• No se han encontrado resultados

Gerencia no realiza seguimiento de las operaciones

NORMAS ECUATORIANAS DE CONTABILIDAD 19 Principios básicos: 21 Clasificación y

During the 1970s it was realized that expensive computer hardware could be used most cost-efficiently (by the maximum number of people) if it was available remotely, i.e. if one could communicate with the computer from a distant location and gain access to its resources. Inter-system communication became possible, in stages, through the use of modems and UUCP batch transfer and later through real-time wide area networks.

The large mainframe computers which served sometimes hundreds of users were painfully slow for interactive tasks, although they were efficient at batch processing. As hardware became cheaper many institutions moved towards a model of smaller computers coupled to file-servers and printers by a network. This solution was relatively cheap but had problems of its own. At this time the demise of the mainframe was predicted. Today, however, mainframe computers are very much alive for computationally intensive tasks, while the small networked workstation provides access to a world of resources via the Internet.

Dealing with networks is one of the most important aspects of system administration today.

The network is our greatest asset and our greatest threat. In order to be a system administrator it is necessary to understand how and why networks are implemented, using a world-wide protocol: the Internet protocol family. Without getting too heavily bogged down in details which do not concern us at an elementary level, we shall explore these themes throughout the remainder of this material.

3.7.1 Constraints on infrastructure

Different operating systems support different ideas about how network services should be used. We are not always free to use the hardware resources as we would like. Operating system technologies restrict the kind of infrastructures it is possible to build in practice (see figure 2).

87

Figure 2: Some network infrastructure models single out a special server-host, which is used to consolidate network services and resources. Such centralization has many administrative advantages, but it concentrates load and can create a bottleneck.

Unix: Much, if not all, of Unix’s success can be attributed to the astonishing freedom which is granted to its users and administrators. Without a doubt, Unix-like operating systems are the most configurable and adaptable ever created. This has kept Unix at the forefront of new technology but has also created a class of operating systems rather like disorganized piles of treasure in Aladdin’s cave.

Unix-like operating systems are not tied to any specific model for utilizing network resources, though vendors sometimes introduce specific technologies for sharing, which favor a particular kind of model. (This is almost viewed as a treasonable offence and is usually quickly rejected in favor of software which offers greater freedom.) Unix lets us decide how we want the network to look. Any Unix system can perform any function, as server, client or both. A Unix network is fully distributed; there is no requirement about centralization of resources, but central models are commonly used.Unix contains troves of tools for making many hosts work together and share resources, but each host can also be configured as a stand-alone system. Each host either has a fixed IP address, or can be assigned one dynamically at boot-time by a service such as BOOTP or DHCP.

Windows: Windows networks are built around a specific model. There are two types of Windows system with separate software licenses: workstations and servers. Windows can work as a stand-alone system or as a workstation, integrated into a system of network services. Windows revolves around a model in w-hich programs are run on a local workstation, but where network services and resources are kept and run on a centralized server. IP addresses may be fixed or may be assigned automatically by a network service such as BOOTP or DHCP. Several Windows servers can coexist. Each server serves a logical group of hosts, users and services called a domain. Domains are now merged into an Active Directory model that provides a logical directory structure for network services. Client machines subscribe to as many domains as they wish or have permission to join. Windows supports two kinds of organizational groups:

workgroups in which hosts share a simple peer-to-peer network, perhaps with Windows 88

9x machines, and domains which have a central authority through a domain server.

Domains provide a common framework including user-ids (SIDs in Windows language), passwords and user profiles. Domains have a common user-database and a common security policy. Any host which subscribes to a domain inherits the users and the security policy of the domain. Windows domains can be simulated by Unix-like hosts.

Windows 2000 is a reincarnation of Windows NT 4.0. It redresses many of the shortcomings of the NT domain model by moving towards Novell-like directory services as its new model for resource organization. It allows remote services such as terminal login, which was only introduced as an afterthought in NT.

3.7.2 User preference storage

Software packages often allow users to store preferences about the way that software should look and behave. Such data are stored in some kind of information repository. Another issue for networked systems is where software preference data should be stored for users. There are two possibilities here which correspond approximately to the Unix approach and the Windows approach.

Windows: Under earlier Windows, each user was assumed to have his or her own personal workstation which would not normally be used by other users. Current Windows versions allow logins by multiple users. Configuration data or preferences which the user selects are stored locally on the system disk in a location provided by the operating system. Later versions of Windows (NT, 2000, XP etc.) solve these problems by maintaining user profiles which are stored on the domain server in a \profiles subdirectory of the system-root. These data are copied into the local workstation when a user logs on to a domain server.

• Unix/Shared: Under Unix, each user sets up personal preferences in his or her personal dot files which are stored in private user space. More general global preferences are stored in a directory of the administrator’s choice. Traditionally this has been the directory /etc.

The difficulties associated with having a fixed location for the configuration information which lies in the system files are several. In any single-user operating system, one user can overwrite another user’s preferences simply by changing them since the system is not capable of telling the difference between users. This is a fundamental problem which indicates that single-user operating systems are basically unsuited to a shared environment. For a multi-user, networked world, the following points must be considered:

• When the operating system is reinstalled, configuration information can easily be lost or overwritten if they are stored in an operating system directory.

• In a distributed environment, where users might not sit at the same physical workstation day after day, the user’s personal configuration data will not follow him or her from machine to machine.

89

On a Unix system, it is easy to specify the locations of configuration files in software and these can then be kept separate from operating system files, e.g. on a different disk partition so that they are immune to accidental deletion by system reinstallation. The main problem with Unix is the lack of any uniformity of approach. In the future, there might be a semblance of uniformity.