Methods for combining LANs. Merging devices

6. Bridges. Routers. Gateways.

Bridges are software and hardware devices that provide connection between several local networks or several parts of the same network working with different protocols. Bridges are designed to logically structure a network or to connect essentially identical networks that have some physical differences. A bridge isolates traffic on one part of the network from traffic on another part, improving overall data transfer performance.

Routers. This is communications equipment that provides a choice of data transmission route between several networks that have different architectures or protocols. Routers are used only for connecting homogeneous networks and in branched networks with several parallel routes. Routers and network operating system software modules implement network layer functions.

Gateways are communications equipment (for example, a computer) that serves to connect heterogeneous networks with different exchange protocols. Gateways completely transform the entire data flow, including codes, formats, control methods, etc.

3.4.1 Reasons for LAN extension and devices used for it

LPs tend to outgrow initial projects. As companies grow, so do LANs. A change in a company's business profile or organization may require network reconfiguration. This becomes obvious when:

Documents wait in line for a network printer for an unacceptably long time;

database query time increased;

information protection requirements have changed, etc.

Networks cannot expand by simple addition workstations and cable routing. Any topology or architecture has its limitations. However, there are devices that can:

segment the LAN so that each segment becomes an independent LAN;

combine two LANs into one;

connect the LAN to other networks to connect them to the Internet.

TO such devices include: repeaters, bridges, routers, bridge routers and gateways.

3.4.2 Repeaters

Connecting a repeater to a LAN

Repeaters transmit all traffic in both directions and operate at the physical layer of the OSI model. This means that each segment must use the same packet formats, protocols, and access methods. That is, with the help of a repeater you can combine single network two Ethernet segments and impossible Ethernet and Token Ring.

However, repeaters make it possible to connect two segments that use different physical signal transmission media (cable - optical, cable - pair, etc.). Some multiport repeaters act as multiport hubs, connecting different types of cables.

The use of repeaters is justified in cases where it is necessary to overcome the limitation on the length of the segment or the number of PCs. Moreover, none of the network segments generates increased traffic, and the cost of the LAN is the main factor. This is due to the fact that repeaters do not perform the functions of isolation and filtration.

Thus, transmitting every bit of data from segment to segment, they will transmit both corrupted packets and packets not intended for this segment. As a result, the problems of one segment will affect others. Those. The use of repeaters does not provide the segment isolation function. In addition, repeaters will distribute all broadcast packets throughout the network. And if a device does not respond to all packets, or packets are constantly trying to reach devices that never respond, then network performance drops, i.e., repeaters do not filter signals.

3.4.3 Bridges

A bridge is a device for integrating a computer network. These devices, like repeaters, can:

increase the size of the network and the number of PCs in it;

connect dissimilar network cables. However, their fundamental difference is that they work on link level OSI models, i.e. at a higher level than repeaters and take into account more features transmitted data, allowing:

restore the shape of signals, but doing it at the packet level;

connect dissimilar network segments (for example, Ethernet and Token Ring) and transfer packets between them;

improve the performance, efficiency, security and reliability of networks (which will be discussed below).

Operating principles of bridges

The operation of the bridge is based on the principle that all network nodes have unique network addresses, and the bridge transmits packets based on the address of the destination node (Fig. below).

Integration example LAN segments using bridges

When controlling access to the network, the bridge:

listens to all traffic;

checks the source and destination addresses of the packet;

builds a routing table;

transmits packets based on the destination host address.

The bridge has some "intelligence" because it learns where to send the data. When packets are sent across a bridge, the transmitter addresses are stored in the bridge's memory and a routing table is created based on them. At the start of work, the table is empty. Then, when nodes transmit packets, their addresses are copied into the table. With this data, the bridge studies the location of computers on network segments. Listening to the traffic of all segments and receiving the packet, the bridge looks for the transmitter address in the routing table. If the source address is not found, it adds it to the table. Then it compares the recipient address with the routing table database.

If the destination address is in the table and the destination is in the same segment as the source, the packet is discarded. This filtering reduces network traffic and isolates network segments.

If the destination address is in the table, but the destination and source are on different segments, the bridge forwards the packet through the appropriate port to the desired segment.

If the address is not in the table, the packet is relayed to all segments, excluding the one from which it was received.

In short, if the bridge knows the location of the destination node, it forwards the packet to it. Otherwise, it broadcasts the packet to all segments.

The considered option corresponds to the simplest, so-called transparent bridges. Currently, bridges with a spanning tree algorithm, bridges with source routing, etc. are used.

Purpose of bridges

1. Bridges allow you to increase the coverage range of your network by working as repeaters. In this case, a cascaded LAN connection via bridges is allowed. Moreover, these LANs can be heterogeneous.

2. The use of bridges improves network performance due to the possibility of network segmentation. Since bridges are able to filter packets according to certain criteria, a large network is divided into several segments connected by bridges. Two small segments will work faster than one large one because traffic is localized within each segment.

3. The use of bridges increases the efficiency of the network, because for each subnet (segment) you can use different topologies and transmission media, and then connect them with bridges. So, for example, if in individual departments the PCs are connected by twisted pairs, then these subnets can be bridged to the corporate LAN with an optical backbone. Since twisted pair cables are cheap, this will save money, and the core backbone (which carries most of the traffic) will use a high-bandwidth environment.

4. Bridges allow you to increase the security (protection) of data due to the fact that they can be programmed to transmit only those packets that contain the addresses of specific senders and recipients. This allows you to limit the circle of PCs capable of sending and receiving information from another subnet. For example, in a network serving accounting, you can install a bridge that will allow only some external stations to receive information.

5. Bridges increase the reliability and resiliency of the network. When network segmentation fails any subnet will not stop all others. In addition, when a single file server goes down, the entire network stops working. If, using internal bridges, you connect two file servers that protect each other, then:

 network fault tolerance will increase;

 traffic levels will decrease.

There are local and remote bridges. Remote bridges are used in large networks when its individual segments are connected by telephone (or other) communication channels.

However, if only one local bridge is used to connect two cable LAN segments, then large networks you have to use two remote bridges connected via synchronous modems to a dedicated communication channel (Fig. below).

Using two remote bridges

3.4.4 Routers

A router is a device for connecting networks using different architectures and protocols. Working for network level OSI models, they can:

switch and route packets across multiple networks;

determine the best way for their transfer;

bypass slow and faulty channels;

filter broadcast messages;

act as a security barrier between networks.

A router, unlike a bridge, has its own address and is used as an intermediate destination.

How the router works

The operation of the router is based on a table stored in its memory. However, this table differs significantly from bridge tables in that it does not contain node addresses, but network addresses (Fig. below). For each protocol used in the network, its own table is built, which includes:

all known network addresses;

methods of communication with other networks;

possible routing paths;

the cost of data transmission along these routes.

Routers, when receiving packets, do not check the address of the destination host, but allocate only the network address. They pass the packet if the network address is known, passing it on to the router that serves the destination network.

Example of a LAN connection using routers

By accepting only addressed network packets, they prevent invalid and broadcast packets from entering the network, thereby reducing the load on the network. The router can “listen” to the network and determine which parts of it are busiest. It sets the number of transits between LANs. Using this information, the router selects a transmission route. If one is overloaded, it will indicate another. Are used various algorithms routing:

based on link state (in IPX);

distance vector (TCP/IP);

Open Shortest Path Preference Protocol (OSPF and TCP/IP), which calculates a route taking into account the number of hops, line speed, traffic, and cost.

Types of routers and their differences from bridges

Just like bridges, routers can be local or remote. Static and dynamic routers are distinguished by type of operation:

static ones require the network administrator to manually create and configure the routing table and specify each route;

dynamic ones automatically determine routes and therefore require minimal setup and configuration. They are more complex and more expensive, because they make a separate decision for each package. The difference between bridges and routers is that:

The bridge operates at the data link level and “sees” only the host address; recognizing it, transmits it to the desired network segment; without determining the address, forwards to all segments;

The router works at the network layer, determining both what needs to be transmitted and where it needs to go; that is, it recognizes not only the address (but already the network!), but also the protocol type; In addition, a router can set the addresses of other routers and decide which packets to forward to which routers.

A bridge can recognize only one path between networks, but a router finds the best one out of many. Currently, bridges - routers - devices have begun to be used that combine the best properties of bridges and routers: for some protocols they act as bridges; for others - like routers.

3.4.5 Gateways

Gateways are devices that enable communication between different architectures and environments. Their main purpose is to communicate between the PC and the environment of minicomputers or mainframes (Fig. below).

Communication between a LAN and a large computer via a gateway

Typically, the role of gateways in a LAN is performed by dedicated servers, and all other workstations on the LAN work with the mainframe as easily as with their own resources. The gateway connects two systems that use different:

communication protocols;

data structures and formats;

languages ​​and architectures.

Gateways accept data from one environment, remove the protocol stack, and repackage it into the protocol stack of the destination system (figure below). When processing data, the gateway performs the following operations:

Gateway operation

1. extracts data from incoming packets, passing them from bottom to top through the full stack of protocols of the transmission medium;

repackages the received data, passing it from top to bottom through the protocol stack of the destination network.

7. Defining host names. HOSTS file. DNS formation service. WINS service. Definition of NetBIOS name.

DNS Domain Name Service Protocols and addresses

DNS (Domain Name System) is a distributed database that maintains a hierarchical naming system for identifying hosts on the Internet. DNS service intended for automatic search IP addresses by known symbolic host name. The DNS specification is defined by RFC 1034 and 1035. DNS requires a static configuration of its tables that map computer names to IP addresses.

The DNS protocol is an application layer service protocol. This protocol is asymmetrical - it defines DNS servers and DNS clients. DNS servers store part of a distributed database of symbolic names and IP addresses. This database is distributed across administrative domains of the Internet. Clients DNS server know the IP address of the DNS server of their administrative domain and send a request via the IP protocol in which they report a well-known symbolic name and ask to return the corresponding IP address.

If the data about the requested match is stored in the database of this DNS server, then it immediately sends a response to the client, but if not, then it sends the request to a DNS server of another domain, which can process the request itself or transfer it to another DNS server. All DNS servers are connected hierarchically, in accordance with the Internet domain hierarchy. The client queries these name servers until it finds the mappings it needs. This process is accelerated because name servers continually cache the information provided by queries. Client computers can use the IP addresses of several DNS servers in their work to increase the reliability of their work.

The DNS database has a tree structure called a domain namespace, in which each domain (node ​​in the tree) has a name and can contain subdomains. The name of a domain identifies its position in this database in relation to the parent domain, with dots in the name separating the parts corresponding to the nodes of the domain.

The root DNS database is managed by the Internet Network Information Center.

Domains top level are assigned for each country, as well as for organizational basis. The names of these domains must follow the international standard ISO 3166. Three-letter and two-letter abbreviations are used to designate countries, and various types organizations use the following abbreviations:

com - commercial organizations (for example, microsoft.com);

edu - educational (for example, mit.edu);

gov - government organizations (for example, nsf.gov);

org - non-profit organizations (for example, fidonet.org);

net - organizations that support networks (for example, nsf.net).

Each DNS domain is administered by a separate organization, which typically splits its domain into subdomains and delegates the administration of those subdomains to other organizations. Each domain has a unique name, and each of the subdomains has a unique name within its domain. The domain name can contain up to 63 characters. Each host on the Internet is uniquely identified by its full domain name(fully qualified domain name, FQDN), which includes the names of all domains in the direction from the host to the root.

Defining the WINS Service

WINS (Windows Internet Name Service) provides a distributed database for registering and querying dynamic NetBIOS name mapping for computers and groups on a network. WINS maps NetBIOS names to IP addresses and was designed to eliminate the difficulties encountered in resolving NetBIOS names in routed environments. WINS service is the most convenient means NetBIOS name resolution on routed networks using NetBIOS

via TCP/IP.

NetBIOS names have been used in previous versions operating rooms Microsoft systems® Windows® to identify and discover computers and other shared or group resources that require registration and name resolution for use on a network.

NetBIOS names are required to establish network services in earlier versions of Microsoft operating systems. Although the NetBIOS naming protocol can be used with protocols other than TCP/IP, WINS was designed specifically to support NetBIOS over TCP/IP (NetBT).

WINS simplifies management of the NetBIOS namespace on TCP/IP-based networks. The following figure shows typical sequences of events associated with WINS clients and servers.

IN In this example, the following events occur.

1. The HOST-A WINS client registers any of its local NetBIOS names on its

WINS server WINS-A.

2. Another WINS client, HOST-B, asks the WINS-A server to find the IP address of the computer HOST-A on the network.

3. The WINS-A server returns 192.168.1.20 - the IP address of the HOST-A computer.

WINS reduces the use of local IP broadcasts for NetBIOS name resolution and allows users to easily find computers on remote networks. Because WINS registrations are performed automatically every time you start the client and log on to the network, the WINS database is automatically updated when changes are made to the dynamic address configuration. For example, when a DHCP server assigns a new or changed IP address to a client computer running

WINS service, the WINS client information is updated. It does not require manual changes by the user or administrator.

Notes

The WINS protocol is based on and compatible with the protocols defined for the NetBIOS Name Service in RFC1001 and RFC1002. Therefore, it interoperates with other implementations of these RFCs.

Replication of NetBIOS name data to WINS is a Microsoft proprietary technology. It is not compatible with other NetBIOS name servers.

WINS service

WINS (Windows Internet Name Service) provides distributed database support for dynamic registration and resolution of NetBIOS names. The WINS service maps the NetBIOS namespace and IP address space to each other and is designed to resolve NetBIOS names on routed networks that use NetBIOS over TCP/IP. It should be recalled that NetBIOS names are used by early versions of Windows operating systems as the main way to name network resources. WINS was designed to simplify the process of managing the NetBIOS namespace on TCP/IP-based networks.

The main purpose of the WINS service is to resolve NetBIOS names to IP addresses. The resolution process is based on a WINS server database containing mappings from the NetBIOS namespace to the IP address space. When a client logs on to the network, it registers its name in the WINS server database. When the client shuts down, it sends a message to the WINS server notifying it that it has released the registered name. In Fig. Figure 13.21 shows the process of interaction between a WINS client and a WINS server.

Rice. 13.21. Interaction between WINS client and WINS server

Implementing the WINS service in Windows Server 2003 is characterized by the functionality listed below.

Permanent connections. Each WINS server can be configured to serve a persistent connection with one or more

Network aggregation devices provide communication between local network segments, individual LANs and subnets of any level. These devices in the most general form can be attributed to certain levels of the reference model of open systems interaction.

There are the following classes of devices for combining LAN segments and networks (see Table 1.1):

repeaters unite networks at the physical level;

Bridges and switches connect networks at the data link layer and use the functionality of the physical layer. Bridges are executed on the basis of a computer equipped with appropriate software. The difference between switches and bridges is that they implement their functions in hardware and therefore have significantly higher performance;

routers connect networks at the network layer and use the functionality of layers 1 and 2;

Gateways, or gateways, connect networks at the application layer and use the functionality of all underlying layers.

1.6. Requirements for the quality of services and performance of computer networks

1.6.1. Criteria for assessing the quality of service

The main requirement is to ensure that all users have access to shared network resources with a given quality of service (QoS - Quality of Service). The main criteria for assessing the quality of service are performance, reliability And safety. The performance indicators used are response time, throughput And transmission delay.

Reaction time is the time interval between the occurrence of a user request to a network service and the receipt of a response. The response time depends on the load on the transmission medium segments and active network equipment (switches, routers, servers).

Bandwidth– this is the amount of data transmitted per unit of time (bit/s, packets/s). The bandwidth of a composite path in a network is determined by the slowest element (typically a router).

Transmission delay– this is the time interval between the moment a packet arrives at the input of a network device and the moment it appears at the output of the device.

Detailed information about quality of service indicators and methods for calculating network parameters are given in Section 8. In this section, we will consider a technique for approximate calculation of network parameters for the interactive mode.

1.6.2. Calculation of network parameters for interactive mode

To analyze the interactive mode of network operation, we use the model of a closed queuing network (CLN). Data from application usage graphs by different types of users and data on the number of users allow us to determine server load graphs in the form of the number of users using the server at each interval.

Let's consider solving the problem for the interactive mode of operation of a local computing system(see Fig. 1.7) with the following parameters:

reaction time dialogue subscriber (thinking time) 1/;

decision time tasks (response time to a request from the terminal) should not exceed T extra for 90% of tasks;

number of usersn.

This model is constantly circulating n applications (transactions).

Need to find:

the value of the queuing network parameters  and , at which t T extra for 90% of interactive requests, i.e. P( t 10 s ) = 0.9;

using the found  and , calculate the system and network characteristics of SeMO;

determine the appropriate type of computing system and its performance indicators that provide the required response time to a request from the terminal.

To solve the problem, an approximate method is used, based on the decomposition of the computing system into a processing subsystem and a terminal subsystem (and their “independent” consideration) with the subsequent balance of flows in these subsystems. Then, to find the unknowns  and  (for exponential flows of demands and service), we can create a system of equations:

, (1.1)

. (1.2)

From the first equation  -  = - ln(1- P) / T extra. Substituting  -  into the equation, we get:

(1.3)

and
, Where – intensity of requests per user. Then you can determine the required performance of the computing system (network and servers) =  - ln(1- P) / T extra, system load factor  =  /, as well as the average request latency on the server.

1 datagram – a packet transmitted through the network independently of other packets

2 host (English host) – a computer permanently connected to the network

3 This interface is known as the serial port. Later, other asynchronous transmission standards emerged. Currently RS-232-C has been replaced modern standard RS-232-D.

4 SDLC – Synchronous Data Link Control, LAPB – Link Access Protocol-Balanced, HDLC – High-Level Data Link Control.

5 The meaning of the term “synchronism” for synchronous and asynchronous lines is different from its meaning for SONET and SDH networks.

This article is dedicated to local network basics, the following topics will be covered here:

• The concept of local network;
• Local network device;
• Equipment for local network;
• Network topology;
• TCP/IP protocols;
• IP addressing.

The concept of a local network

Net - a group of computers connected to each other using special equipment that allows the exchange of information between them. The connection between two computers can be direct ( point-to-point connection) or using additional communication nodes.

There are several types of networks, and a local area network is just one of them. A local area network is essentially a network used within a single building or individual space, such as an apartment, to allow the computers and programs used within them to communicate. Local networks located in different buildings can be connected to each other using satellite communication channels or fiber optic networks, which allows you to create global network, i.e. a network that includes several local networks.

The Internet is another example of a network that has long since become worldwide and pervasive, containing hundreds of thousands of different networks and hundreds of millions of computers. Regardless of how you access the Internet, through a modem, local or global connection, every Internet user is in fact network user. A wide variety of programs are used to surf the Internet, such as Internet browsers, FTP clients, email programs and many others.

A computer that is connected to a network is called workstation (Workstation). As a rule, a person works with this computer. There are also computers on the network that no one works on. They are used as control centers in the network and as information storage devices. Such computers are called servers,
If computers are located relatively close to each other and connected using high-speed network adapters, then such networks are called local networks. When using a local network, computers are usually located within the same room, building, or in several nearby houses.
To connect computers or entire local networks that are located at a considerable distance from each other, modems are used, as well as dedicated or satellite channels communications. Such networks are called global. Typically, the data transfer speed in such networks is much lower than in local ones.

LAN device

There are two types of network architecture: peer-to-peer ( Peer-to-peer) and client/server ( Client/Server), At the moment, the client/server architecture has practically replaced the peer-to-peer architecture.

If a peer-to-peer network is used, then all computers included in it have the same rights. Accordingly, any computer can act as a server that provides access to its resources, or a client that uses the resources of other servers.

In a network built on a client/server architecture, there are several main computers - servers. The remaining computers that are part of the network are called clients, or workstations.

Server - it is a computer that serves other computers on the network. There are various types of servers, differing from each other in the services they provide; database servers, file servers, print servers, mail servers, web servers, etc.

Peer-to-peer architecture has become widespread in small offices or home local networks. In most cases, to create such a network, you will need a couple of computers that are equipped with network cards and a cable. The cable used is twisted pair cable of the fourth or fifth category. Twisted pair gets its name because the pairs of wires inside the cable are twisted ( this avoids interference and external influences). You can still find fairly old networks that use coaxial cable. Such networks are obsolete, and the information transmission speed in them does not exceed 10 Mbit/s.

After the network has been created and the computers are connected, you need to configure all the necessary parameters programmatically. First of all, make sure that the computers you are connecting to have operating systems that support networking ( Linux, FreeBSD, Windows)

All computers in a peer-to-peer network are united into workgroups that have their own names ( identifiers).
In the case of a client/server network architecture, access control is carried out at the user level. The administrator has the opportunity to allow access to the resource only to certain users. Let's assume that you make your printer available to network users. If you do not want anyone to print on your printer, then you should set a password for working with this resource. With a peer-to-peer network, anyone who knows your password can gain access to your printer. In a client/server network, you can restrict certain users from using the printer, whether they know the password or not.

To gain access to a resource on a local network built on a client/server architecture, the user must enter a username (Login) and password (Password). It should be noted that the username is public information, and the password is confidential.

The process of verifying a username is called authentication. The process of checking whether the entered password matches the username is authentication. Together, identification and authentication constitute the authorization process. Often the term " authentication" - used in a broad sense: to mean authentication.

From all that has been said, we can conclude that the only advantage of the peer-to-peer architecture is its simplicity and low cost. Client/server networks provide higher levels of performance and security.
Quite often, the same server can perform the functions of several servers, for example, a file server and a web server. Naturally, the total number of functions that the server will perform depends on the load and its capabilities. The higher the server power, the more clients it can serve and the more services it can provide. Therefore, a powerful computer with a large amount of memory and a fast processor is almost always assigned as a server ( As a rule, multiprocessor systems are used to solve serious problems)

Equipment for local network

In the very simple case Network cards and a cable are enough for the network to work. If you need to create a fairly complex network, you will need special network equipment.

Cable

Computers within a local network are connected using cables that transmit signals. A cable connecting two network components ( for example, two computers), is called a segment. Cables are classified depending on the possible values ​​of information transfer speed and the frequency of failures and errors. There are three main categories of cables most commonly used:

• Twisted pair;
• Coaxial cable;
• Fiber optic cable,

Nowadays it is most widely used for building local networks. twisted pair . Inside, such a cable consists of two or four pairs of copper wire twisted together. Twisted pair also has its own varieties: UTP ( Unshielded Twisted Pair - unshielded twisted pair) and STP ( Shielded Twisted Pair - shielded twisted pair). These types of cable are capable of transmitting signals over a distance of about 100 m. As a rule, UTP is used in local networks. STP has a braided copper strand jacket that has a higher level of protection and quality than UTP cable jacket.

In the STP cable, each pair of wires was additionally shielded ( it is wrapped in a layer of foil), which protects the data being transmitted from external interference. This solution allows you to support high speeds transmission over longer distances than when using a UTP cable. The twisted pair cable is connected to the computer using an RJ-45 connector ( Registered Jack 45), which closely resembles an RJ-11 telephone jack ( Regi-steredjack). Twisted pair cable is capable of providing network operation at speeds of 10,100 and 1000 Mbit/s.

Coaxial cable consists of a copper wire covered with insulation, shielding metal braiding and outer shell. The central wire of the cable transmits signals into which the data has previously been converted. Such a wire can be either solid or multi-core. To organize a local network, two types of coaxial cable are used: ThinNet ( thin, 10Base2) and ThickNet ( thick, 10Base5). At the moment, local networks based on coaxial cable are practically not found.

At the core fiber optic cable There are optical fibers (light guides), through which data is transmitted in the form of light pulses. Electrical signals are not transmitted over a fiber optic cable, so the signal cannot be intercepted, which virtually eliminates unauthorized access to the data. Fiber optic cable is used for transportation large volumes information at the highest available speeds.

The main disadvantage of such a cable is its fragility: it is easy to damage, and can only be mounted and connected using special equipment.

Network cards

Network cards make it possible to connect a computer and a network cable. The network card converts the information that is intended to be sent into special packets. A packet is a logical collection of data that includes a header with address information and information itself. The header contains address fields that contain information about the origin and destination of the data. The network card analyzes the destination address of the received packet and determines whether the packet was actually sent to a given computer. If the output is positive, the board will transmit the packet to the operating system. Otherwise, the package will not be processed. Special software allows you to process all packets that pass within the network. This opportunity is used system administrators, when analyzing the operation of the network, and attackers to steal data passing through it.

Any network card has an individual address built into its chips. This address is called the physical or MAC address ( Media Access Control - access control to the transmission medium).

The order of actions performed by the network card is as follows.

1. Receiving information from the operating system and converting it into electrical signals for further sending via cable;
2. Receiving electrical signals over a cable and converting them back into data that the operating system can work with;
3. Determining whether the received data packet is intended specifically for this computer;
4. Controlling the flow of information that passes between a computer and a network.

Hubs

Hub (hub) - a device capable of combining computers into a physical star topology. The hub has several ports that allow you to connect network components. A hub with only two ports is called a bridge. A bridge is required to connect two network elements.

The network together with the hub is " common bus" Data packets when transmitted through the hub will be delivered to all computers connected to the local network.

There are two types of concentrators.

Passive hubs. Such devices send the received signal without it pre-treatment.
Active hubs ( multi-post repeaters). They receive incoming signals, process them and transmit them to connected computers.

Switches

Switches are needed to organize a closer network connection between the sending computer and the destination computer. During data transfer through the switch, information about the MAC addresses of computers is recorded in its memory. Using this information, the switch compiles a routing table, in which for each computer it is indicated that it belongs to a specific network segment.

When the switch receives data packets, it creates a special internal connection ( segment) between two of its Ports using a routing table. It then sends a data packet to the appropriate port on the destination computer, based on the information described in the packet header.

Thus, this connection turns out to be isolated from other ports, which allows computers to exchange information at the maximum speed that is available for a given network. If a switch has only two ports, it is called a bridge.

The switch provides the following features:

• Send a packet with data from one computer to the destination computer;
• Increase data transfer speed.

Routers

A router is similar in principle to a switch, but has a greater range of functionality. It studies not only the MAC, but also the IP addresses of both computers involved in data transfer. When transporting information between different network segments, routers analyze the packet header and try to calculate the optimal path for the packet to travel. The router is able to determine the path to an arbitrary network segment using information from the route table, which allows you to create general connection to the Internet or global network.
Routers allow packets to be delivered in the fastest way, which increases the throughput of large networks. If some network segment is overloaded, the data flow will take a different path,

Network topology

The order in which computers and other elements are located and connected on a network is called network topology. A topology can be compared to a network map, which shows workstations, servers and other network equipment. The selected topology affects general features networks, protocols and network equipment that will be used, as well as the possibility of further expansion of the network.

Physical topology - it is a description of how the physical elements of the network will be connected. Logical topology defines the routes for data packets to take within a network.

There are five types of network topologies:

• Common bus;
• Star;
• Ring;

Common bus

In this case, all computers are connected to one cable, which is called a data bus. In this case, the packet will be received by all computers connected to this network segment.

Network performance is largely determined by the number of computers connected to the common bus. The more such computers there are, the slower the network works. In addition, such a topology can cause various collisions that occur when several computers simultaneously try to transmit information to the network. The likelihood of a collision increases with the number of computers connected to the bus.

Advantages of using networks with a topology " common bus» the following:

• Significant cable savings;
• Easy to create and manage.

Main disadvantages:

• the likelihood of collisions occurring as the number of computers on the network increases;
• a cable break will shut down many computers;
• low level of protection of transmitted information. Any computer can receive data that is transmitted over the network.

Star

When using a star topology, each cable segment coming from any computer on the network will be connected to a central switch or hub. All packets will be transported from one computer to another through this device. Both active and passive hubs can be used. If the connection between the computer and the hub is lost, the rest of the network continues to work. If the hub fails, the network will stop working. With the help of a star structure, even local networks can be connected to each other.

Using this topology is convenient when searching for damaged elements: cables, network adapters or connectors, " Star" more comfortable " common bus"and in case of adding new devices. It should also be taken into account that networks with transmission speeds of 100 and 1000 Mbit/s are built according to the topology “ star».

If in the very center " stars» position the hub, the logical topology will change to a “common bus”.
Advantages " stars»:

• ease of creation and management;
• high level of network reliability;
• high security of information transmitted within the network ( if there is a commutator in the center of the star).

The main disadvantage is that a failure of the hub leads to the cessation of operation of the entire network.

Ring topology

When using a ring topology, all computers on the network are connected to a single ring cable. Packets pass along the ring in one direction through all the network cards of computers connected to the network. Each computer will amplify the signal and send it further along the ring.

In the presented topology, packet transmission along the ring is organized using the token method. A marker is a specific sequence of binary bits containing control data. If network device has a token, then it has the right to send information to the network. Only one token can be transmitted within the ring.

The computer that is about to transport the data takes the token from the network and sends the requested information around the ring. Each subsequent computer will transmit data further until this packet reaches the recipient. Once received, the recipient will return an acknowledgment of receipt to the sending computer, and the latter will create a new token and return it to the network.

The advantages of this topology are as follows:

• Large volumes of data are serviced more efficiently than in the case of a shared bus;
• each computer is a repeater: it amplifies the signal before sending it to the next machine, which allows you to significantly increase the size of the network;
• the ability to set different network access priorities; in this case, a computer with a higher priority will be able to hold the token longer and transmit more information.

Flaws:

• a broken network cable leads to the inoperability of the entire network;
• any computer can receive data that is transmitted over the network.

TCP/IP protocols

TCP/IP protocols ( Transmission Control Protocol/Internet Protocol - Data Transmission Control Protocol/Internet Protocol) are the main internetworking protocols and manage data transfer between networks of different configurations and technologies. It is this family of protocols that is used to transmit information on the Internet, as well as in some local networks. The TPC/IP protocol family includes all intermediate protocols between the application layer and the physical layer. Their total number is several dozen.

The main ones are:

• Transport protocols: TCP - Transmission Control Protocol ( transmission control protocol) and others - manage data transfer between computers;
• Routing protocols: IP - Internet Protocol ( internet protocol) and others - ensure the actual transfer of data, process data addressing, determine the best path to the recipient;
• Network address support protocols: DNS - Domain Name System ( domain name system) and others - provides determination of the unique address of the computer;
• Application service protocols: FTP - File Transfer Protocol ( file transfer protocol), HTTP - HyperText Transfer Protocol, TELNET and others - are used to gain access to various services: file transfer between computers, access to WWW, remote terminal access to the system, etc.;
• Gateway protocols: EGP - Exterior Gateway Protocol ( external gateway protocol) and others - help transmit routing messages and network status information over the network, as well as process data for local networks;
• Postal protocols: POP - Post Office Protocol ( mail reception protocol) - used to receive messages Email, SMPT Simple Mail Transfer Protocol ( mail transfer protocol) - used to transmit mail messages.

All major network protocols ( NetBEUI, IPX/SPX and TCIP) are routed protocols. But you only have to manually configure TCPIP routing. Other protocols are routed automatically by the operating system.

IP addressing

When building a local network based on the TCP/IP protocol, each computer receives a unique IP address, which can be assigned either by a DHCP server - a special program installed on one of the network computers, or using Windows tools, or manually.

A DHCP server allows you to flexibly distribute IP addresses to computers and assign permanent, static IP addresses to some computers. The built-in Windows tool does not have such capabilities. Therefore, if there is a DHCP server on the network, then it is better not to use Windows tools, setting automatic ( dynamic) assigning an IP address. Installing and configuring a DHCP server is beyond the scope of this book.

It should be noted, however, that if you use a DHCP server or tools to assign an IP address Windows boot computers on the network and the operation of assigning IP addresses requires a long time, the more more network. In addition, the computer with the DHCP server must be turned on first.
If you manually assign static networks to computers ( constant, unchanging) IP addresses, then computers will boot faster and immediately appear in the network environment. For small networks, this option is the most preferable, and it is what we will consider in this chapter.

For the TCP/IP protocol bundle, the IP protocol is the basic one, since it is the one that deals with the movement of data packets between computers through networks that use various network technologies. It is thanks to the universal characteristics of the IP protocol that the very existence of the Internet, consisting of a huge number of heterogeneous networks, became possible.

IP protocol data packets

The IP protocol is the delivery service for the entire TCP-iP family of protocols. Information coming from other protocols is packaged into IP protocol data packets, an appropriate header is added to them, and the packets begin their journey through the network

IP addressing system

Some of the most important fields in the IP data packet header are the packet's source and destination addresses. Each IP address must be unique on the internetwork where it is used in order for the packet to reach its intended destination. Even on the entire global Internet, it is impossible to find two identical addresses.

IP address, as opposed to a regular one postal address, consists exclusively of numbers. It occupies four standard computer memory cells - 4 bytes. Since one byte (Byte) is equal to 8 bits (Bit), the length of the IP address is 4 x 8 = 32 bits.

A bit represents the smallest possible unit of information storage. It can only contain 0 ( bit cleared) or 1 ( bit set).

Although an IP address is always the same length, it can be written in different ways. The format for recording an IP address depends on the number system used. At the same time, the same address can look completely different:

Numerical format	Meaning
Binary
Hexadecimal(Hexadecimal)	0x86180842
Decimal	2249721922
Dotted decimal(Dotted Decimal)	134.24.8.66

The binary number 10000110 is converted to decimal as follows: 128 + 0 + 0 + 0 + 0 + 4 + 2 + 0 =134.
The most preferable option, from the point of view of human readability, is the format of writing the IP address in dotted decimal notation. This format consists of four decimal numbers, separated by dots. Each number, called an Octet, represents the decimal value of the corresponding byte in the IP address. An octet is so called because one byte in binary consists of eight bits.

When using dotted decimal notation to write octets in an IP address, keep the following rules in mind:

• Only integers are valid;
• Numbers must be in the range from 0 to 255.

The most significant bits in the IP address, located on the left, determine the class and number of the network. Their collection is called a subnet identifier or network prefix. When assigning addresses within the same network, the prefix always remains unchanged. It identifies the ownership of an IP address on a given network.

For example, if the IP addresses of computers on the subnet are 192.168.0.1 - 192.168.0.30, then the first two octets define the subnet ID - 192.168.0.0, and the next two - host IDs.

Exactly how many bits are used for certain purposes depends on the class of the network. If the host number is zero, then the address does not point to any one specific computer, but to the entire network as a whole.

Network classification

There are three main classes of networks: A, B, C. They differ from each other by the maximum possible number of hosts that can be connected to a network of a given class.

The generally accepted classification of networks is shown in the following table, which shows the largest number network interfaces available for connection, which octets of the IP address are used for network interfaces (*), and which remain unchanged (N).

Network class	Largest quantity hosts	Variable IP address octets used for host numbering
	16777214	N *.*.*
	65534	N.N.*.*
		N.N.N.*

For example, in the most common class C networks there cannot be more than 254 computers, so only one, the lowest byte of the IP address, is used to number network interfaces. This byte corresponds to the rightmost octet in dotted decimal notation.

A legitimate question arises: why can only 254 computers be connected to a class C network, and not 256? The fact is that some intranet IP addresses are intended for special use, namely:

O - identifies the network itself;
255 - broadcast.

Network segmentation

The address space within each network can be divided into smaller subnets based on the number of hosts ( Subnets). The process of subnetting is also called segmentation.

For example, if the class C network 192.168.1.0 is divided into four subnets, then their address ranges will be as follows:

• 192.168.1.0-192.168.1.63;
• 192.168.1.64-192.168.1.127;
• 192.168.1.128-192.168.1.191;
• 192.168.1.192-192.168.1.255.

IN in this case For host numbering, not the entire right octet of eight bits is used, but only the least significant 6 of them. And the remaining two most significant bits determine the subnet number, which can take values ​​from zero to three.

Both regular and extended network prefixes can be identified using a subnet mask ( Subnet Mask), which also allows you to separate the subnet identifier from the host identifier in the IP address, masking with a number the part of the IP address that identifies the subnet.

The mask is a combination of numbers that resembles an IP address in appearance. The binary representation of the subnet mask contains zeros in the bits that are interpreted as the host number. The remaining bits set to one indicate that this part of the address is a prefix. The subnet mask is always used in conjunction with the IP address.

In the absence of additional subnetting, standard network class masks have the following meanings:

Network class	Mask
	binary	dotted decimal
	11111111.00000000.00000000.00000000	255.0.0.0
	11111111.11111111.00000000.00000000	255.255.0.0
	11111111.11111111.11111111.00000000	255.255.255.0

When the subnetting mechanism is used, the mask is modified accordingly. Let us explain this using the already mentioned example of dividing a class C network into four subnets.

In this case, the two most significant bits in the fourth octet of the IP address are used to number subnets. Then the mask in binary form will look like this: 11111111.11111111.11111111.11000000, and in dotted decimal form -255.255.255.192.

Private network address ranges

Each computer connected to the network has its own unique IP address. For some machines, such as servers, this address does not change. Such Permanent Address called static. For others, such as clients, the IP address can be permanent (static) or assigned dynamically each time they connect to the network.

To obtain a unique static, that is, permanent IP address on the Internet, you need to contact a special organization InterNIC - Internet Network Information Center ( Internet Network Information Center). InterNIC assigns only a network number, and the network administrator must handle further work on creating subnets and numbering hosts independently.

But official registration in InterNIC in order to obtain a static IP address is usually required for networks that have constant communication with the Internet. For private networks that are not part of the Internet, several blocks of address space are specially reserved, which can be freely used to assign IP addresses without registering with InterNIC:

Network class	Number of available network numbers	IP address ranges used for host numbering
		10.0.0.0 - 10.255.255.255
		172.16.0.0-172.31.255.255
		192.168.0.O-192.168.255.255
LINKLOCAL		169.254.0.0-169.254.255.255

However, these addresses are used only for internal addressing of networks and are not intended for hosts that directly connect to the Internet.

The LINKLOCAL address range is not a network class in the usual sense. It is used by Windows to automatically assign personal IP addresses to computers on the local network.

I hope you now have an idea about the local network!

In any organization where there are two or more computers, it is advisable to combine them into local network. The network allows employees to quickly exchange information and documents with each other, serves for sharing shared Internet access, equipment and storage devices.
To combine computers we need a certain network hardware. In today's article we will look at what equipment is used to create wired LAN.

network hardware – devices that make up a computer network. There are two types of network equipment:

• Active network equipment is equipment that is capable of processing or converting information transmitted over the network. Such equipment includes network cards, routers, and print servers.
• Passive network equipment is equipment used for simple signal transmission at the physical level. These are network cables, connectors and network sockets, repeaters and signal amplifiers.

To install a wired local network, we first need:

• network cable and connectors (called connectors);
• network cards - one in each PC on the network, and two on the computer serving as a server for accessing the Internet;
• a device or devices that ensure the transfer of packets between computers on a network. For networks of three or more computers you need special device– which unites all computers on the network;
• additional network devices. The simplest network is built without such equipment, however, when organizing a general Internet connection, using common network printers additional devices can make solving such problems easier.

Now let’s take a closer look at all the equipment listed above:

Network Explorers

This group includes various network cables(twisted pair, coaxial cable, fiber optic).

Coaxial cable - This is the first cable that was used to create networks. From its use in building local computer networks They gave up a long time ago.

Fiber optic cable – the most promising in terms of speed performance, but also more expensive compared to coaxial cable or twisted pair. In addition, the installation of fiber optic networks requires high qualifications, and expensive equipment is needed to terminate the cable. For these reasons, this type of cable has not yet become widespread.

twisted pair – the most common type of cable used to build local networks today. The cable consists of pairs of intertwined copper insulated conductors. A typical cable has 8 conductors (4 pairs), although cables with 4 conductors (2 pairs) are also available. The colors of the internal insulation of conductors are strictly standard. The distance between devices connected by twisted pair cables should not exceed 100 meters.
There are several categories of twisted pair cables, which are labeled CAT1 to CAT7. In local networks of the Ethernet standard, twisted pair cables are used CAT5.

Connectors are used to work with twisted pair cables RJ-45.

Network cards

Network cards are responsible for transferring information between network computers. A network card consists of a connector for a network conductor (usually a twisted pair cable) and a microprocessor that encodes/decodes network packets. A typical network card is a card that plugs into a slot PCI buses. Almost all modern computers contain electronics network adapter soldered directly to the motherboard. Instead of an internal network card, you can use external USB network adapter: It is a USB-LAN adapter and has similar functions to its PCI counterparts. The main advantage of USB network cards is their versatility: without opening the case system unit This adapter can be connected to any PC that has a free USB port. Also USB adapter will be indispensable for a laptop in which the only built-in network connector has failed, or there is a need for two network ports.

Network switches

Not so long ago, network networks were used to build local networks. hubs (or, in common parlance, hubs ). When a network card sends a data packet from a computer to the network, the hub simply amplifies the signal and transmits it to all network participants. Only the network card to which it is addressed receives and processes the packet; the others ignore it. Essentially, a hub is a signal amplifier.

Currently used in local networks (or as they are called, switches ). These are more “intelligent” devices, which have their own processor, internal bus and buffer memory. If the hub simply forwards packets from one port to all others, then the switch analyzes the addresses of the network cards connected to its ports and forwards the packet only to the desired port. As a result, unnecessary traffic on the network is sharply reduced. This allows you to significantly increase network performance and provides higher data transfer speeds in networks with a large number of users. The switch can operate at speeds of 10, 100 or 1000 Mbps. This, as well as the network cards installed on computers, determines the speed of the network segment. Another characteristic of a switch is the number of ports. This determines the number of network devices that can be connected to the switch. In addition to computers, they include print servers, modems, network drives and other devices with a LAN interface.

When designing a network and choosing a switch, you need to take into account the possibility of expanding the network in the future - it is better to purchase a switch with a slightly larger number of ports than the number of computers in your network at the moment. In addition, one port must be kept free in case it is combined with another switch. Currently, switches are connected by ordinary twisted pair cable of the fifth category, exactly the same one that is used to connect each computer on the network to the switch.

There are two types of switches - managed and unmanaged. Managed ones have additional functionality. Thus, it becomes possible to manage the switch using the web interface, combine several switches into one virtual one with its own packet switching rules, etc. The cost of managed switches is much higher than the cost of unmanaged switches, which is why unmanaged switches are used in small and medium-sized networks.

Additional network equipment

In a local network, you can use various additional equipment, for example, to combine two networks or to protect the network from external attacks. Let's briefly look at the network equipment that is used to build computer networks.

Print server , or print server is a device that allows you to connect a printer that does not have its own network port to the network. Simply put: a print server is a box to which a printer is connected on one side, and a network cable on the other side. In this case, the printer becomes available at any time, since it is not tied to any computer on the network. There are print servers with different ports: USB and LPT; There are also combined options. Repeater Designed to increase network connection distance by amplifying the electrical signal. If you use a twisted pair cable more than 100 meters long in your local network, repeaters should be installed in the cable break every 100 meters. Repeaters are usually powered via the same cable. Using repeaters, you can connect several separate buildings with a network cable. Router (or ) is a network device that, based on information about the network structure, uses a certain algorithm to select a route for sending packets between different network segments.

Routers are used to connect networks of different types, often incompatible in architecture and protocols (for example, to connect Ethernet to a WAN network). The router is also used to provide access from the local network to the global Internet, while performing the functions of a firewall. A router can be presented not only in hardware form, but also in software form. Any computer on the network with the appropriate software installed can serve as a router.

Connecting networks using bridges, switches and routers

Network aggregation devices provide communication between local network segments, individual LANs and subnets of any level. These devices in the most general form can be attributed to certain levels of the reference model of open systems interaction.

The relationship between the functions of these devices and the layers of the OSI model is shown in Figure 31.

Figure 31 - Correspondence of functions of communication equipment to the OSI model

There are the following classes of devices for connecting segments and networks. A repeater, which regenerates signals, thereby allowing you to increase the length of the network, operates at the physical level.

The network adapter also works at the physical and partly at the data link layers. The physical layer includes that part of the functions of the network adapter that is associated with the reception and transmission of signals over the communication line, and gaining access to the shared transmission medium and recognizing the MAC address of the computer is already a function of the link layer.

Bridges and switches connect networks at the data link layer and use the functionality physical level. Bridges are executed on the basis of a computer equipped with appropriate software. The difference between switches and bridges is that they implement their functions in hardware and therefore have significantly higher performance;

For bridges, the network is represented by a set of device MAC addresses. They extract these addresses from headers added to packets at the data link layer and use them during packet processing to decide which port to send a particular packet to. Bridges do not have access to network address information related to more high level. Therefore, they are limited in making decisions about possible paths or routes for packets to travel through the network. .

Routers operate at the network layer of the OSI model. For routers, a network is a set network addresses devices and multiple network paths. Routers analyze all possible paths between any two network nodes and choose the shortest one.

Figure 32 shows another type of communication device - a gateway, which can operate at any level of the OSI model. Gateway(gateway) is a device that performs protocol translation. A gateway is placed between communicating networks and serves as an intermediary, translating messages coming from one network into the format of another network. The gateway can be implemented either purely by software installed on a regular computer, or on the basis of a specialized computer.

A fragment of a computer network (Figure 32) includes the main types of communication equipment for the formation of local networks and connecting them through global connections with each other.

Figure 32 - Network fragment

To connect local networks to global relations special outputs (WAN ports) of bridges and routers are used, as well as data transmission equipment over long lines - modems (when working over analog lines) or devices connecting to digital channels(TA - ISDN network terminal adapters, digital leased channel service devices such as CSU/DSU, etc.).