Application Architecture and Process Design

Application Architecture and Process Design

Introduction

The "information architecture" is an iterative process, transverse, which occurs throughout the entire design of the site and in each of its phases, to ensure that the objectives of its production and development of the interface fulfilled effectively.

The Information Architecture as a discipline does not seek to define a universal design methodology but to articulate a set of techniques to help the development and production of information spaces such as websites.

Discussion

The information system to be designed will be multi-tier system. It will be client-server architecture. This gives you one main advantage - the opportunity to get a graphical user interface without upgrading the server. Compared with the file-server architectures, in this case is sharply reduced traffic (but only if properly designed).

There are two key points of any architecture:

The number of pairs of exchange operations (request, response), especially where used slow and unstable channels.

The client application must identify the name of the target server to which it sends a request, and the search for a physical server is based on the rules of name resolution. These functions take on the server and client components of the DBMS, which provide network services. All name resolution issues are the responsibility of the settings of the software, not the actual client application. In fact, the client application communicates directly with the client software network driver database, but not with the database engine. In this case:

Client and server processes can be on one or on different machines.

Client and server processes can work in different environments (hardware, operating system). The architecture ensures the independence of the client software on the database server software. Independence is achieved through a well-defined communication interface, which, in turn, is carried out by sending messages to a strictly defined format.

The client process does not know which server will handle the result of his inquiry. Funds transfer requests from service to service controlled by a database or other server processes (for example, the request was processed by Web-server, convert it to a request for a transaction monitor, which is applied directly to the database, and the reply to go in reverse - as a result client request is processed sequentially by three servers.)

All these provisions are, in fact, describe the encapsulation, which is a prerequisite for the successful design of this architecture.

An important factor is the network bandwidth, and throughput of the selected communications protocol. Most DBMSs support protocol Named Pipes, TCP / IP, NetBIOS, IPX / SPX. All these protocols were not designed to share a database. Some vendors implement their own communication protocols - for example, APPC, provided by IBM DB2, specifically optimized for the exchange operations specific to the database. The bandwidth of the protocol becomes critical only when the packets are large, such as the transmission of images and sound. Bandwidth Protocol - one of the possible parameters to be optimized to improve the speed of information transfer.

The main task of designing the architecture of the "client-server" - on the client ...