• No se han encontrado resultados

Calidad de la relación de pareja y bienestar físico

CAPÍTULO 3. RELACIÓN DE PAREJA Y BIENESTAR PERSONAL

3.2. Calidad de la relación de pareja y bienestar físico

On completion of this chapter, you should be able to:

n Define a distributed system

n Analyse the organizational benefits and drawbacks of distributed systems

n Appraise the degree of distribution in a system in terms of processing power and data

n Describe basic techniques used to transmit signals across networks

n Compare and evaluate techniques for creating communications channels across networks

n Describe topologies for local area networks

n Outline the OSI model and cite issues relating to standards in networking

n Discuss EDI communications between organizations and evaluate the benefits of EDI.

Introduction

Major developments over the last 30 years have been achieved in information technology.

It is not uncommon to view this purely as the advent and development of computer systems. However, this ignores the significant impact that improvements and innova-tions in telecommunicainnova-tions have had as an enabling technology for information systems.

This chapter begins with a consideration of the way in which centralized and dis-tributed systems can be seen as alternatives for handling information provision. The impact of a distributed system on the organization, its benefits and its acceptability, are analysed. In order properly to appreciate issues in networks and distributed com-puting it is necessary to have at least a basic understanding of the underlying tech-nology. The ‘language of networks’ is explained, various types of public and local area networks are examined, and the way that issues concerning standards bedevil the full integration of systems is covered. The Internet and the World Wide Web are import-ant enough to merit chapters on their own (Chapters 5 and 6, respectively), although concepts key to their understanding are covered in this chapter. Finally, the impact of electronic data interchange and its effect on the competitive position of the firm in the marketplace is assessed.

4.1 Networks and distributed systems

In the last two decades, many organizations have adopted the policy of installing sev-eral geographically distinct computers within their organizations and linking these using telecommunications. The computers may be desktop computers linked together locally in one site or even one office. Or it might be the linking of minicomputers or mainframe computers across large geographical distances. As well as these traditional network connections, the growth of the Internet has introduced a simple and relatively inexpensive way for users to establish temporary connections between computers. The issues involved in this distribution of computing power and the linking networks are the subject of this section.

It used to be believed that computing benefited from economies of scale. The pre-vious chapter introduced Grosch’s law; this stated that the computational and data-processing power of a computer increases with the square of its cost. It therefore made financial sense for an organization to centralize its computer systems in order to get the most power for its money. Under centralization, an organization that is located on several geographically distant sites would then incur a large communications cost.

Terminals at each site needed to interchange data constantly with the centralized central processing unit.

With the development of much cheaper computing hardware and, in particular, the development of the microchip, Grosch’s law has broken down. There are no longer the same economies of scale to be gained by centralization. Local computers can carry out local processing needs, and the need to communicate between different sites in an organization is reduced to those occasions where data held at one location is needed at another. This is called distributed computing.

An example of a distributed system is shown in Figure 4.1. A tyre and car battery manufacturer purchases materials and produces goods for sale throughout the coun-try. The headquarters, factory and a warehouse are located at one site. In order to cut Figure 4.1 An example of functions in a hierarchical distributed system

distribution costs and satisfy retail orders quickly, the organization maintains two other warehouses in different parts of the country to which the manufactured goods are distributed for storage prior to sale. The headquarters’ mainframe computer takes care of centralized accounting, purchasing, production scheduling, wages and salaries, local stock control and local sales order processing. Each of the two warehouses has a small minicomputer to handle its own local stock control and local sales order pro-cessing. These two minicomputers are connected to the mainframe computer so that an enquiry can be made to the other warehouses for products not held in the local warehouse that are needed for local retail outlets.

Most of the stock control enquiries and updates will therefore be on the locally held data stores. On the occasions when the local warehouse cannot satisfy a customer demand, the data held at the other warehouses is interrogated via the telecommunications links.

As the accounting is carried out centrally, although the sales order processing is local, it is necessary to ensure that sales order and delivery details are exchanged between the local computers and the mainframe. As this is not required immediately on a sale then the data can be transferred at the end of each day in one operation. Although accounting, wages and salaries are handled centrally in this organization, an organiza-tion with a different structure might grant greater independence to its branches. These functions would then be the responsibility of each site, and headquarters would receive consolidated accounting reports.

Compare this with a centralized system, as shown in Figure 4.2. Here all the func-tions are carried out centrally at headquarters. Each time there is a need to access the data store or carry out any processing the interaction between the local sites and head-quarters will involve a telecommunications link – even though the processing of data only concerns stock held at the local site. This involves a heavy telecommunications cost. Moreover, unless the links involve high-speed connections the response times in the interaction will be slow. At the headquarters, the mainframe will need to be able Figure 4.2 An example of functions in a centralized system

4.2

to accept transactions from many sites and will need to give over some of its process-ing time to the maintenance and servicprocess-ing of queues. This problem will be larger the greater the number of sites and the greater the traffic. In this scenario, it is unlikely that computer personnel will reside at each of the sites. It would be more common to have a centralized team at the headquarters responsible for applications development and the day-to-day running of computer operations. It is easy for users at the local sites to feel isolated – particularly if help is required or difficulties are encountered with the operation of the system. As can be seen from the two treatments of essentially the same set of functions, a distributed approach has much to commend it.

However, it would be simplistic to suggest that there were only two possible approaches – distributed or centralized. In the above case there is a hybrid. In the

‘distributed’ example certain functions are in fact centralized: the distribution of the stock control system, particularly that component dealing with the update of stock data relating to another warehouse held at another site, involves considerable technical com-plexity as the database itself is distributed. A variation on this is to hold copies cen-trally of the stock data on each of the sites. Downloading to each site of all the stock data on all of the sites occurs early in the morning. Local processing of data on stock held locally occurs during the day. However, information on stocks at other warehouses is obtained by interrogating the early morning copies received locally. These may be out of date – but only by a maximum of 24 hours. Requests for stock from other sites together with the end-of-day copy of the local stock data are transferred to the cen-tral mainframe at the end of the day. The cencen-tral mainframe carries out overnight processing and produces up-to-date stock data for each site, which is downloaded the following morning. This escapes the complexity of requiring a truly distributed database at the expense of forfeiting the immediate update of all stock transactions.

It should be clear from the above that the simple idea of distributed versus central-ized does not apply. Rather the question that is addressed nowadays is to what extent and how should the organization decentralize its functions and data?