Methods of data management in computer networks. Classification of computer networks

Today's article opens a new section on the blog, which will be called " networks". This section will cover a wide range of issues related to computer networks. The first articles of the rubric will be devoted to explaining some of the basic concepts that you will encounter when working with the network. And today we will talk about what components will be required to create a network and which ones exist. types of networks.

Computer network is a set of computer and network equipment connected via communication channels into a single system. To create a computer network, we need the following components:

• computers that have networking capabilities (for example, a network card that is in every modern PC);
• transmission medium or communication channels (cable, satellite, telephone, fiber-optic and radio channels);
• network equipment (for example, a switch or router);
• network software (usually included with operating system or supplied with the network equipment).

Computer networks are usually divided into two main types: global and local.

Local networks(Local Area Network - LAN) have a closed infrastructure before reaching Internet service providers. The term "local area network" can describe both a small office network and the network of a large factory covering several hectares. In relation to organizations, enterprises, firms, the term is used corporate network - a local network of a separate organization (legal entity), regardless of the territory it occupies.
Corporate networks are networks of a closed type, access to them is allowed only to a limited circle of users (for example, company employees). Global networks are focused on serving any users.

Global network(Wide Area Network - WAN) spans large geographic regions and consists of many local area networks. Everyone is familiar with the global network, which consists of several thousand networks and computers - this is the Internet.

The system administrator has to deal with local (corporate) networks. A typical user computer connected to a local network is called workstation . A computer that shares its resources with other computers on a network is called server ; and the computer accessing the shared resources on the server is client .

There are various types of servers: file (for storing shared files), database servers, application servers (providing remote operation of programs on clients), web servers (for storing web content) and others.

Network load is characterized by a parameter called traffic. Traffic is the flow of messages in a data network. It is understood as a quantitative measurement of the number of data blocks passing through the network and their length, expressed in bits per second. For example, the data transfer rate in modern local networks can be 100Mbps or 1Gbps

Currently, the world has a huge amount of all kinds of network and computer equipment that allows you to organize a variety of computer networks. The whole variety of computer networks can be divided into several types according to various criteria:

By territory:

• local - cover small areas and are located inside individual offices, banks, corporations, houses;
• regional - are formed by combining local networks in separate territories;
• global (Internet).

By way of connecting computers:

• wired (computers are connected via cable);
• wireless (computers exchange information via radio waves, for example, using WI-FI or Bluetooth technology).

Control method:

• with centralized management - one or more machines (servers) are allocated to manage the process of data exchange in the network;
• decentralized networks - do not contain dedicated servers, network management functions are transferred in turn from one computer to another.

According to the composition of computing facilities:

• homogeneous - combine homogeneous computing tools (computers);
• heterogeneous - combine various computing tools (for example: PCs, trading terminals, webcams and network storage).

By type of transmission medium networks are divided into fiber-optic, with the transmission of information via radio channels, in the infrared range, via a satellite channel, etc.

You may come across other classifications of computer networks. As a rule, the system administrator has to deal with local wired networks with centralized or decentralized control.

The combination of the above components into a network can be done in various ways and means. According to the composition of their components, the methods of their connection, the scope of use and other features, networks can be divided into classes in such a way that the belonging of the described network to one or another class could quite fully characterize the properties and qualitative parameters of the network.

However, this kind of classification of networks is rather conditional. The most widespread today is the division of computer networks on the basis of territorial location.

On this basis, networks are divided into three main classes:

LAN - local networks (Local Area Networks);

MAN - city networks (Metropolitan Area Networks).

WAN - global networks (Wide Area Networks);

The local network(LS) is a communication system that supports, within a building or some other limited area, one or more high-speed digital information transmission channels provided to connected devices for short-term exclusive use. The territories covered by the LA can vary significantly.

The length of communication lines for some networks can be no more than 1000 m, while other LANs are able to serve the whole city. Served territories can be both factories, ships, aircraft, and institutions, universities, colleges. Coaxial cables are typically used as the transmission medium, although twisted pair and fiber optic networks are becoming more common, and wireless LAN technology is also rapidly developing in recent years, using one of three types of radiation: broadband radio signals, low power radiation ultrahigh frequencies (microwave radiation) and infrared rays.

Small distances between network nodes, the transmission medium used and the associated low probability of errors in the transmitted data make it possible to maintain high exchange rates - from 1 Mbps to 100 Mbps /from).

City networks, as a rule, cover a group of buildings and are implemented on fiber optic or broadband cables. According to their characteristics, they are intermediate between local and global networks. Recently, in connection with the laying of high-speed and reliable fiber-optic cables in urban and intercity areas, and new promising network protocols, for example, ATM (Asynchronous Transfer Mode - asynchronous transfer mode), which in the future can be used both in local and wide area networks.

global networks, unlike local ones, as a rule, cover much larger territories and even most regions of the globe (the Internet can serve as an example). Currently, analogue or digital wired channels, as well as satellite communication channels (usually for communication between continents) are used as a transmission medium in global networks. Transmission rate restrictions (up to 28.8 Kbps on analog channels and up to 64 Kbps on user sections of digital channels) and the relatively low reliability of analog channels, which requires the use of error detection and correction tools at the lower levels of the protocols, significantly reduce the exchange rate data in global networks compared to local ones.

There are other classification features of computer networks.

By area of ​​operation networks are divided into:

banking networks,

Networks of scientific institutions,

University networks;

According to the form of operation can be distinguished:

Commercial networks;

free networks,

Corporate networks

Public networks;

By the nature of the implemented functions networks are divided into:

Computational, designed to solve control problems based on computational processing of the initial information;

Informational, designed to obtain reference data at the request of users; mixed, in which computing and information functions are implemented.

By way of management computer networks are divided into:

Networks with decentralized control;

Centralized management;

Mixed management.

In the first case, each computer that is part of the network includes a complete set software tools to coordinate network operations. Networks of this type are complex and quite expensive, since the operating systems of individual computers are developed with a focus on collective access to the common memory field of the network.

In the conditions of mixed networks under centralized control, tasks with the highest priority and, as a rule, are associated with the processing of large amounts of information are being solved.

Compatibility software there are networks:

Homogeneous;

Homogeneous (consisting of software compatible computers)

Heterogeneous or heterogeneous (if the computers included in the network are programmatically incompatible).

Local networks

There are two approaches to building local networks and, accordingly, two types: client/server networks and peer-to-peer networks.

Client/server networks

In client/server networks, a dedicated computer (server) is used, on which shared files are concentrated and which provides a print service for many users (Fig. 1).

Rice. one.Client/server networks

Server -a computer connected to a network and providing its users with certain services.

Servers can perform data storage, database management, remote job processing, job printing, and a number of other functions that network users may need. The server is the source of network resources. There can be quite a lot of servers on the network, and each of them can serve its own group of users or manage certain databases.

Work station – Personal Computer A connected to the network through which the user accesses its resources. The network workstation operates in both network and local modes. It is equipped with its own operating system (MSDOS, Windows, etc.) and provides the user with all the necessary tools for solving applied problems. Workstations connected to the server are called clients. Both powerful computers for resource-intensive spreadsheet processing and low-power PCs for simple word processing can be used as clients. In contrast, powerful computers are usually installed as servers. Due to the need to ensure simultaneous processing of requests from a large number of clients and good protection of network data from unauthorized access, the server must run a specialized operating system.

Examples: Novell Net Ware, Windows NT Server, IBM OS/2 Lan Server, Banyan Vines.

Peer-to-peer networks

Dedicated servers are not used in peer-to-peer networks (Fig. 2). At the same time as serving the user, a computer in a peer-to-peer network can take on the functions of a server, executing print jobs and responding to file requests from other workstations on the network. Of course, if a computer does not share its disk space or its printer, then it is only a client in relation to other workstations that act as a server. Windows 95 has built-in capabilities for building a peer-to-peer network. If you need to connect to other peer-to-peer networks, then Windows 95 supports the following networks:

NetWare Lite

Artisoft LANtastic.

Rice. 2.Location of computers in peer-to-peer networks.

Network topology

Under topology refers to the description of the network properties inherent in all its homomorphic transformations, i.e. such changes appearance network, the distances between its elements, their mutual arrangement, at which the ratio of these elements to each other does not change.

The topology of a computer network is largely determined by the way computers are connected to each other. Topology largely determines many important properties of a network, such as reliability (survivability), performance, etc. There are different approaches to classifying network topologies. According to one of them, local area network configurations are divided into two main classes: broadcast And consecutive.

In broadcast configurations, each PC (transceiver physical signals) transmits signals that can be perceived by other PCs. Such configurations include "common bus", "tree", "star with passive center" topologies. A star-with-passive-centre network can be thought of as a kind of "tree" that has a root with a branch to each connected device.

In serial configurations, each physical sublayer transmits information to only one PC. Examples of sequential configurations are: arbitrary (arbitrary connection of computers), hierarchical, "ring", "chain", "star with an intelligent center", "snowflake" and
other.

The most optimal in terms of reliability (the possibility of the network functioning in the event of failure of individual nodes or communication channels) is fully connected network, i.e. a network in which each network node is connected to all other nodes, however, with a large number of nodes, such a network requires a large number of communication channels and is difficult to implement due to technical difficulties and high cost. Therefore, almost all networks are incompletely connected.

Although for a given number of nodes in a non-mesh network, there may be a large number of options for connecting network nodes, in practice, the three most common (basic) LAN topologies are usually used:

1. common bus;

2. ring;

3. star.

Bus topology (Fig. 3), when all network nodes are connected to one open channel, usually called a bus.

Fig 3.Bus topology.

In this case, one of the machines serves as a system server that provides centralized access to shared files and databases, printers, and other computing resources.

networks of this type gained great popularity due to low cost, high flexibility and data transfer speed, ease of network expansion (connection of new subscribers to the network does not affect its main characteristics). The disadvantages of bus topology include the need to use rather complex protocols and the vulnerability to physical damage to the cable.

Ring topology (Fig. 4), when all network nodes are connected to one closed ring channel .

Fig 4.Topology "Ring".

This network structure is characterized by the fact that information can be transmitted along the ring in only one direction and all connected PCs can participate in its reception and transmission. In this case, the recipient subscriber must mark the received information with a special marker, otherwise “lost” data may appear that interfere with the normal operation of the network.

As a daisy chain configuration, the ring is especially vulnerable to failures: the failure of any cable segment leads to the termination of service for all users. LAN designers have put in a lot of effort to deal with this problem. Protection against damage or failure is provided either by closing the ring on the return (backup) path, or by switching to a spare ring. In both cases, the general ring topology is preserved.

Star Topology (Figure 5), when all network nodes are connected to one central node, called the host ( host) or hub ( hub).

Fig 5.Topology "Star".

A configuration can be seen as a further development of a rooted tree structure with a branch to each connected device. A switching device is usually located at the center of the network, ensuring the viability of the system. LANs of this configuration are most often used in automated office management systems that use a central database. Star LANs tend to be less reliable than shared bus or hierarchical networks, but this problem is solved by duplicating the central site hardware. The disadvantages include significant cable consumption (sometimes several times higher than the consumption in LANs of similar capabilities with a common bus or hierarchical ones).

Networks can also be of mixed topology ( hybrid) when separate parts of the network have different topologies. An example is the FDDI local network, in which the main (backbone) nodes are connected to the ring channel, and the remaining nodes are connected to them in a hierarchical topology.

The basis TVS classification the most characteristic functional, informational and structural features are laid down.

According to the degree of territorial dispersal network elements (subscriber systems, communication nodes) distinguish between global (state), regional and local computer networks (WAN, RCS and LAN).

By the nature of the implemented functions networks are divided into computing (the main functions of such networks are information processing), informational (for obtaining reference data at the request of users), information and computing, or mixed, in which computing and information functions are performed in a certain, non-constant ratio.

By way of management TVS are divided into networks with centralized(the network has one or more governing bodies), decentralized(each AS has the means to manage the network) and mixed management, in which, in a certain combination, the principles of centralized and decentralized control are implemented (for example, under centralized control, only tasks with the highest priority associated with the processing of large amounts of information are solved).

On the organization of information transfer networks are divided into networks with information selection and information routing. In networks with a selection of information, built on the basis of a monochannel, the interaction of the AU is carried out by selecting (selection) the data blocks (frames) addressed to them: all the AUs of the network have access to all the frames transmitted in the network, but only the AUs to which they are intended take a copy of the frame. In networks with information routing Multiple routes can be used to transfer frames from a sender to a receiver. Therefore, with the help of communication systems of the network, the problem of choosing the optimal (for example, the shortest time to deliver a frame to the addressee) route is solved.

By type of data transfer organization networks with information routing are divided into networks with circuit (channel) switching, message switching and packet switching. Networks that use mixed data transmission systems are in operation.

According to the topology those. configurations of elements in TVS, networks are divided into two classes: broadcast (Fig. 11.1) and serial (Fig. 11.2). Broadcast configurations and a significant part of the serial configurations (ring, star with an intelligent center, hierarchical) are characteristic of the LAN. For global and regional networks, the most common is arbitrary (mesh topology). The hierarchical configuration and the “star” have also found application.

IN broadcast configurations at any time, only one workstation (subscriber system) can work to transmit a frame. Other PCs in the network can receive this frame, i.e. such configurations are typical for a LAN with information selection. The main types of broadcast configuration are common bus, tree, star with passive center. The main advantages of a LAN with a common bus are the ease of network expansion, the simplicity of the management methods used, the absence of the need for centralized management, and the minimum cable consumption. A tree topology LAN is a more advanced version of a bus topology network. A tree is formed by connecting several buses with active repeaters or passive multipliers (“hubs”), each branch of the tree is a segment. The failure of one segment does not lead to the failure of the others. In a LAN with a star topology, there is a passive connector or an active repeater in the center - quite simple and reliable devices. To protect against violations in the cable, a central relay is used, which turns off failed cable beams.

Rice. 11.1. Broadcast network configurations: but - common bus; b- tree; in - star with passive center

Rice. 11.2. Sequential network configurations: a - arbitrary (mesh); b- hierarchical; in - ring; G - chain; e - a star with an intellectual center; e - Snowflake

In sequential configurations, typical for networks with information routing, data transmission is carried out sequentially from one PC to a neighboring one, and different types of physical transmission medium can be used in different parts of the network.

The requirements for transmitters and receivers are lower than in broadcast configurations. Sequential configurations include: arbitrary (cellular), hierarchical, ring, chain, star with an intelligent center, snowflake. In the LAN, the ring and the star are most widely used, as well as mixed configurations - star-ring, star-bus.

On a LAN with a ring topology, signals travel in only one direction, usually counterclockwise. Each PC has a memory of up to a whole frame. When a frame moves around the ring, each PC receives the frame, analyzes its address field, makes a copy of the frame if it is addressed to this PC, and retransmits the frame. Naturally, all this slows down the data transfer in the ring, and the duration of the delay is determined by the number of PCs. Removing a frame from the ring is usually done by the sending station. In this case, the frame makes a full circle around the ring and returns to the sending station, which perceives it as a receipt - confirmation of the receipt of the frame by the addressee. Removing a frame from the ring can also be carried out by the receiving station, then the frame does not complete a full circle, and the sending station does not receive confirmation receipts.

Ring.structure provides quite wide functionality LAN with high efficiency of using a monochannel, low cost, simplicity of management methods, the possibility of monitoring the performance of a monochannel.

In broadcast and most serial configurations (with the exception of the ring), each cable segment must provide signal transmission in both directions, which is achieved: in half-duplex communication networks - using one cable for alternate transmission in two directions; in duplex networks - using two unidirectional cables; in broadband systems - the use of different carrier frequencies for simultaneous transmission of signals in two directions.

Global and regional networks, like local ones, in principle can be homogeneous (homogeneous), in which software-compatible computers are used, and heterogeneous (heterogeneous), including software-incompatible computers. However, given the length of the hot water supply and RCS and the large number of computers used in them, such networks are more often heterogeneous.

Depending on the scalability of the network, it will depend on how the network will be managed in a given enterprise. There are several ways to manage. Local area networks are divided into two subgroups according to the control method: peer-to-peer and hierarchical (multilevel) networks.

Peer-to-peer networks

In a peer-to-peer network, all computers are equal: there is no hierarchy among computers and there is no dedicated server. Typically, each computer functions as both a client and a server; in other words, there is no single computer responsible for administering the entire network. All users decide for themselves what data on their computer to make publicly available over the network.

Peer-to-peer networks are also called workgroups. Working group this is a small team, so peer-to-peer networks most often have no more than 30 computers. Peer-to-peer networks are relatively simple.

Since each computer is both a client and a server, there is no need for a powerful central server or other components required for more complex networks.

Peer-to-peer networks are usually cheaper than server-based networks, but require more powerful (and more expensive) computers. In a peer-to-peer network, the performance and security requirements for network software are typically lower than in dedicated server networks.

Figure 5. Peer-to-peer network

Hierarchical networks

Hierarchical networks have one or more servers that store information that is shared between different users. In order to increase the reliability of information storage on the server, two disks working in parallel and duplicating each other can be installed,

at the same time, in case of failure of one of them, the other is automatically included in the work. Depending on how the server is used in hierarchical networks, the following types of servers are distinguished:

File server. In this case, the shared files and/or shared programs are located on the server. One example of using a file server is to host MS Office programs on it. In this case, only a small (client) part of these programs is located on the workstations, requiring insignificant resources. Programs that allow this mode of operation are called network installable programs.

Database server. In this case, the database is hosted on the server (for example, Consultant Plus, Guarantor, Bank customer accounts, etc.). The database on the server can be updated from various workstations and/or provide information upon request from the workstation.

Hierarchical network clients can use operating systems: Windows XP, Windows Vista, Windows 7, servers require special server versions of operating systems.

Figure 6. Hierarchical network

Our service center will use a hierarchical network. For our case, this is the most suitable option. So that our network does not turn into an information "garbage dump", and also, in order to increase the reliability of information storage, it is necessary to have several servers. In this case, a file server, an Internet server, and a database server. The server will host MS Office, 1C and other programs, and the workstations will host only a small (client) part of these programs, requiring insignificant resources. It is also necessary for each user to allocate his rights in the local network.

LAN classification

Local networks can be classified according to:

• management level;
• appointment;
• homogeneity;
• administrative relations between computers;
• topology;
• architecture.

According to the level of control, the following LANs are distinguished :

• Workgroup LANs, which consist of multiple PCs running the same operating system. In such a LAN, as a rule, there are several dedicated servers: a file server, a print server;
• LAN of structural subdivisions (departments). These LANs contain several dozen PCs and servers such as: file server, print server, database server;
• LAN of enterprises (firms). These LANs can contain over 100 computers and servers such as file server, print server, database server, mail server, and other servers.

Networks are divided into :

• computer networks intended for settlement work;
• information and computing networks, which are designed both for conducting settlement work and for providing information resources;
• information-advising, which, based on data processing, generate information to support decision-making;
• information and control networks, which are designed to manage objects based on information processing.

The types of computers used are:

• homogeneous networks that contain the same type of computers and system software;
• heterogeneous networks that contain heterogeneous computers and system software.

According to the administrative relations between computers, we can distinguish:

• LAN with centralized management (with dedicated servers);
• LAN without centralized management (decentralized) or peer-to-peer (single-level) networks.

In local networks with centralized management, the server provides interaction between workstations, performs the functions of storing public data, organizes access to this data and transmits data to the client. The client processes the received data and provides the processing results to the user. It should be noted that data processing can also be carried out on the server.

Local networks with centralized management, in which the server is intended only for storing and issuing information to clients upon request, are called networks with a dedicated file server. Systems in which information is processed on the server along with storage are called "client-server" systems.

It should be noted that in server local networks, only server resources are directly accessible to the client. But workstations that are part of a centrally managed LAN can simultaneously organize a peer-to-peer local network with all its capabilities.

The software that controls the operation of a centrally managed LAN consists of two parts:

• network operating system installed on the server;
• software on a workstation, which is a set of programs running under the operating system that is installed on a workstation. At the same time, different operating systems can be installed on different workstations in the same network.

Large hierarchical LANs use UNIX and LINUX as network operating systems, which are more reliable. For medium-sized LANs, the most popular network operating system is Windows 2008 Server.

Depending on how the server is used in hierarchical networks, the following types of servers are distinguished:

• File Server. In this case, the shared files and/or shared programs are located on the server.
• Database server. The server hosts the network database.
• Print server. A sufficiently productive printer is connected to the computer, on which information can be printed from several workstations at once.
• Mail server . The server stores information sent and received as if over a local network.

Advantages:

• higher data processing speed;
• has a reliable system of information protection and secrecy;
• easier to manage than peer-to-peer networks.

Disadvantages:

• the network is more expensive due to a dedicated server;
• less flexible than a peer-to-peer network.

All computers in the local network are connected by communication lines. The geometric arrangement of communication lines relative to network nodes and the physical connection of nodes to the network is called the physical topology. Depending on the topology, networks are distinguished: bus, ring, star, hierarchical and arbitrary structures.

Distinguish between physical and logical topology. The logical and physical network topologies are independent of each other. The physical topology is the geometry of the network, and the logical topology determines the direction of data flows between network nodes and the methods of data transmission.

All existing configurations can be divided into two main classes: broadcast and serial.

In the case of a broadcast LAN configuration, signals transmitted by one physical medium connection device are received by all others. In a broadcast LAN, only one station can operate at any given time. All workstations can communicate directly with any workstation on the network.

To build a broadcast configuration, it is necessary to use relatively powerful receivers and transmitters. Therefore, there is a need to limit the length of cable segments and the number of connections. If limits are exceeded, an analog amplifier or digital repeater is used. In addition, the means of connection to the physical medium are chosen such that they do not cause significant signal attenuation.

The main types of broadcast topologies "bus", "tree" and "star" are shown in the diagrams (Figure 3).

Figure 3 - Types of broadcast topologies:

a) "tire"; b) "tree"; c) "star"

In a daisy-chained LAN configuration, each physical media connection device communicates information to only one device. This reduces the requirements for transmitters and receivers, since all stations actively participate in the transmission.

The main types of sequential topologies: "ring", "chain", "snowflake" and "grid" are shown in the diagram (Figure 4).

Figure 4 - Types of sequential topologies "ring", "chain", "snowflake" and "grid"

Consider the following physical topologies:

• physical "bus" (bus);
• physical "star" (star);
• physical "ring" (ring);

Bus topology

• easy to connect a new PC;
• there is a possibility of centralized management;
• the network is resistant to failures of individual PCs and to disconnections of individual PCs.

Disadvantages of star topology networks:

• failure of the hub affects the operation of the entire network;
• high cable consumption;

Topology "ring"

In a network with a ring topology, all nodes are connected by communication channels into an inextricable ring (not necessarily a circle) through which data is transmitted. The output of one PC is connected to the input of another PC. Starting a movement from one point, the data eventually gets to its beginning. Data in a ring always moves in the same direction.

The receiving workstation recognizes and receives only the message addressed to it. A network with a physical ring topology uses token access, which grants a station the right to use the ring in a specific order. The logical topology of this network is a logical ring.

This network is very easy to create and configure. The main disadvantage of ring topology networks is that damage to the communication line in one place or PC failure leads to the inoperability of the entire network.

As a rule, in its pure form, the “ring” topology is not used because of its unreliability, therefore, various modifications of the ring topology are used in practice.

In general, the IT infrastructure of various enterprises can be distinguished by:

• scale;
• the composition of the components;
• equipment level, etc.

Based on this, certain types of IT infrastructures can be represented in the form of basic configurations, which are shown in Figures 5, 6 and 7.

Figure 5 - Small local network.

Small local network. Usually consists of 1-3 servers, network switches, 5-30 workstations.

Figure 6 - Local network and telephone network with MiniATS.

Local network and telephone network with MiniATS. Includes all components of a "small local area network" with the addition of an internal Mini-PBX for switching telephones within the office

Figure 7 - Local network and digital telephone network at several sites.

Local area network and digital telephone network at several sites, united in a virtual private network. The organization's local network is used for IP telephony. It is possible to combine digital telephone networks organizational units over the Internet using virtual private networks.

The level of training of specialists serving the IT infrastructure of enterprises should be very high, requiring responsibility for the work on which the functioning and security of corporate computer networks will depend.