CLIENT/SERVER TECHNOLOGY:C / S SYSTEM ARCHITECTURES

3. C / S SYSTEM ARCHITECTURES

Functional Elements of C / S Systems

Since vendors began releasing RDBMS (relational database management system) products in the 1970s, the processing model in which various business data are divided into distributed databases and are accessed via network has been widely adopted. Client / server systems are composed of the various functional elements associated with data processing.

The logical functions of C / S systems are roughly divided into three layers:

1. Presentation logic: This function contains all the logic needed to manage screen formats, the content of windows, and interaction with the user, that is, the user interface. In recent years, the use of graphical user interface (GUI) features such as buttons has become general.

2. Application logic: This is a generic name for application functions that do not belong to other layers. It includes business logic and the flow of control among application components.

3. Data logic: This contains all the logic relating to the storage and retrieval of data, and enforcing business rules about data consistency. Generally, databases are treated and are maintained by a DBMS (database management system).

In addition to these application-functional layers, some functions are needed to support interac- tions between clients and servers via networks. All the distributed software that supports interaction between application components and network software is called middleware. Middleware is a generic term for all the software components that allow us to connect separate layers or components and put them into a complete distributed system. It provides an application programming interface (API) that isolates application codes from the underlying network communication formats and protocols. It also supplies intermediate system services such as security, naming, directory, messaging, and transaction management services:

1. Naming service: In C / S systems, names of clients and servers must be unique within the range in which they are used. A naming service uniquely names a physical entity such as a client or server and associates logical structure such as a tree with physical entities.

2. Directory service: This is a directory service that provides a way for clients to locate servers and their services on the network and controls the address of messages and processing requests through a tree structure.

3. Security service: This service provides authentication, authorization, access control, and user account management.

4. Messaging service: This service provides the ability to send and receive messages between and among applications and users.

5. Transaction service: This service manages distributed transaction processing, such as consis- tency control, that maintains a database in a consistent state and enables recovery control from failures.

Two-Tier Architecture

Client / server architectures can be characterized by how the applications are distributed between the client and the server. In a two-tier architecture, the application logic is placed on either the client or the server undivided or split into two parts, which are placed on the client and the server respectively. In most cases, the presentation logic and the application logic are placed on the client side and the data logic is placed on the server side. Examples of two-tier C / S systems are file servers and database servers with stored procedures. Figure 5 shows a two-tier architecture.

The advantage of the two-tier architecture is ease of development. Because the two-tier architec- ture is simple, a developer can create applications quickly using a 4GL (fourth-generation language) development environment such as Microsoft’s Visual Basic or Inprise’s Delphi. However, as C / S systems grew up to run mission-critical or enterprise-wide applications, shortcomings of the two-tier architecture emerged:

1. Performance: Performance may deteriorate markedly as the number of clients, the size of the database, or the volume of transferred data increases. This deterioration is caused by a lack of capacity of resources such as the processing unit, the memory area, and the network bandwidth

when the processing load is concentrated on the database server. Placing a part of application logic on the database server allows network traffic between the client and the database server to be reduced and the performance even improved. For example, a set of database access commands is compiled and saved in the database of the server. This scheme is called stored procedure. Some DBMS packages support this scheme. But there remains the problem that the application maintenance becomes complicated.

2. Maintenance: Because the business logic is placed on the client side in most two-tier C / S systems, version control of applications becomes complex. For example, in the case of an online shopping system, the business logic (or rule), such as tax rate, is implemented on the client side. Although this structure is effective in reducing network traffic between the client and the server, a system manager has to replace the application of all the clients every time the business rule or version of application is changed. This becomes a big problem in the total cost of ownership of two-tier C / S systems: costs of maintaining, operating, and managing large-scale systems accommodating many users and services.

Three-Tier Architecture

To solve some problems of the two-tier architecture and improve the reliability and performance of the system, a three-tier architecture has been adopted. In a three-tier architecture, the business logic is separated from the presentation and data layers and becomes the middle layer between them. The presentation logic resides in the first tier (the client side), and the data logic in the third tier (the server side), and the business logic in the second tier (the middle) between both tiers. This solves the problems occurring with the two-tier architecture. Figure 6 shows a three-tier architecture where the business logic and some connection functions are placed in the second tier.

In a three-tier architecture, the second tier plays the most important role. Many functions such as data access and connection with other systems are located in the second tier and can be used by any client. That is, the second tier becomes the gateway to other systems. The second tier is often called the application server.

The advantages of three-tier architecture derived from the application server are:

1. Reduction of network traffic: Concentrating the function of accessing database and other sys- tems on the application server makes it possible to reduce network traffic between the client and the server. That is, instead of interacting with the database directly, the client calls the business logic on the application server, and then the business logic accesses the database on the database server on behalf of the client. Therefore, only service requests and responses are sent between the client and the server. From the viewpoint of network traffic, comparisons of two-tier architecture and three-tier architecture are shown in Figure 7.

2. Scalability: Adding or removing application servers or database servers allows the system to be scaled incrementally according to the number of clients and volume of requests.

3. Performance: Because workloads can be distributed across application servers and database servers, this architecture can prevent an application server from becoming a bottleneck and can keep the performance of the system in a permissible range.