Title: Performance Analysis Model of Cloud Center with Batch Arrival
First, the background of the topic
Cloud computing is a network-based computing model. Users apply for computing resources, such as operating system, running environment or software package, from providers through the network. In fact, when allocating resources, users are not clear about the real running environment and the specific details of allocation. In other words, the cloud is an abstraction layer between users and the computing environment. In 1969, L.Kleinrock once said that computer network is still in its infancy, but with its growth, we will see the same "computing service" as power system and telephone system, and it will be fully applied in personal homes and offices. This view based on "computing service" predicts the great changes of the whole computing industry in the 2 1 century. Cloud is a computing service model, which always provides services on demand like other infrastructure services. Cloud computing has become the fifth largest public infrastructure after relay, water, gas and telephone. At present, customers no longer need to invest too much energy and financial resources in building and maintaining large and complex IT infrastructure. Instead, they only need to pay for the computing services they use. The service model of cloud computing can be divided into three layers: Device as a Service (laaS), in which devices refer to hard disks, memory, servers and network devices, which can be accessed through the network; Platform as a Service (PaaS) includes some computing platforms, such as hardware with operating system and virtual server. Software as a Service (SaaS) includes software applications and other corresponding service applications. The definition of cloud computing is not unique, among which H.Khazad put forward it in 20 10. "Cloud computing is a new computing field. Resources such as physical devices, hardware platforms and application software provide users with on-demand services through network services." [2] This definition expounds several important features of cloud computing.
(1) large-scale infrastructure. The cloud computing platform based on very large-scale hardware devices has super computing power. Well-known enterprises in the world, such as roM, Amazon, Microsoft, etc., all have cloud service platforms with hundreds of thousands of servers, and the number of servers in Google cloud computing platform is more than one million. Even an ordinary private cloud will generally purchase hundreds or even thousands of servers.
(2) Based on virtualization technology. The resources that users get from the cloud computing platform are all virtualized. From the running end, users can just host the application in the cloud without knowing the specific location where the application runs. As far as the terminal is concerned, the user can get the required service through the terminal equipment at any place. In short, users will always face the interface of a cloud platform, not a tangible and fixed entity.
(3) High reliability. Cloud computing adopts data multi-copy fault-tolerant technology and isomorphic exchange strategy of computing nodes to ensure the reliability of cloud center. The reliability of cloud computing is unmatched by local computing.
(4) universality. Cloud computing does not specifically provide services for any specific application. In fact, a user can create multiple different applications according to his own needs in the cloud computing platform, and a cloud computing platform can also run different applications of multiple different users.
(5) Easy to expand. The scale of cloud computing platform can be contracted and expanded according to the actual needs to meet the changes in the scale of platform requests and the number of users.
(6) On-demand service. The usage fee that users should pay is calculated according to how many computing resources are used. Pay more if you use more, pay less if you use less, and pay nothing if you don't use it. This can completely reduce the idleness of user resources in leisure time.
(7) Low cost. By adopting fault-tolerant technology, large-scale and cheap server clusters can be used as hardware infrastructure to build a cloud computing platform, which greatly reduces the cost investment for cloud computing service providers. For users, it is very cost-effective to exchange a small amount of rent for computing resources that originally needed high investment without considering the maintenance cost of software and hardware.
Second, the purpose and significance of the study
The existing physical model of cloud center is usually task-oriented, while the performance evaluation and index change of service model for batch tasks are in the preliminary research stage, and there is no mature model. Therefore, this paper uses ikT/G/m/w+t queuing system to describe the cloud center facing batch tasks, and uses embedded Markov chain method to model the+queuing system, thus achieving accurate modeling and analysis of the cloud center.
Third, the main theories involved in this study.
Queuing is a common social phenomenon in daily life. Waiting for the bus needs to line up, going to the hospital to see a doctor needs to line up, going to the cafeteria to cook needs to line up and so on. The emergence of queuing phenomenon needs to have two aspects at the same time. The individuals in the queue need to get services and have service providers. The so-called queuing theory is to imitate this queuing phenomenon, first abstract it into a physical model, and then further establish a theoretical system of mathematical model. Obviously, queuing theory studies the various states of a system when it provides a certain service to users. In queuing theory, people or things that need services are usually called customers, while people or institutions that provide services are called service desks. Customers and service desks form a queuing system. Although there are many kinds of queuing systems, judging from the main factors that determine the progress of queuing systems, they are mainly composed of customer arrival, queuing process and service process.
(1) Customer Arrival: The customer arrival process describes the law of customer arrival. Customers usually arrive one by one, and others arrive in batches, which is also called collective arrival. Customers may arrive one after another or in batches, and the interval between customers' arrival is not unique. But there will always be a certain arrival mode. This arrival law refers to the distribution of arrival process or arrival time. The main contents of the customer arrival process research include the probability distribution of the arrival time intervals of adjacent customers, the parameters of the probability distribution and whether the arrival time intervals are independent of each other.
(2) Queuing process: In the process of queuing, we mainly discuss two issues, one is the queuing length, and the other is the queuing rules. Queue length can be divided into finite and infinite. The length of the queue is different, and the difficulty and conclusion of discussing the problem are also different. In many cases, the queue length capacity is set to infinity to deal with this problem. Queuing rules include two parts: queue form and waiting system. Queue forms include single queue, parallel multi-queue, serial multi-queue and chaotic queue. Parallel multi-queue is to allow a queue to be formed in front of each window of multiple windows. Arriving customers can choose a queue to queue at the beginning according to the length of the queue. As the name implies, a serial queue is a series of queues. After customers receive services in one queue, they will go to the next queue to receive services. Random queues mean that serial and parallel queues will be randomly distributed.
The main purpose of queuing model simulation is to find the best configuration between service settings and service objects, so that the system has the most reasonable configuration and the best service efficiency. Markov process is the main method to study queuing system. Markov process is a special random process with no aftereffect, and its state space is finite or countable infinite. The process of this system jumping from one state to another depends only on the current departure state, and has nothing to do with the previous historical state. Markov chain is widely used as an important tool to study queuing system. However, not all queuing systems can be abstracted into Markov processes in a strict sense. Therefore, with the development of queuing process, Markov chains have many expansion models and regeneration methods, which make Markov chains more widely used, such as embedding Markov chains, supplementary variable method, quasi-birth and death process and so on. Firstly, this section introduces Markov chains in the strictest sense. Markov chains can be divided into two categories according to time, discrete-time Markov chains and continuous-time processes.
Fourthly, the main content of this paper.
From the standpoint of the government, this paper focuses on how to successfully apply REITs to public rental housing construction financing, and takes the operation of REITs in public rental housing construction financing as the main research object. In addition to the introduction and conclusion, the main content of this paper focuses on the second chapter to the fifth chapter, which is divided into four parts: The first part studies the application experience of REITs at home and abroad and the successful experience of combining with affordable housing. Foreign countries mainly investigate typical countries and regions in the United States and Asia, including Japan, Singapore and Hong Kong. Due to the lack of domestic experience, we mainly investigate two typical cases: CITIC-Kaide Science Park Investment Fund and Huixian Industrial Trust. The second part deeply studies the necessity and feasibility of developing REITs in public rental housing in China. The necessity analysis points out that REITs is an important way to expand the financing channels and improve the management efficiency of public rental housing construction. The feasibility is analyzed in detail from two aspects of economic and financial environment and laws and regulations. In the third part, aiming at the current situation of public rental housing management in China, the operation status of REITs in public rental housing construction financing is expounded in detail, including the basic mode and operation process of REITs, and the power and responsibility relationship and interest distribution of REITs participating parties are further studied, so as to put forward the agent selection mechanism and incentive mechanism. It is worth pointing out that at this time, the government is not only the payer of REITs subsidies, but also the representative of REITs investors. In different stages of REITs operation, the government participates in the internal game of REITs in different capacities. The fourth part, from the perspective of the government as a regulator, puts forward a series of policy suggestions on the application of REITs in the financing of public rental housing construction in China, including that the government should improve the laws and regulations related to REITs and public rental housing and establish a sound REITs supervision system.
Five, writing outline
Acknowledgement 5-6
Chinese abstract 6-7
Abstract 7
Chapter 1 Introduction 10- 17
1. 1 research background and significance10-1/
1. 1 research background10-1/
1. 1.2 research significance 1 1
1.2 research status 1 1- 15
1.2. 1 foreign research status 1 1- 12
1.2.2 domestic research status 12- 15
1.3 content and structure of the paper 15- 17
1.3. 1 main content of the paper 15
1.3.2 paper structure 15- 17
Chapter II Application Experience and Enlightenment of REITs at Home and Abroad 17-35
2. 1 American REITs 17-26 Application Experience
2. 1. 1 American REITs and their application in low-rent housing construction 17-2 1
2. 1.2 Operation mode of REITs in the United States 2 1-26
2.2 REITs application experience in typical European countries and regions 26-29
2.2. 1 Japanese REITs operating mode 26-27
2.2.2 Operation Mode of Singapore Real Estate Investment Trust Fund 27-28
2.2.3 Operation Mode of Hong Kong Real Estate Investment Trust Fund 28-29
2.3 REITs application experience in China 29-32
2.3. 1 CITIC-Kaide Science Park Investment Fund 29-30
2.3.2 Huixian Industrial Trust 30-32
2.4 REITs at home and abroad experience comparison and enlightenment 32-35
2.4. 1 Comparison of REITs at home and abroad 32-33
2. 4. 2 China public rental housing construction financing experience inspiration 33-35
Chapter III Analysis on the Necessity and Feasibility of Financing for Public Rental Housing Construction in China 35-43
3. 1REITs Analysis on the Necessity of Financing in the Construction of Public Rental Housing 35-37
3. 1. 1REITs is an important way to expand the financing channels for public rental housing construction 35-36
3. 1.2REITs is an important way to improve the efficiency of public rental housing construction management 36-37
3.2 Feasibility Analysis of Real Estate Investment Trust Fund Applied to Public Rental Housing Construction Financing 37-43
3.2. 1 The economic and financial environment is relaxed and private capital is abundant 37-4 1
3.2.2 Guided by laws and regulations, policies will continue to be favorable 4 1-43
The fourth chapter is the operation of REITs in the financing of public rental housing construction in China, 43-64 pages
4. 1 REITs financing mode of public rental housing construction in China 43-47
4. 1. 1 design principle
4. 1.2 basic form selection 44-45
4. 1.3 Organizational structure construction 45-47
4.2 Defects in the financing operation process of public rental housing construction in China 47-50
4.2. 1 Establishment and issuance stage 47-48
4.2.2 Operation Management Stage 48-49
4.2.3 Termination of liquidation phase 49-50
4.3 Defects of financing mechanism for public rental housing construction in China 50-64
4.3. Agency Problems in the Operation of1REITs 50-52
4.3.2 Agent Selection Mechanism 52-56
4.3.3 Incentive Mechanism for Agents 56-64
Chapter V Policy Suggestions on the Use of REITs in the Financing of Public Rental Housing Construction in China 64-68
5. 1 Improve the system of laws and regulations 64-66
5. 1. 1 Perfecting the legal system of real estate investment trust 64-65
5. 1.2 Perfecting laws and regulations related to public rental housing 65-66
5.2 The establishment of real estate investment trust fund supervision system
5.2. 1 Clarify the subject and responsibility of government supervision 66
5.2.2 Establishment of information disclosure system of real estate investment trust funds
5.2.3 Guide the public to supervise.
Chapter 6 Conclusion and Prospect 68-70
6. 1 main work and conclusion 68
6.2 Problems to be further studied 68-70
References 70-73