a region zone
b.
Setup permissions
c.
Allows network traffic

Step 2:
Select Server Pool
Step 3:
Select VM template
Step 4:
Enter the Configure information for VM
Step 5:
Confirm

Table 2
Member Server roles and responsibility

Table 3
Applications used in experimental

Table 4
Virtual Machine with different instances