Cloud computing model based on natural cloud

Cloud computing model based on natural cloud tangmin Beijing posts and Telecommunications Pcn&cad Center, Peking (100876) e-mail:tangminbupt@gmail.com

Absrtact: In network autonomous systems such as Web Service, a system with better communication capability, robustness and security is designed according to cloud computing model. This model is mainly composed of the quantity dynamic controllable sub middleware and the unified management platform, through the dynamic management and allocation of the pair middleware to meet the needs of different types and sizes, and shielding the function of the low-level implementation, so as to improve the performance of the Web Service system. This paper describes the cloud computing model based on simulating the life cycle and shape of cloud in nature, introduces the life cycle model and working model of cloud computing model, and finally introduces the latest results of application cloud computing model: Amazon Elastic Computing Cloud System

Keywords: cloud computing model, WEB Service, autonomous system

1. Overview

Computer network Since its inception, its main role is the sharing of information and communication, the greatest value of the network is that users can really manage the business and share information, rather than simply using the network itself. Through years of development, the network also from the initial simple file sharing development to the current multimedia content sharing, computing resource sharing. The size and function of computer networks are expanding over time.

As a result of this demand, new technologies have been applied to the network, and the efficiency and ease of use of new technologies have further facilitated the development of the network. Higher service requirements depend to a large extent on the effectiveness of Network Computing [1]. The emergence of more effective methods of organization and computation has always shocked people and began to rethink the fundamentals of dealing with things.

The service-oriented framework (SOA, service-oriented architecture) is one of the trends in the current network organization approach [2], a network system built using the SOA approach (hereinafter referred to as SOA System by providing a unified middleware products to provide services for more and more network products, at the same time relying on the scale and integration of the system, so that the system can support the scale of services multiplied to meet the increasingly developed network service requirements.

However, the diversity of network services also poses a great challenge to SOA and Web service [1,3], in order to be able to support different sizes and types of services, SOA systems need to dynamically allocate system resources based on changing requirements [4], an SOA that meets this requirement The system is called Dynamic autonomous SOA system. To meet this requirement, dynamic autonomous SOA systems must have the following characteristics [5]:

Unified external interface and the low-level implementation of the component-independent, collaborative sub-middleware Group [6] dynamically increase and decrease the ability of Web services

The cloud computing model described in this paper has the characteristics its structure and life cycle are similar to those of nature, and in SOA systems designed with cloud model architecture, component-based sub-middleware has the independence of time and space (similar to the water molecule in nature) when it does not provide service. By consolidating these sub middleware (similar to the process of cloud formation in nature), cloud computing models provide a unified, scaled service (similar to the cloud), while the size of the service is proportional to the size of the cloud and can be dynamically adjusted. This is the conceptual source of the cloud computing model.

2. Cloud computing Model

The cloud model is the model concept which is applied to the computer network design according to the generation Way and form of cloud in nature, and then the cloud computing model is derived by introducing the life cycle and shape of the natural cloud.

2.1 Work Model

To enable the system to be dynamically tuned and autonomous in the design process of the SOA system, SOA systems must have modular capabilities in both vertical and horizontal systems [5], and vertical modularity enables the system to be abstracted at any level so that when the system is not adapted to new requirements, it can be redefined by a minimum cost. 5,6]. On the basis of vertical abstraction, the lateral modularity is composed of distributed services, that is, the system is designed as a combination of many of the same functional modules, according to the needs of users dynamic allocation of the number of modules, so that the system has dynamic adjustment and autonomy capacity. The final system diagram is shown in Figure 1.

From the system shown in the diagram, we can see that the cloud computing model has the following characteristics:

The system has a unified interface, and the WEB service customer connects to the SOA system through the cloud model boundary to obtain services. Cloud computing model for the bottom of the external shielding system implementation, in any Web service customer service angle, the system is represented as a unified, to meet the needs of the single logical entity. The cloud computing model consists of independent logical components, each with space and time independence from each other. At the same time, the components and the system itself are also independent, components can be dynamically joined the system or uninstall from the system. By dynamically provisioning components to meet different requirements, the system has the ability to provide higher levels of service at a macro level.

Comparing the cloud computing model with the real cloud in nature (Fig. 2), the above features are also characteristic of the cloud [7]: A uniform appearance of the external, shielding the internal water molecule form, consisting of many independent droplets, forming rain under certain conditions. Similar to these properties, the cloud computing model is named to Figure 2 Natural cloud shape schematic source.

2.2 2 Lifecycle Model

The cloud computing model has the possibility to satisfy the dynamic autonomous SOA system with the feature of longitudinal and transverse component in structure, and in order to obtain the dynamic adjustment, the system must have a reasonable life cycle model which is composed of system, system adjustment and system decomposition. At this point, the cloud computing model and the natural cloud still have a good coincidence. Before introducing the model of cloud computing, let's introduce the life cycle model of the cloud in nature [7]. First, water in the land is exposed to the sun, and water molecules get out of the attraction of other water molecules and into the atmosphere, and this process is what we normally call evaporation.

Water vapor from the Earth's surface into the lower atmosphere, where the temperature is high, the moisture contained more, if these hot and humid air is lifted, the temperature will gradually reduce, to a certain height, the air vapor will reach saturation. This phase can be called a bonding phase.

If the air continues to uplift, there will be excess vapor precipitation. If the temperature is higher than 0°c, the excess vapor condenses into small water droplets, and if the temperature is below 0°c, the excess water vapor is condensed into small ice crystals. As these small water droplets and small ice crystals gradually increase and reach the level of human eye recognition, the cloud forms.

Finally, when the droplets in the cloud grow large enough, the atmosphere cannot keep it rising, so it starts falling into rain or falling into other clouds. The cloud ended his life cycle.

The above is the whole process of the life cycle of the cloud in nature, and it can be seen that it contains the four important steps shown in Figure 3: evaporation, bonding, condensation, rain [3,7].

Like the life cycle of a natural cloud, the computing Cloud's lifecycle model contains four important steps [3]:

first, find the available resources in the system based on requirements-discovery phase [2].  Integrate the found available resources into our existing systems to form a new system that meets the requirements--the combination phase. Deploy new systems for users--molding phase. When we don't need some features in this new system, we can do some of these functions

Decomposed from the whole system--the decomposition phase.

In contrast to the natural cloud formation process and cloud computing model described earlier, the natural cloud formation process consists of four
Steps: evaporation, bonding, condensation and rain, and the instantiation of the cloud computing model also includes four steps: Discovery, combination,
Type, decomposition, as shown in Figure 4, there is a similarity between the steps and the processing of the content:

resources available on the Internet are analogous to water on the Earth's surface. The process of discovering the available resources is analogous to the evaporation of the surface water. The process of combining available resources with each other is analogous to the process of combining water molecules into droplets. Combinatorial logic is analogous to the condensation process of water molecules. The process of decomposition is analogous to the rain of clouds.

Therefore, from the working model and the life cycle model two, the cloud computing model provides an easy to understand and effective solution for the design of the SOA system by simulating the composing ways and characteristics of the natural cloud, and through this low coupling system design method [3].

2.3 System Analysis

Compared to Figure 5 and Figure 6, you can find that the main advantage of the SOA system using cloud computing model is that, using the independence and reusability of sub-middleware, all middleware services are used as a combination of sub-middleware-"cloud" to provide services externally, so when the system needs to be redeployed for new services, The main process added is to add or find available sub-middleware (as shown in Figure 4), which in turn constitutes a new middleware service. When a general SOA system adds new services, the basic unit to be added is the entire middleware [1,2,4].

So from the point of view of the cloud, in the dynamic change of the system, the SOA system with cloud computing model needs only to make necessary modifications to the internal computing cloud, while maintaining consistency on the external interface and system level while not interfering with other components, which is called the dynamic characteristic of cloud computing model.

At the same time, the miniaturization of constituent units also brings the miniaturization of fault units. By designing the corresponding strategy, when a child middleware fails, the system can look for the same functional unit to replace, without affecting the middleware level of service.

At the same time, you can also set the trust of the shift or backup strategy, these sound mechanisms are called cloud computing model autonomy.

Therefore, we say that the SOA system with cloud computing model is a system with dynamic autonomous capability.

2.4 Amazon Elastic Computing Cloud

According to the design idea of cloud computing model, many excellent network autonomy systems are designed and promoted the improvement and development of cloud computing model to some extent. Amazon Elastic Compute Cloud (Amzon EC2), or Amazon elastic Computing cloud, is the pioneer of cloud computing model application [8,10]. Amzon EC2 provides the ability to dynamically resize cloud computing, providing users with a network virtual computing environment through Web Service. It is designed to make network computing easier for developers. For example, the Kettle cluster is based on Amazon elastic Computing Cloud did a test, a single server output 4000 rows/sec, the database is MySQL. Got a good cluster effect [9].

Its design innovation is to allow users to quickly build a virtual machine AMI1 in the data center provided by Amazon, while installing any operating system in this virtual machine, and can use this machine to provide a variety of business. This service allows users to freely control the size of their applications in accordance with their own needs, this avoids the need to purchase, install, maintain additional servers as a result of business changes, as well as the resulting compatibility issues and the additional workload required for system tuning, while also bringing the system the scale advantage and relative stability of distributed network computing. Amazon EC2 is the second SOA system developed using cloud computing model after Amazon simple Storage Service,amazon S3 uses cloud computing models to provide storage services. It allows calculations to be performed in a cloud computing model. Amazon's simple Web services make it easy to have the ability to acquire and configure, allowing you to run calculations in Amazon's computing environment with full control of the computing resources. Amazon EC2 reduces the time it takes to get and configure a new server instance, and can expand or shrink the size of the server instance at any time, depending on the computational requirements.

2.4.1 Amazon EC2 Features

Amazon EC2 describes a real virtual computing environment in this environment, you can use its Network service interface to apply for the virtual machine you need, then you can download your application in this virtual machine, configure your network access License, and use one or more components to provide services according to your needs.
Using Amazon EC2, you can [8]:

can create an AMI that contains the required applications, class libraries, data, and related configuration information. Users can quickly create multiple identical instances using the AMI. Or use preconfigured template mirroring to build the system and run it quickly.

The 1 AMI is an important concept in Amazon EC2, and Amazon EC2 allows users to install and configure instances according to their needs. An AMI (Amazon Machine Image) is a simple packaging environment that contains the information needed to install and configure the city. An AMI is a unit of User Service deployment where a user may need only one AMI or multiple Ami to form a system (such as a network server, application server, and database). Amazon EC2 provides command-line tools that make it easier to create an AMI. If the user builds an AMI, the user needs to upload it to Amazon S3. Amazon EC2 uses Amazon S3 reliable storage users to upload AMI. Or, if the user needs it, the user does not have to start building his AMI from scratch. He can select an appropriate instance from the available Ami.

For example, if a user simply needs a simple Linux server, you can choose a standard Linux AMI.

to upload the Ami to Amazon S3. Amazon EC2 provides a range of tools for storing AMI. Amazon S3 provides a secure, reliable, and fast "warehouse" for storing mirrors. Use the Amazon EC2 Network Service to configure security options and access licenses. Depending on your needs, you can use the Amazon EC2 Network Service to start, terminate, and monitor instances that use your AMI configuration. Pay based on real time and bandwidth.

2.4.2 Amazon EC2 Cloud computing model and its characteristics

As can be seen from the above description, Amazon EC2 The design elements of the cloud computing model: The Amazon EC2 System consists of distributed computing units, which are combined into a complete virtual machine to serve users according to specific user requirements. The virtual host Ami is displayed on the client side. At the same time, users can also according to their own needs, dynamic change the required service capacity rationing. It can be said that the Amazon EC2 has a complete cloud computing model features:

Unified External interface: Virtual host AMI. It has nothing to do with the underlying implementation: the specific implementation of the system is not transparent to the user, and it is similar to using the Amazon EC2 system and using a common virtual machine from the user experience. Modular, collaborative sub-middleware group: Distributed computing units dynamically increase and decrease the ability of Web services: Users can also dynamically change the required service capacity according to their own needs

2.4.3 Amazon EC2 Advantage

Amazon EC2 has the following advantages because of its EC2 cloud computing model:

Elastic--amazon EC2 allows you to add or decrease functionality within a few minutes. You can also delegate one, hundreds, or even thousands of server instances at the same time. Of course, your application in the Network Service application interface under the control, according to the requirements, can automatically expand or reduce the size. Complete control-You can completely control your instance ami. You can use any machine and you can interact with any machine when you need it. Each instance can provide 1.7ghz*86 processor, 1.75GB RAM,160GB Local disk and 250mb/s network bandwidth [8]. Reliable ――amazon EC2 provides a highly reliable environment in which users can efficiently and reliably replace existing instances. This substitution does not involve the underlying artifacts and data centers of the Amazon network. Security ――amazon EC2 provides a network service interface to control network security. You can define a set of instances and their access level. Economic ――amazon EC2 has brought you great economic benefits. You only need to pay very low fees according to the computing function you actually use. This approach is much cheaper than the traditional way of buying and maintaining hardware, both internal and host, [8]. Such systems are designed to liberate users from a number of complex functional plans, transforming their already large fixed costs into relatively small and proportional floating costs, and eliminating the need to buy "extra secure network devices" to handle cyclical network intrusion.

3. Looking ahead

We are now entering the era of computing clouds. The computational cloud provides a dynamic system for the development of SOA systems by component middleware, so that complex problems can be solved easily and quickly. It also enhances the autonomy of the system. The success of Amazon EC2 is a good proof, while promoting the virtualization of computing environment as a trend of many small and medium enterprises and enterprise management software development [8,10].

Of course, computing clouds are not perfect, and there are a lot of problems to be solved, such as a network service that cannot be used offline [11]. However, the computing cloud provides us with a correct path: A dynamic composition system, a simple way of handling, and a greater degree of resource sharing. Cloud computing model has attracted the attention of many important industry leading enterprises, which is considered as one of the important development directions of computer technology in the 2007 [12]. It can be said that the successful use of cloud computing model will make the network application into a new chapter.