A Unified Framework of the Cloud Computing Service Model

Wen-Lung Shiau and Chao-Ming Hsiao

1. Introduction

The information technology is in a state of constant change[1]-[3]. For example, computing has evolved from distributed computing, grid computing, and utility computing to cloud computing. Thus, cloud computing is emerging as a new computational method and service[4]. A dramatic proliferation of research has recently emerged concerning cloud computing. Increasingly more aspects of work and life are conducted online and the popularity of smart phones has resulted in using computing resources anywhere, or cloud computing. Cloud computing is not a product, it is a computing service to be delivered[5]. Cloud computing provides Internet software, computation, storage resources, and data access, of which the user does not need to know any details of the computing infrastructure. Cloud computing generally involves three categories: software as a service (SaaS), platform as a service (PaaS), and infrastructure as a service (IaaS)[6], which provide various services for daily life.

According to a Gartner Executive Programs worldwide survey[7], more than 2000 chief information officers (CIOs)identified cloud computing as a top technology priority for CIOs in 2011. This finding indicates that companies must start thinking of approaches for using cloud computing to improve company effectiveness because companies will want to learn about cloud computing and determine what type of cloud computing technology they can use. Currently,many cloud service providers can provide services[8].However, a unified cloud framework/architecture does not exist; only an architecture that cloud service providers use for their own services currently exists. Most people are unable to understand the full range of cloud computing.Therefore, after investigating the core technology of cloud computing based on a thorough review of literature, this study proposes a unified cloud computing architecture and provides a comparison of Microsoft, Trend Micro, and the proposed unified cloud architecture. This research addresses the following questions: (1) What is the core knowledge of cloud computing? (2) Is a unified cloud computing service architecture applicable for business?

The remainder of this paper is organized as follows.The first section provides background information on cloud computing. The second section is a literature review of cloud computing technology architectures and related research. The third section describes the methodology and procedures for collecting data on a cloud computing technology architectures. The fourth section presents the results for the unified cloud computing technology architecture. The fifth section compares results among the cloud architecture of Microsoft, Trend Micro, and the proposed unified cloud technology architecture. Finally,conclusions are presented and suggestions are provided for further research.

2. Literature Review

2.1 Cloud Computing

Cloud computing refers to network computing. The term “cloud” is used as a metaphor for the Internet, based on a previous cloud drawing used to represent the telephone network, and later to depict the Internet in computer network diagrams as an abstraction of the underlying infrastructure. Cloud computing is not simple, and researchers have offered many definitions of cloud computing, which are schematically represented in Table 1.

Table1: Table of cloud computing definitions

The collective view of cloud computing promoted by scholars and research institutions indicates only a partial range of difference. Based on these various definitions, this work defines cloud computing as a large-scale virtualized resource that combines several hardware resources to conduct parallel and distributed computing. Users can dynamically adjust computing resources for rapid deployment according to their needs. The payment includes pay-as-you-go and various Internet applications to deliver services.

Cloud computing is not a new technology, but has evolved from the old technology and ideas. In the 1980s,computing evolved from distributed computing to grid computing, which was used to solve large-scale computing problems. Grid computing is a virtual supercomputer that combines many computers from different locations with the Internet. Each computer in the virtual supercomputer is a node. The computing method divides a large-scale computing problem into many small computing problems,which are assigned to different nodes to compute. Grid computing has two advantages: the full use of idle computing resources and having powerful computing capabilities.

SETI@home[17]is the search for extra terrestrial intelligence at home and a typical application of grid computing. The project started in early 1999 and was combined with a worldwide distribution of computers to become a computer array for detecting intelligent extraterrestrial life. With over 5.2 million participants worldwide, the venture is a distributed computing project with the most participants to date. The original intent of SETI@home was to use 50000 to 100000 home computers.Since its launch on May 17, 1999, the project has logged over 2 million years of aggregate computing time. On September 26, 2001, SETI@home had performed 1021 floating-point operations and is acknowledged by the Guinness World Records as the largest computation in history.

The origin of utility computing began as early as 1961,when John McCarthy mentioned in a speech delivered at the Massachusetts Institute of Technology: “... maybe one day, computing will be as a public utility for organizations to use, just like the telephone system. Utility computing may become a new and important industrial base.” In other words, computing resources may become general services similar to water, electricity, and gas. Users pay according to the actual amount of resources used according to demand.Therefore, users lease resources rather than purchasing them. In 1998, HP practically applied utility computing by establishing utility computing departments, and allowed customers to register for services and perform self-administration online[18].

SaaS is a computing model that offers software as a service to users through the Internet, according to their needs and customized services. Users do not need to purchase software to install on their computers, but can rent software from software vendors through the Internet. The hardware requirements for running the software are lower than those for using the traditional approach, and users do not have to be concerned about software maintenance or upgrades. SaaS first appeared in 2000, with the vigorous development of the Internet. Various Internet-based business models have continued to emerge, such as the Salesforce customer-relationship management system.Salesforce provides custom relationship management(CRM) system services to users through the Internet to facilitate integrating the system with the existing enterprise enterprise resource planning (ERP) system. Users can view the results to achieve a rate of customer contacts, historical information, opportunities, statuses, and other information on the Internet. Cloud computing has evolved from these computing modes, and combines their merits. By combining distributed computing resources into one large virtual computing resource, such as grid computing, users can apply to suppliers of customized services such as SaaS,which involves using virtualization technology to provide computing resources through the Internet, based on user demand. The user pays the supplier according to used services, resources, and the amount of time used in utility computing. Therefore, users can improve hardware resource usage and, to reduce costs, eliminate both the need to purchase high-specification hardware and the cost of software maintenance and updates.

2.2 Cloud Computing Architecture

Many cloud computing providers have developed their own cloud architecture. This study collected and compared cloud architectures from cloud providers and researchers.The cloud computing service was first launched in 2006 by Amazon Web Service, which began by providing an Amazon IaaS to allow users to rent computing, storage, and other resources applicable to businesses of any size.Amazon later added a wide range of cloud services. The Amazon cloud computing architecture includes computing,content delivery, database, deployment and management,messaging, networking, payment and billing, storage,support, web traffic, and workforce.

In February 2010, Microsoft launched a cloud computing platform, Windows Azure[19]. Microsoft also offers private cloud solutions for enterprises. Enterprise customers can import cloud computing using three applications: (1) move to the Cloud, an enterprise application system which transfers all data to the Microsoft data center; (2) use the Cloud, in which the user uses Microsoft cloud services without developing or importing;and (3) be a Cloud, in which the enterprise creates the company’s own private cloud that both internal and external customers can use. Microsoft cloud computing architecture includes Windows Azure, SQL Azure,Windows Azure AppFabric, and Windows Azure Marketplace, as shown in Table 2.

Table 2: Microsoft cloud computing architecture

Microsoft cloud computing services mainly focus on PaaS and SaaS. The PaaS is Windows Azure, which is designed to simplify the daily operations of developers and system administrators, rather than provide an in-house deployment environment. The application can have one or more role structures, such as web, background work, or virtual machine (VM) created through Windows Azure. A web role is used to create the environment necessary for implementing web-based applications, by using various programming technologies such as PHP, Java, and ASP.net to create an application. The background work role is used to perform various codes to perform simulations, video processing, or other operations. The VM role can run the Windows Server 2008 R2 image, and allows users to move internal deployment of Windows Server applications to Windows Azure.

SQL Azure is compatible with existing data access technologies and tools based on the existing Microsoft SQL Server. SQL Azure provides database management system(DBMS) to store data in the Microsoft server in the data center. SQL Azure also provides SQL server reporting services (SSRS) run in the cloud, for establishing and distributing the standard SSRS report. The SQL Azure database with the SQL server can synchronize data processing in Microsoft data centers at different locations.Each SQL Azure account has one or more logical servers used to organize account information and billing.

Windows Azure AppFabric is based on Windows Azure,and the components include service bus, access control, and caching. Service bus is used to open the application on the Internet. Access control allows users to obtain digital identification, control the content of different users who can access caching by creating a cache to accelerate the application to access speed, and reduce applications to query the number of databases.

Windows Azure Marketplace allows users to find the applications and data they need and includes two components, AppMarket, and DataMarket. Users and applications can access information through DataMarket at a single location to make purchases and access various commercial data sets. AppMarket is a trading platform for cloud applications.

The Gartner research report released in January 2011 revealed that business-to-cloud will move toward developing a hybrid cloud that can help companies use internal and external IT resources flexibly to create a balance point of resource applications[20]. Microsoft also provides private cloud solutions to the public cloud, which provides private cloud-based access to create a seamless match and operation of the hybrid cloud, in which the workload between public and private cloud obstacles shifts back and forth. The private cloud MCloud includes mainly self-service operation management, service management,provision engine, and collection of all types of data for resource performance. The self-service provides users with self-service management through the self-service portal.Operation management allows data center managers to audit open VM equipment resources. Service management users can manage the VM service life cycle and monitor health statuses. Provision engine establishes a default virtual operating system and application services. Resource performance data is collected through the collector and the performance-owned performance data, including VM CPU,RAM, and disk space. Reporting capabilities provide statistical reports of virtual machine applications.

The MCloud software chart shows Microsoft private cloud solutions based on virtualization of the basic building and its use of Microsoft Hyper-V virtualization technology.This forms the basis for providing THE OA cloud, database cloud, desktop cloud, App cloud AND compute cloud as private cloud services. Microsoft offers cloud services focused on PaaS and SaaS to produce the core value of cloud computing and IaaS private cloud solutions. This paper presents a cloud computing architecture that uses the core of Microsoft cloud computing technology, from which Microsoft cloud services are derived.

Trend Micro used THE cloud computing technology in the field of information security in 2008 to provide a cloud security solution. In 2010, Trend Micro established a cloud computing department independent of the establishment of a subsidiary: TCloud computing[21]. The Trend Micro cloud computing architecture includes Elaster and a Elaster cloud application platform (CAP), as shown in Table 3.

Trend Micro is the leading computer antivirus and network security vendor worldwide and has recently positioned itself as the cloud-network security expert. Their business purpose is to “develop, provide, and support the most flexible and powerful customized Internet security solutions to deal with the ever-changing security threats on the network.” To develop the smart protection network through cloud computing, Trend Micro Investment established TCloud computing in 2010, with complete ownership, to promote cloud computing technology.TCloud computing provides a cloud operating system,technical support, systems integration, a cloud computing platform, education training and consultancy, and other related services.

In the development of the initiative smart scan technology, Trend Micro faced several challenges: the need to manage an enormous amount of daily data, the priority of program, and focus on highly relevant to the threat.Trend Micro’s solution is the open source Hadoop platform to construct the cloud architecture for powerful computing capability.

Table 3: Trend micro cloud computing architecture

A subsidiary of Trend Micro, TCloud Computing,committed to creating a leading open cloud computing platform, allows enterprises to have greater business agility and to create more cost-effectiveness. TCloud Computing provides two solutions: the IaaS solution Elaster and the PaaS solution Elaster CAP. Elaster is a cloud-in-a-box solution, which provides system scalability and business agility through integration within enterprise management,application, and IT environments. Elaster provides computing, storage, and network services as the IT infrastructure and supports provision, multi-tenant,optimization, security, configuration, and monitoring technologies. Elaster CAP is a solution of performance and flexibility of cloud computing application development and implementation, in which users can quickly develop SaaS.The Elaster CAP can be divided into two phases: Elaster CAP SDK and the Elaster CAP Application Run-time Environment. The Elaster CAP SDK provides the user with four packages for development and testing, including the Elaster CAP API Library, the Eclipse Plug-in for IDE Development, the Command Line and the Script for non-GUI Development, and the Standalone Run-time Environment. The Elaster CAP application Run-time Environment provides a web application and a MapReduce application suite to deploy in the cloud, implementation,and operation action. The web application kit contains Java(JSP 2.1, Servlet 2.5, J2SE 5, Java SE 6), PHP 5.3, an HTML, JavaScript, and other programming language technology.

The Trend Micro cloud computing architecture in IaaS still uses the abstraction of several server virtualization technology resources for further management and configuration. The Hadoop technology in PaaS is used to construct the platform, using the HBase and Hadoop distributed file system (HDFS) as the database and file system architecture to achieve multi-tenant configuration through VM technology. The software development kit is used as a tool to create the application and execution environment. The MapReduce technology is used as the SaaS computing base to establish monitoring and analysis applications. The cloud technology architecture proposed by this study uses the core technology used in the cloud computing architecture created by Trend Micro, which is based on these core technologies in accordance with the services and solutions provided by the amplification of cloud application technology.

Cloud computing is based on utility computing to deliver services and integrate Smart Home with the cloud architecture. In contrast to general cloud computing, web services and peer-to-peer technologies are combined into a cloud. The most critical aspect of this architecture includes a service interface, service management, and service catalogues. The platform layer includes a resource management module and a security management module.The infrastructure layer provides cloud computing services to develop management and a large number of physical resources through the virtualization technology[22].

Litoiu offered a cloud computing architecture for optimized solutions of cloud computing. The IaaS layer used as the virtualization of hardware resources includes storage, computing power, and memory. The PaaS layer provides platform services such as the web, applications,and database environments, in which the user can deploy and test operations in this environment. The SaaS layer provides end users single or composite application services[4].

Traditional information security products cannot ensure that cloud service providers will provide services within the scope of user needs. Therefore, the study has established a trusted cloud computing environment between cloud service providers and users through a multi-tenancy trusted computing environment model. The multi-tenancy trusted computing environment model is a delivery model designed for IaaS, which is used to increase user confidence in cloud computing. The Cloud Security Alliance cloud computing architecture is used for this study reference. The IaaS includes hardware abstraction, whereas PaaS includes the integration and middleware, and SaaS includes applications[23].

The resource interface layer is an agreement between hardware resources, which provides virtualization, a database, and the distributed systems technology. The system layer is responsible for the daily operation of the cloud management system, such as submitting jobs,managing user accounts, and monitoring service quality.The business layer with the prices and quality of service agreements regulates the supply and demand of resources,and allows users to save and manage a personal VM. The user interface layer provides different interfaces that allow users to operate and access the cloud system.

Findings from the cloud-architecture literature review and analysis show that the cloud computing architecture can be classified into three parts: the IaaS, PaaS, and SaaS.The current study compiles the cloud architecture and lists the technology used by more than three researchers or enterprises, which are defined as key technologies and listed in Table 4.

Table 4: Key technologies of cloud computing

3. Methodology

This paper reports a thorough literature review and studies the compiled cloud computing technology architecture, and describes the evolution of a single technology by showing the overall evolution of the cloud computing architecture. The following steps are shown in Fig. 1.

The primary purpose of technology is to understand and create. The ability to leverage the cloud computing technology to solve problems and adapt to new technological environments must derive from a macro viewpoint to understand the evolutionary history of cloud computing and to learn the necessary skills and knowledge for using the cloud computing technology efficiently.Therefore, this work investigated the evolutionary process of relevant technologies in the architecture through literature review.

4. Results

4.1 Technology Evolution

A. Distributed Computing

The concept of distributed computing has a close relationship with the development of a computer network.Various computers connected through the communication line that allow the terminal to access multiple computers that can also pass data to each other. In 1979, the U.S.Department of Defense established The Advanced Research Projects Agency Network (ARPANET) to connect to a remote computer, thus exhibiting the prototype of today’s Internet. However, the computer network focused on connecting different computers and operating systems and therefore needed to develop a protocol. In 1983,ARPANET changed the network core protocol to TCP/IP,which was proposed by Bob Kahn and Vinton Gray Cerf in 1974. In the 1990s, the entire network was open to the public. In 1996, the term Internet became widespread,referring to the entire World Wide Web (WWW). The development of the Internet has resulted in a client-server computing mode. The personal computer can communicate with other computers through the Internet. The client makes a request over the network to the server-side, and the server-side responds to client information through the Internet. In the client-server architecture, the client must install special software. Consequently, the client can balance server-processing capacity; however, the hardware requirement of the platform is relatively high, because the client must install special software.

Fig. 1. Research steps of this study.

The European Organization for Nuclear Research defined grid computing as follows: “Grid computing is through the Internet to share computing power and data storage capacity.” Grid computing is a “virtual supercomputer” conducted through many combined computers from different locations on the Internet.SETI@home is an example of grid computing.

In 2004, Google released a new programming model,MapReduce, which is used to handle large amounts of data and applications in the Google search engine. The basic concept of MapReduce is to divide the work into two parts,Map and Reduce. Map is responsible for dividing the problem into several problems that are sent to different machines, and Reduce collects the results of each machine processing and combines them into a single answer that is distributed computing.

In 2006, the Apache Software Foundation released the Hadoop project, which is an open source. Hadoop is a distributed computing programming tool developed by Apache, in which the content includes the HDFS and MapReduce. In 2008, Yahoo used 10000 microprocessor cores of the Linux computer clusters running Hadoop applications. Currently, Hadoop is often used to create a cloud computing platform environment.

B. Virtualization

The virtualization technology can be traced to Mainframe (1960-1970). Mainframe was expensive,because of the need to operate a single machine and provid time sharing, not only to allow different users to test the software in the same hardware platform but also to save costs. Strachey[25]first used the time-sharing concept to implement multi-programming in time sharing in large, fast computers. Using the underlying hardware architecture abstraction, the VM simulates underlying hardware resources, and implements the application on a VM. This method is similar to implementation in a real hardware platform, and has an independent operating system. The virtualization technology of an operating system was invented by IBM, and includes many operating systems simultaneously used by Mainframe. In 1997, Connectix created PC virtualization on Macintosh, and the v4.0 of this product supported the Windows operating system in 2001.In 1999, VMware announced the VMware Virtual Platform using x86-platform virtualization technology. Currently,VMware is the global leader of the x86 virtual platform,which has more than 250000 users. IBM has been the leader in virtualization, developing the DLPAR, which distributes computing resources to independent partitions,such as CPU and RAM without restarting the system, and providing flexible resources and convenient management.Milind Yedkar is a solution plan marketing director at Dell who developed virtualization through five stages on the IT news website in 2008. The first stage integrates the server to enhance the utility rate, reduce costs, and simplify the IT infrastructure. The second stage achieves dynamic workload migration through virtualization, in which IT resources dynamically allocation and adjustment does not affect the business premise. The third stage improves service quality through virtualization, and backs up or restores files without increasing IT costs. The fourth stage configures the resources required by departments or units through virtualization technology to achieve workload management. The fifth stage is based on virtualization technology to achieve disaster recovery operations,according to the needs of a local or remote backup and fast recovery operation in a tolerable time, without the need to invest in high-level IT equipment. Virtualization not only enables management of resources but can also improve service quality further to provide a higher value.

The virtualization technology includes computer virtualization, network virtualization, storage virtualization,computer virtualization of actual hidden computing resources (such as CPU, memory), and virtual budgetary resources available for software execution. Network virtualization combines different network hardware and software resources into a virtual whole, or a network into several independent LANs. Combined storage resources of storage virtualization cause more than one entity to become a large-scale storage resource, or a physical storage resource, and are divided into multiple independent storage resources.

C. Database

The database can be regarded as an electronic file cabinet that is a computerized data storage system used to store electronics. Users can add, delete, modify, and query operations for filing data in the file. The database can be divided into two parts: database storage, that is, the organization of data and storage methods, and a database management system (DBMS). The DBMS typically includes storage, retrieval, security, backup, and other infrastructural functions. A database typically refers to the database management system. Before the invention of computers, data were incompatible to file management, and were archived through paper records. This information-processing mode has many disadvantages and inconveniences, such as difficulty in searching, no simultaneous access, and limited storage and space. After the emergence of computers, data processing entered a sequential computer file system using a punch card, tape,paper tape, and other media to record information. With the development of the drive, direct access replaced the old generation of tape, paper tape, and other media to become the direct access file system. The file system of direct access can directly access each file, but the data processing files still lack flexibility. A database system for processing units to records began to appear, including a file subdivided into unit record data. These data were stored in the computer, thus forming a “centralized database system.”

Database development at this stage was maturing gradually and using the concept of a centralized database system to extend different data models: the hierarchical data model, network data model, and relational data model.The hierarchical database system organized records by using a tree structure, an appropriate description of the one-to-many information relationship. The concept of the network data model involves using a network-like structure to organize records. However, these two data models use a link method to string records together. Codd[26]proposed a solution to this problem, the “relational data model,”emphasizing that no dependencies should exist between the internal structure of data and the application, or “data independence.” In the relational organization, the application faces only the record of the form composition.With the development of object-oriented programming language, the concept of object-oriented applications on the database involves using objects to represent all information in the database, which represent specialization,generalization, and aggregation through an object in the real world. The three-tier application architecture of the Internet (client, web server, and database server), gradually distributes the data to all stations in the network, improves the processing speed of the data points, and improves the reliability of data access; that is, distributed database systems. In 2004, Google used their own distributed file system to build a BigTable used to store structured data. In 2006, Hadoop also formulated a plan called the Hbase,which is similar to the BigTable. Using a distributed file system, the table is split into many parts, and a different server is responsible for accessing different parts to achieve a high-performance database.

D. Web Service

In accordance with Curbera[27], the main Internet open technology constitutes a network service of four standards:XML, SOAP, WSDL, and UDDI. The Internet service is a software component, building on HTTP, XML, SOAP,WSDL, and other standard protocols, having openness and dispersion characteristics. Integrating different machines and heterogeneous platforms can work together through the Internet operations. UDDI (universal description discovery and integration) provides a standard for registration and for searching web service information. The WSDL (web service description language) is the mode of operation that describes a network service, indicating the interaction of client and web services. The SOAP (simple object access protocol) is a simple protocol for exchanging structured and type information on a network, jointly developed by IBM,Microsoft, and other vendors to provide a standard set of specifications for message encoding and writing.

The early development stage of network services(1999-2001) included many Internet service standards, in which various software companies developed their own web service development tools (Microsoft, IBM, Sun,Oracle, etc.). In 2002, numerous commercial Internet services began to appear in the business to consumer(business to the consumer; B2C) Internet service. Adopting the common consumer as the object, such as in 2002,Microsoft launched the work service, “My Services,” based on Microsoft.NET. In 2003, UDDI announced the official opening of private registered network services, and many public network services and the government Internet service began to appear. In 2004, the prevalence of private registered network services had become a new venture pipeline. In 2005, Business to Business (B2B) Internet services began developing, and dynamic network services also began to receive attention. After 2005, mature mobile Internet technologies, coupled with the popularity of smart phones, have caused the Internet to expand to a wide range of Internet services introduced for personal and innovative network services (online buying, various types of applications APP). Several traditional methods provide web services such as Internet telephony. Because Internet use has created the rise of social networking sites, it is not only a convenient tool but also a vital channel of communication between people. The current concept of cloud computing is spreading rapidly; hence, performing an Internet operation has become increasingly large and complex, substantially more complexity than using the original PC.

4.2 Unified Cloud Computing Technology Architecture

This study extracted key technologies in cloud computing and combined them into a unified cloud computing architecture through the literature review. Fig. 2 shows the evolution of cloud computing architecture.

Fig. 2. Cloud computing architecture.

According to different types of cloud services, the cloud computing architecture is divided into three parts: IaaS,PaaS, and SaaS, all of which require security technology to maintain service quality. Therefore, the outermost layer is security-related technologies. An IaaS is mainly provided by various computing resources, such as storage, computing,and service management. PaaS provides a development and testing platform, database, business intelligence, integrated systems, and other platform services. SaaS provides a wide range of services, and the major suppliers can be broadly classified as file management, collaboration, billing,customer relationship management systems, backup and recovery, finance, sales, social networks, and content management, demonstrating a wide range of applications.

Physical resources must be virtualized to provide an IaaS, such as computing, storage, and network through the server management, platform management, and application management techniques to the planning of virtualized resources. Providing PaaS requires using IaaS technology through distributed computing. The platform should contain the Hadoop cloud environment platform, the HDFS distributed file system, the HBase database environment,and a virtual machine, which can be run using a virtual machine operating system and software development tools to the index technology. Offering SaaS requires using IaaS and PaaS technologies that employ MapReduce to analyze and synthesize the results, and uses control and status analysis techniques to manage the application. Signaling functions, such as e-mail and electronic messages technology are also necessary. To meet the demand for solutions to develop a suitable software system to provide services requires information systems, applications and other related technologies.

The cloud computing technology architecture diagram in this paper is quite complex and combines different fields of technology; therefore, the cloud vendor should provide cloud services and construct a technical architecture of its services that are often locked within a specific function.

5. Discussion

5.1 Mapping between Microsoft Cloud Computing and the Proposed Architecture

The Microsoft cloud computing can be classified as public and private clouds. Windows Azure is mainly a public cloud, which provides PaaS and SaaS. The public cloud includes Windows Azure, SQL Azure, Windows Azure AppFabric, and Windows Azure Marketplace. In Windows Azure, Microsoft used virtualization of computing, storage, and a network to build a large resource base. Users can use these virtual resources as physical resources and work on Windows Azure. Therefore, users of Microsoft technologies can map to IaaS and PaaS in the proposed architecture. The technologies include a software development kit, a virtual machine, operating systems, and a distributed file system. In SQL Azure, Microsoft uses technologies based on the Microsoft SQL Server, which process data through distributed computing that can be mapped to PaaS in the proposed architecture. Windows Azure AppFabric controls user privileges and accelerates the access speed of the application by the application of caching services. The technologies of Windows Azure AppFabric can be mapped to the management of IaaS and security in the proposed architecture. Windows Azure Marketplace allows users to exchange applications and data,and the technologies of Windows Azure Marketplace can be mapped to the index technology of PaaS in the proposed architecture. For using a private cloud, Microsoft has a Mcloud solution that provides IaaS. In Mcloud, most Microsoft technologies belong to management technologies and resource virtualization, thus enabling them to map to IaaS in the proposed architecture. Enterprises can use a private cloud (Mcloud) and a public cloud (Windows Azure)to have a hybrid cloud.

5.2 Mapping between Trend Micro Cloud Computing and the Proposed Architecture

Trend Micro cloud computing can also be classified as public and private clouds. The smart protection network is a SaaS belonging to a public cloud, whereas the Elaster and Elaster CAP are PaaS, which belongs to a private cloud.The smart protection network uses Hadoop to build a platform and create a database and file system by using HDFS and HBase. Based on Hadoop, the smart protection network collects various types of information from the HTTP, DNS, and FTP by using Clawer, and analyzes network threats through a relational approach. Most technologies that use the smart protection network can map to the proposed architecture, including virtualization,platform and application management, Hadoop, HBase,HDFS, and MapReduce. The solution Elaster of IaaS and the PaaS solution Elaster CAP both belong to a private cloud. Elaster provides numerous virtualized resources to provide enterprises with system scalability and business agility. Elaster uses virtualization of computing, storage,and networks to build an IT infrastructure, and uses multi-tenant, optimization, information security,configuration, monitoring, and technology supply. All of these technologies can map to the proposed architecture.Elaster CAP is a solution of application development and implementation. During the development phase, Elaster CAP uses a development tool, such as the Elaster CAP API library, eclipses plug-in for IDE development, command line and script for non-GUI development, and the standalone run-time environment. During the operational phase, Elaster CAP provides a web applications package and a MapReduce application package. Most of these technologies can map to the proposed architecture.

5.3 Comparisons of Cloud Computing Among Microsoft, Trend Micro, and the Proposed Architecture

This paper presents a cloud computing architecture in IaaS, which includes computing resources and virtualization of computing resources (computing, storage,and network), and three service levels of management (the server, platform, and application). In the proposed architecture, PaaS includes Hadoop for a cloud environment, HDFS for a distributed file system, HBase for a database environment, a VM for running a virtual operating system and software development tools, and the index technology for the service index. SaaS decomposes the problem and combines the results through MapReduce,and uses information systems and application technologies to develop SaaS. Table 5 summarizes the comparisons of cloud computing among Microsoft, Trend Micro, and the proposed architecture.

6. Conclusions

This study presents a cloud computing architecture and explains the core technologies of cloud computing. This work also reviews the evolution of cloud computing through a discussion on several key technologies. After comparing the proposed cloud computing architecture with the architectures of Microsoft and Trend Micro, the proposed cloud computing architecture is shown to be able to cover the core technologies of the cloud computing architectures of both Microsoft and Trend Micro. These findings support the assertion that the proposed architecture is practical and can help both researchers and enterprises understand cloud computing. For researchers, this architecture can clearly define the scope of cloud computing research focused on the core technology of cloud computing. For enterprises, the architecture can help businesses assess whether to import the cloud computing technology, or to determine the type of cloud computing services to use.

Table 5: Comparison of Microsoft, Trend Micro, and the proposed architecture

[1] E. Laszlo, “Information technology and social change: An evolutionary systems analysis,” Behavioral Science, vol. 37,no. 4, pp. 237-249, 1992.

[2] Z. Li, S. S. Chaudhry, and S. Zhao, “Designing ERP systems with knowledge management capacity,” Systems Research and Behavioral Science, vol. 23, no. 2, pp. 191-200, 2006.

[3] L. Zhang and Y.-C. Li, “Theory and practice of systems methodology in ERP implementation,” Systems Research and Behavioral Science, vol. 23, no. 2, pp. 219-235, 2006.

[4] M. Litoiu, M. Woodside, J. Wong, J. Ng, and G. Iszlai, “A business driven cloud optimization architecture,” in Proc. of the 2010 ACM Symposium on Applied Computing, Sierre,2010, pp. 380-385.

[5] G. Zervas, V. Martini, Y. Qin, E. Escalona, R. Nejabati, D.Simeonidou, F. Baroncelli, B. Martini, K. Torkmen, and P.Castoldi, “Service-oriented multigranular optical network architecture for clouds,” Journal of Optical Communications and Networking, vol. 2, no. 10, pp. 883-891, 2010.

[6] M. Creeger. (Feburary 2009). Cloud computing: an overview.[Online]. Available: http://queue.acm.org/detail.cfm?id=1554608.

[7] Gartner. (Feburary 2011). Gartner executive programs worldwide survey of more than 2,000 CIOs identifies cloud computing as top technology priority for CIOs in 2011.[Online]. Available: http://www.gartner.com/ it/page.jsp?id=1526414

[8] OpenCrowd. (Feburary 2012). Cloud computing taxonomy.[Online]. Available: http://cloudtaxonomy.opencrowd.com/taxonomy/

[9] P. Mell and T. Grance. (Feburary 2011). The NIST definition of cloud computing. [Online]. Available: http://www.nist.gov/itl/cloud/upload/cloud-def-v15.pdf

[10] IBM. (Feburary 2012). Cloud computing overview. [Online].Available: http://www.ibm.com/cloud-computing/us/en/

[11] Wikipedia. (Feburary 2012) Cloud computing. [Online].Available: http://en.wikipedia.org/wiki/Cloud_computing

[12] L. M. Vaquero, L. Rodero-Merino, J. Caceres, and M.Lindner, “A break in the clouds: towards a cloud definition,”SIGCOMM Comput. Commun. Rev., vol. 39, no. 1, pp.50-55, 2008.

[13] R. Buyya, C. S. Yeo, S. Venugopal, J. Broberg, and I.Brandic, “Cloud computing and emerging IT platforms:Vision, hype, and reality for delivering computing as the 5th utility,” Future Generation Computer Systems, vol. 25, no. 6,pp. 599-616, 2009.

[14] M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. Katz, A.Konwinski, G. Lee, D. Patterson, A. Rabkin, I. Stoica, and M.Zaharia, “A view of cloud computing,” Commun. ACM, vol.53, no. 4, pp. 50-58, 2010.

[15] S. Marston, Z. Li, S. Bandyopadhyay, J. Zhang, and A.Ghalsasi, “Cloud computing—the business perspective,”Decision Support Systems, vol. 51, no. 1, pp. 176-189, 2011.

[16] T. J. Bittman. (Feburary 2010). Clarifying private cloud computing. [Online]. Available: http://blogs.gartner.com/thomas_bittman/2010/05/18/clarifying-private-cloud-comput ing/

[17] SETI @ Home. [Feburary 2012]. [Online]. Available:http://setiathome.berkeley.edu

[18] Wikipedia. (Feburay 2012). Utility Computing. [Online].Available: http://en.wikipedia.org/wiki/Utility_computing

[19] D. Chappell. (Feburary 2010). Introducing Windows azure.[Online]. Available: http://go.microsoft.com/?linkid=9682907

[20] J. Brodkin. (Feburary 2008). Cloud computing hype spurs confusion, gartner says. [Online]. Available: http://www.computerworld.com/s/article/9115904/Cloud_computing_hy pe_spurs_confusion_Gartner_says

[21] Trend Micro. [Feburary 2012]. TCloud computing. [Online].Available:http://www.tcloudcomputing.com

[22] Z. Wei, S. Qin, D. Jia, and Y. Yang, “Research and design of Cloud architecture for smart home,” presented at the IEEE Int. Conf. on Software Engineering and Service Sciences,Beijing, 2010.

[23] X.-Y. Li, L.-T. Zhou, Y. Shi, and Y. Guo, “A trusted computing environment model in cloud architecture,” in Proc. of the 2010 Int. Conf. on Machine Learning and Cybernetics, Qingdao, 2010, pp. 2843-2848.

[24] F. Doelitzscher, A. Sulistio, C. Reich, H. Kuijs, and D. Wolf,“Private cloud for collaboration and e-learning services:from IaaS to SaaS,” Computing, vol. 91, no. 1, pp. 23-42,2011.

[25] C. Strachey, “Time sharing in large, fast computers,” in Proc.of Int. Federation for Information Processing Congress,Paris, 1959, pp. 336-341.

[26] E. F. Codd, “A relational model of data for large shared data banks,” Communications of the ACM, vol. 13, no. 6, pp.377-387, 1970.

[27] F. Curbera, M. Duftler, R. Khalaf, W. Nagy, N. Mukhi, and S. Weerawarana, “Unraveling the Web services web: an introduction to SOAP, WSDL, and UDDI,” IEEE Internet Computing, vol. 6, no. 2, pp. 86-93, 2002.