Along with that China has joined in the WTO, the competition of telecom industry is more and more intense, and therefore the demand of informatization construction is much higher. Although, most of the telecom sections has built a number of operation system, such as billing system, network management system, business system, office automation system, and so on. However these systems are facing to application and affairs, and they can only simply classify and summarize the business data. With the plentiful increase of the data, how to better utilize these data become a great task to the administrators. CRM was the new type management strategy which was developed from the west in 1990s, it used the information technology means to reach the management objectives. In short, CRM is a sort of business strategy which chooses and manages the valuable customers and the relationship between them. It requires using the business philosophy and the corporate culture that takes customers as the center to support the effective marketing management and service process. This article presents to take advantage of JavaEE to design the CRM, and to organize the operational data in process, respective independent multi data sources together uniformly. Moreover, it looks for the therein law and relation, which aims to provide efficient and accurate operating strategy for telecom industry. Meanwhile, it can prevent operational risk for telecom industry preferably, and can help to open up new telecom service market, and then to achieve bigger economic benefit.
According to the existed information system and the operation flow, based on the existed informatization facilities, it is necessary to build a networked service system in the view of Intemet/Intranet, and therefore on the one hand, to share and release the company's information through public information service, and on the other hand, to provide a kind of interactive mode for self-service based on the network through Web. Moreover, to take advantage of the information technology to centralized manage the customer resources, and using data mining to link the customers and their related business scope together from the dealt customers' information. And then let the market, marketing, customer service and other sections to share the customer resources. And as a result, let the company to track customers' demand in time, and therefore to supply the products and services, and then to increase the level of fast response to customers and service, and to increase customer's satisfaction and loyalty. Consequently, attract more customers, and to maximize the profit, and therefore to reafize the CRM's function of customer management, marketing management, marketing process management, after-sale service management, market management and so on.
When design the telecom CRM system, this article referred to the layering design model of JavaEE, and therefore separated the data access, business logic and view interface. Client tier is a set of views; it contains the pages of all functions, and can be realized by JSP. The client tier is the key part of the whole system, it not only need to do the interaction between users and system, and as a result to finish the relevant functions, and also it is in charge of providing friendly interface to the users. The Struts is applied in the design of boundary layer to achieve the MVC, and therefore, it can separate the input, dispose and output according to the MVC, and what's more, it can weaken the coupling between business logic interface and data interface. The client tier is mainly used to form the logic and to accept the feedback from users, and further, based on the accepted client's request to generate the relevant response. This tier is consisted of the view of Struts and the controller. View make use of tiles tag base from Struts to create the Html menu, and it separates the implemented logic and the displayed logic, and then to build the connection between view and model (business delegate) through action form bean. Controller, as the central controller, it uses the action servlet and the action to call the business delegate, and to update the status of business delegate, and to control the application program, therefore to separate the request logic and the business logic.
Business logic layer makes up the main body of the system, it represents the functions of the system, and it has good portability and expansibifity. During the design of the business logic layer, the business delegate model in JavaEE has been applied to do the capsulation to business logic. At the same time, it used the Spring structure to uniformly allocate and manage the service locator object. As a result, it on one hand, simplified the exploitation of business object, and one the other hand it higher abstracted the business rule according to the AOP of Spring structure.Therefore, the system can amend the business rule during the actual motion according to the various users. The perfection degree of business logic layer directly decides the movement and efficiency of the whole system.
The persistence layer separates the business logic and details of data access, it contains all the possible operations to data base in the process of realizing the business function, such as insert, update, select and delete, etc. Persistence layer is called by business object, and it will return the result of data base's operation to business object. The specially designed persistence layer simplified the interface of data access, it can not only to increase the flexibility and expansibility of system, and also can raise up the performance of data access throughout specially optimized data base access code. In the design of persistence layer, the DAO was applied to package the process of data access.
Now take the sub-system in telecom CRM system as an example to describe the system realization of diagram 3. The data sheets that used in the sub-system of customer management are: customer contract management table, user trace table, and user remind table, user relation table, user information table, user loss table and user data table.
Data persistence layer encapsulated the data access details, and provide service to business logic layer. It used DAO model to access the data base, and abstracted, encapsulated the data source. And created the persistence interface, user compact DAO and its realization of user compact DAO impl, etc. Spring can provide favorable support to DAO of Hibernate, and can provide Hibernate DAO-support, and it is convenient to call session's operation of Hibernate, that is the save, delete,and update. To achieve the quote of Hibernate template by letting user compact DAO Impl to expand the Hibernate DAO support. Next take the get user compact (String user compact Id) in user compact Dao Impl in Java as an example.
The business layer of system separately defined the interface of user compact business and the interface realization class of user compact business Impl, and the relevant component elements will be defined according to the interface of user compact business. To apply the programming of aiming not at classification but the interface can reduce the degree of coupling among the component elements, and it can greatly increase the test's flexibility during the development process.
It is necessary to state that the container managing the dependence among subassembly inverts the control of the dependence among the assembly. In this way, there is no hard code among these subassemblies; therefore, any subassem- bly can be used furthest. Spring injects the dependence through setter, and when it allocates the context file application context-.xml it also easily allocates the quote among the subassembly, and assigns the dependence.
Due to that the office management has a specific transverse concept; therefore, the AOP of the Spring is appropriate. The declarative office management can help to avoid to repeatedly writing a mass of transaction code. In the system, to define the office management on the user compact business, and therefore the center code only need to pay attention to the business logic, and to defiver the office management to the configuration files.
Networks are ubiquitous. Many practical systems in nature and society can be characterized by the network. Examples include (online) social networks, computer networks, Internet, biological networks, transportation networks, and so on. After the seminal work by Watts and Strogatz  and Barabasi and Albert , complex networks have attracted increasing attention in both industry and research communities in the last decade. The studies of complex network mainly focus on investigating the underlying organizing principles, the function, and the dynamics of the network.
However, networks are very often attacked by malicious users. An important issue in network science is to study the robustness of a network subject to nodes or links errors . Many real-world examples are briefly described below. In airline network, an important question is how the operational ability of the network is affected given that a certain airports are closed. In computer networks, a key problem is to study how the communication capacity changes when some computers in the network are attacked by the hackers. In P2P networks, a crucial issue is to study how the communication ability of the network is affected when a certain number of peers depart from the network.
In practice, a network attacker typically has a small budget of knodes (edges) to attack due to resource limitation. The attacker aims at maximizing some utility functions after attacking k nodes (edges). Many previous studies have been focused on such nodes (edges) attack problems. For example, Albert et al. studied the robustness of a network by considering the diameter of the network after deleting a small fraction of nodes. In, the utility function of the attacker corresponds to the diameter of the network, and the attacker aims to maximize the diameter of the network, which will make the network as less cohesive as possible. Many subsequent studies followed this framework to study how a network is affected by nodes or edges errors. However, most of them focus on deriving analytical solutions on the basis of some specific random graph models. More recently, Schneider et al.  investigated how the size of the giant connected component changes subject to nodes deletions. Clearly, in , the utility function of the attacker is the inverse of the size of the maximal connected component of the network, and the attacker aims to minimize the size of the maximal connected component of the network.
Instead of the diameter and the size of maximal connected component, in this paper, we consider another important metric, i.e., the number of residual edges of the network after attacking. Specifically, in our model, the attacker's utility function is equal to the number of edges that are removed. The attacker aims at maximizing the number of edges that are deleted after attacking a small fixed budget of nodes of the network. Or, equivalently, the attacker wants to minimize the number of residual edges of a network after attacking. Clearly, this problem is equivalent to the maximal vertex coverage (MVC) problem on networks. We thus refer to such type of attack as the MVC attack. Note that the number of residual edges is a very natural and intuitive metric to measure the function and performance of the network. Intuitively, after the removal of nodes, the network with a large number of residual edges implies that the function and performance of the network are not extensively damaged. There are many practical applications that can suffer from such MVC attack. For instance, in computer networks, the hacker may want to attack workstations so as to minimize the number of surviving links in the network. In online social networks, the attacker may want to target users by providing some incentives to persuade them to leave the social network so as to minimize the number of residual social ties of the network. Such incentive-based attacks indeed encounter in real-world online social networks.
For example, a recent news in CNET reports an attack event. The fast-food company Burger King developed a Facebook application, namely Whopper Sacrifice, giving a Facebook user a free coupon for a free hamburger if he/she deletes 10 people from his/her friends list. By statistics, 82,771 Facebook users participated in this campaign, and 233,906 Facebook friendships were deleted. After then, Facebook closed this application due to a large number of link deletions. In addition, on the positive side, the MVC attack could be used to fight terrorism. For example, given a terrorist network, we may want to attack a small number of terrorists so that the number of residual ties in the network is minimal. As a consequence, it is important to investigate the impact of MVC attack in a network.