Computer networks and telecommunications
DNS Domain Name System
The correspondence between domain names and IP addresses can be established either by means of the local host or by means of a centralized service. In the early days of the Internet, a text file with the well-known name hosts was manually created on each host. This file consisted of a number of lines, each containing one IP address - domain name pair, such as 102.54.94.97 - rhino.acme.com.
As the Internet grew, hosts files also grew, and building a scalable name resolution solution became a necessity.
This solution was a special service - the domain name system (Domain Name System, DNS). DNS is a centralized service based on a distributed database of domain name - IP address mappings. The DNS service uses a client-server protocol in its work. It defines DNS servers and DNS clients. DNS servers maintain a distributed database of mappings, and DNS clients contact servers with requests to resolve a domain name to an IP address.
The DNS service uses text files that are almost the same format as the hosts file, and these files are also manually prepared by the administrator. However, the DNS service relies on a hierarchy of domains, and each DNS service server stores only a subset of the network names, not all of the names, as is the case with hosts files. With an increase in the number of nodes in the network, the problem of scaling is solved by creating new domains and subdomains of names and adding new servers to the DNS service.
Each name domain has its own DNS server. This server can store mappings "domain name - IP address" for the entire domain, including all its subdomains. However, in this case, the solution turns out to be poorly scalable, since when new subdomains are added, the load on this server may exceed its capabilities. More often, a domain server only stores names that end at the next lower level in the hierarchy than the domain name. (Similar to a file system directory, which contains entries about the files and subdirectories "included" directly in it.) It is with this organization of the DNS service that the name resolution load is distributed more or less evenly among all the DNS servers on the network. For example, in the first case, the DNS server of the mmt.ru domain will store mappings for all names ending in mmt.ru: wwwl.zil.mmt.ru, ftp.zil.mmt.ru, mail.mmt.ru, etc. In the second case, this server stores only mappings of names like mail.mmt.ru, www.mmt.ru, and all other mappings must be stored on the DNS server of the zil subdomain.
Each DNS server, in addition to the name mapping table, contains links to the DNS servers of its subdomains. These links link individual DNS servers into a single DNS service. The links are the IP addresses of the respective servers. To service the root domain, several duplicating each other DNS servers are allocated, the IP addresses of which are widely known (they can be found, for example, in InterNIC).
The procedure for resolving a DNS name is in many ways similar to the procedure for the file system to look up the address of a file by its symbolic name. Indeed, in both cases, the composite name reflects the hierarchical structure of the organization of the corresponding directories - file directories or DNS tables. Here, the domain and domain DNS server are analogous to a file system directory. Domain names, like symbolic file names, are naming independent of physical location.
The procedure for searching for a file address by a symbolic name consists in sequential browsing of directories, starting from the root. This pre-checks the cache and the current directory. To determine an IP address from a domain name, it is also necessary to look at all DNS servers serving the chain of subdomains included in the host name, starting from the root domain. The essential difference is that the file system is located on one computer, and the DNS service is distributed by its nature.
There are two main DNS name resolution schemes. In the first option, the DNS client coordinates the work on finding the IP address:
The DNS client contacts the root DNS server with the FQDN;
The DNS server responds with the address of the next DNS server serving the top-level domain given in the upper part of the requested name;
The DNS client makes a request to the next DNS server, which sends it to the DNS server of the desired subdomain, and so on, until a DNS server is found that stores the match of the requested name to the IP address. This server gives the final answer to the client. Such an interaction scheme is called non-recursive or iterative, when the client itself iteratively performs a sequence of queries to different name servers. Since this scheme loads the client with rather complex work, it is rarely used. In the second option, a recursive procedure is implemented:
The DNS client queries the local DNS server, that is, the server that serves the subdomain to which the client name belongs;
If the local DNS server knows the answer, then it immediately returns it to the client; this may correspond to the case where the requested name is in the same subdomain as the client's name, and may also correspond to the case where the server already knew this match for another client and stored it in its cache;
If the local server does not know the answer, then it makes iterative requests to the root server, etc., in the same way as the client did in the first option; having received a response, it passes it to the client, which all this time was just waiting for it from its local DNS server.
In this scheme, the client delegates work to its server, so the scheme is called indirect or recursive. Almost all DNS clients use the recursive procedure.
TCP/IP protocol stack.
The TCP/IP stack, also called the DoD stack and the Internet stack, is one of the most popular and promising communication protocol stacks. If at present it is distributed mainly in UNIX networks, then its implementation in the latest versions of network operating systems for personal computers (Windows NT, NetWare) is a good prerequisite for the rapid growth in the number of installations of the TCP/IP stack.
The stack was developed at the initiative of the US Department of Defense (DoD) more than 20 years ago to connect the experimental ARPAnet network with other satellite networks as a set of common protocols for a heterogeneous computing environment. The ARPA network supported developers and researchers in the military fields. In the ARPA network, communication between two computers was carried out using the Internet Protocol (IP), which to this day is one of the main ones in the TCP / IP stack and appears in the name of the stack.
The University of Berkeley made a major contribution to the development of the TCP / IP stack by implementing the stack protocols in its version of the UNIX OS. The widespread adoption of the UNIX operating system led to the widespread adoption of the IP protocol and other stack protocols. This stack is also used by the Internet, whose Internet Engineering Task Force (IETF) is the main contributor to the development of the stack's standards, published in the form of RFC specifications.
Since the TCP/IP stack was developed before the advent of the ISO/OSI Open Systems Interconnection model, although it also has a layered structure, the correspondence between the levels of the TCP/IP stack and the levels of the OSI model is rather arbitrary.
The lowest (layer IV) - the level of gateway interfaces - corresponds to the physical and data link layers of the OSI model. This level is not regulated in TCP/IP protocols, but it supports all popular physical and data link level standards: for local channels it is Ethernet, Token Ring, FDDI; point-to-point connections via WAN serial links, and X.25 and ISDN area network protocols. A special specification has also been developed that defines the use of ATM technology as a link layer transport.
The next layer (layer III) is the internetworking layer, which deals with the transmission of datagrams using various local area networks, X.25 territorial networks, ad hoc links, etc. As the main network layer protocol (in terms of the OSI model) in the stack the IP protocol is used, which was originally designed as a protocol for transmitting packets in composite networks, consisting of a large number of local networks, united by both local and global links. Therefore, the IP protocol works well in networks with a complex topology, rationally using the presence of subsystems in them and economically consuming the bandwidth of low-speed communication lines. The IP protocol is a datagram protocol.
The internetworking layer also includes all protocols related to the compilation and modification of routing tables, such as the routing information collection protocols RIP (Routing Internet Protocol) and OSPF (Open Shortest Path First), as well as the Internet Control Message Protocol (ICMP) ). The latter protocol is designed to exchange information about errors between the router and the gateway, the source system and the receiver system, that is, to organize feedback. With the help of special ICMP packets, it is reported about the impossibility of delivering a packet, about exceeding the lifetime or duration of the packet assembly from fragments, about anomalous parameter values, about changing the forwarding route and type of service, about the state of the system, etc.
The next level (level II) is called the main level. The Transmission Control Protocol (TCP) and the User Datagram Protocol (UDP) operate at this layer. The TCP protocol provides a stable virtual connection between remote application processes. The UDP protocol provides the transfer of application packets using the datagram method, that is, without establishing a virtual connection, and therefore requires less overhead than TCP.
The top level (level I) is called the application level. Over the years of use in the networks of various countries and organizations, the TCP / IP stack has accumulated a large number of protocols and application-level services. These include such widely used protocols as the FTP file copy protocol, the telnet terminal emulation protocol, the SMTP mail protocol used in Internet e-mail and its Russian branch RELCOM, hypertext services for accessing remote information, such as WWW and many others. Let us dwell in more detail on some of them, which are most closely related to the subject of this course.
SNMP (Simple Network Management Protocol) is used for organizing network management. The control problem is divided here into two tasks. The first task is related to the transfer of information. Control information transfer protocols define the procedure for interaction between the server and the client program running on the administrator's host. They define the message formats exchanged between clients and servers, as well as the formats for names and addresses. The second task is related to controlled data. The standards govern what data must be stored and accumulated in the gateways, the names of this data and the syntax of these names. The SNMP standard defines the specification of the network management information database. This specification, known as the Management Information Base (MIB), defines the data elements that a host or gateway must store and the allowed operations on them.
The File Transfer Protocol (FTP) provides remote access to a file. In order to ensure reliable transmission, FTP uses the connection-oriented protocol - TCP - as a transport. Besides the file transfer protocol, FTP offers other services. So the user is given the opportunity to work interactively with a remote machine, for example, he can print the contents of its directories, FTP allows the user to specify the type and format of the stored data. Finally, FTP performs user authentication. Users are required by protocol to provide their username and password before accessing the file.
Within the TCP/IP stack, FTP offers the most extensive file services, but it is also the most complex to program. Applications that do not need all the features of FTP can use another, more economical protocol - the simplest file transfer protocol TFTP (Trivial File Transfer Protocol). This protocol implements only file transfer, and the connectionless protocol, UDP, which is simpler than TCP, is used as a transport.
The telnet protocol provides a stream of bytes between processes and between a process and a terminal. Most often, this protocol is used to emulate the terminal of a remote computer.
BGP protocol
The general scheme of how BGP works is as follows. BGP routers of neighboring ASs that have decided to exchange routing information establish BGP connections between themselves and become BGP neighbors (BGP peers).
Further, BGP uses an approach called path vector, which is an evolution of the distance vector approach. BGP neighbors send (announce, advertise) each other path vectors. The path vector, unlike the distance vector, contains not just the network address and distance to it, but the network address and a list of path attributes that describe various characteristics of the route from the source router to the specified network. In what follows, for brevity, we will call the data set consisting of the network address and path attributes to this network a route to this network.
BGP Implementation
A pair of BGP neighbors establishes a TCP connection between themselves, port 179. Neighbors belonging to different ASs must be directly accessible to each other; for neighbors from the same AS, there is no such restriction, since the internal routing protocol will ensure the availability of all necessary routes between nodes of the same autonomous system.
The information flow exchanged between BGP neighbors via TCP consists of a sequence of BGP messages. The maximum message length is 4096 octets, the minimum is 19. There are 4 types of messages.
BGP message types
- OPEN - sent after a TCP connection has been established. The response to OPEN is a KEEPALIVE message if the other party agrees to become a BGP neighbor; otherwise, a NOTIFICATION message is sent with a code explaining the reason for the failure, and the connection is terminated.
- KEEPALIVE - the message is intended to confirm consent to establish neighbor relations, as well as to monitor the activity of an open connection: for this, BGP neighbors exchange KEEPALIVE messages at certain time intervals.
- UPDATE - the message is intended for announcing and revoking routes. Once a connection is established, UPDATE messages send all the routes that the router wants to advertise to the neighbor (full update), after which only data about added or removed routes is sent as they become available (partial update).
- NOTIFICATION - a message of this type is used to inform the neighbor about the reason for closing the connection. After this message is sent, the BGP connection is closed.
BGP message format
A BGP message consists of a header and a body. The header is 19 octets long and consists of the following fields:
marker: in the OPEN message always, and when working without authentication - in other messages, filled with ones. Otherwise, it contains authentication information. A related function of the marker is to improve the reliability of highlighting the message boundary in the data stream.
The length of the message in octets, including the header.
IGRP protocol
The Interior Gateway Routing Protocol (IGRP) is a routing protocol developed in the mid-1980s. by Cisco Systems, Inc. The main goal was to provide a robust protocol for routing within an Autonomous System (AS) having an arbitrarily complex topology and including media with varied bandwidth and delay characteristics.
IGRP is an interior router protocol (IGP) with a distance vector. Distance vector routing protocols require each router to send all or part of its routing table in route update messages to all neighboring routers at regular intervals. As routing information propagates through the network, routers can calculate distances to all nodes in the internetwork.
IGRP uses a combination (vector) of indicators. Internetwork delay, bandwidth (bandwidth), reliability (reliability) and load (load) - all these indicators are taken into account as coefficients when making a routing decision. Network administrators can set weighting factors for each of these metrics. IGRP provides for a wide range of values for its indicators.
To provide additional flexibility, IGRP allows multipath routing. Duplicated lines with the same bandwidth can carry a separate traffic flow in a cyclic manner with automatic switching to the second line if the first line fails.
Packet Format
The first field of the IGRP packet contains the version number.
Operational code field (opcode). This field indicates the type of package. An opcode of 1 denotes an update package (contains a header immediately followed by routing table data records); equal to 2-packet request (used by the source to query the routing table from another router.
Edition field. This release number value is used to allow routers to avoid processing updates containing information they have already seen.
The next three fields indicate the number of subnets, the number of main networks, and the number of external networks in the update package.
Checksum field. The checksum calculation allows the receiving router to verify the validity of the incoming packet.
Stability characteristics
IGRP has a number of features designed to enhance its stability. These include:
Temporary hold changes are used to prevent regular correction messages from illegally reclaiming a route that may have been corrupted. The change retention period is usually calculated to be longer than the time required for the entire network to adjust to any routing change.
Split Horizons The concept of split horizons stems from the fact that it is never useful to send information about a route back in the direction it came from. The split-horizon rule helps prevent route loops.
Route cancellation adjustments are designed to deal with larger route loops. An increase in routing metrics usually indicates the appearance of routing loops. In this case, cancellation updates are sent to remove the route and place it on hold.
IGRP provides a number of timers and variables containing time intervals. This includes
- update timer (determines how often route update messages should be sent),
- dead route timer, determines how long the router should wait in the absence of messages about the correction of a particular route before declaring this route as dead
- change holding time period
- shutdown timer. specifies how much time must elapse before any router must be excluded from the routing table.
Network layer protocols are implemented, as a rule, in the form of software modules and run on end computer nodes, called hosts, as well as on intermediate nodes - routers, called gateways. The functions of routers can be performed by both specialized devices and universal
The concept of internetworking
The main idea behind the introduction of the network layer is as follows. A network is generally considered as a collection of several networks and is called a composite network or internet. (internetwork or internet). The networks that make up a composite network are called subnets. (subnet), constituting networks or simply networks (Fig. 5.1). Subnets are interconnected by routers. The components of a composite network can be both local and global networks. The internal structure of each network is not shown in the figure, as it does not matter when considering the network protocol. All nodes within the same subnet communicate using the same technology for them. So, the composite network shown in the figure includes several networks of different technologies: local networks Ethernet, Fast Ethernet, Token Ring, FDDI and WAN frame relay, X.25, ISDN. Each of these technologies is sufficient to organize the interaction of all nodes in its subnet, but is not able to build an information link between arbitrarily selected nodes belonging to different subnets, for example, between node A and node B in Fig. 5.1. Therefore, to organize the interaction between any arbitrary pair of nodes of this "large" composite network, additional funds are required. Such means are provided by the network layer.
The network layer acts as a coordinator that organizes the work of all subnets that lie in the path of the packet's progress through the composite network. To move data within subnets, the network layer refers to the technologies used on those subnets.
While many LAN technologies (Ethernet, Token Ring, FDDI, Fast Ethernet, etc.) use the same host addressing system based on MAC addresses, there are many technologies (X.25, ATM, frame relay), that use different addressing schemes. Addresses assigned to nodes in accordance with subnetting technologies are called local. In order for the network layer to fulfill its task, it needs its own addressing system, independent of the addressing methods of nodes in individual subnets, which would allow the network layer to identify any node of the composite network in a universal and unambiguous way.
The natural way to form a network address is to uniquely number all the subnets of a composite network, and number all nodes within each subnet. Thus, a network address is a pair: a network (subnet) number and a host number.
The node number can be either the local address of this node (such a scheme is adopted in the IPX / SPX stack), or some number, unrelated to local technology, that uniquely identifies a node within a given subnet. In the first case, the network address becomes dependent on local technologies, which limits its use. For example, IPX/SPX network addresses are designed to work in composite networks that combine networks that use only MAC addresses or addresses of a similar format. The second approach is more general and is specific to the TCP/IP stack. In both cases, each node of the composite network has, along with its local address, one more - the universal network address.
Data that enters the network layer and that needs to be sent over the composite network is provided with a network layer header. The data together with the header form a packet. The network layer packet header has a unified format that does not depend on the link layer frame formats of those networks that may be part of the internetwork, and carries, along with other service information, data on the number of the network to which this packet is intended. The network layer determines the route and moves the packet between subnets.
When a packet is transmitted from one subnet to another, the network layer packet encapsulated in the arriving link frame of the first subnet is stripped of that frame's headers and surrounded by the link layer frame headers of the next subnet. The information on the basis of which this replacement is made are the service fields of the network layer packet. The new frame's destination address field specifies the local address of the next router.
Ethernet hubs
In Ethernet technology, devices that combine several physical segments of a coaxial cable into a single shared environment have been used for a long time and are called "repeaters" for their main function - repeating on all their ports the signals received at the input of one of the ports. In networks based on coaxial cable, two-port repeaters were common, connecting only two cable segments, so the term hub was not usually applied to them.
With the advent of the lOBase-T specification for twisted pair, the repeater became an integral part of the Ethernet network, since without it, communication could only be established between two network nodes. Multiport twisted-pair Ethernet repeaters began to be called hubs or hubs, since connections between a large number of network nodes were really concentrated in one device. An Ethernet hub typically has 8 to 72 ports, with most of the ports dedicated to connecting twisted-pair cables. On fig. Figure 2 shows a typical Ethernet hub designed to form small segments of a shared environment. It has 16 lOBase-T ports with RJ-45 connectors, as well as one AUI port for connecting an external transceiver.
Typically, a transceiver operating on a coaxial or fiber optic is connected to this port. Using this transceiver, the hub is connected to a trunk cable connecting several hubs to each other, or in this way, a station is connected that is more than 100 m away from the hub.
Rice. 15. Ethernet hub.
To connect lOBase-T technology hubs to each other in a hierarchical system, a coaxial or fiber optic cable is not necessary; you can use the same ports as for connecting end stations, subject to one circumstance. The fact is that a regular RJ-45 port, designed to connect a network adapter and called MDI-X (crossed MDI), has an inverted pin assignment of the connector so that the network adapter can be connected to a hub using a standard connecting cable that does not cross contacts.
When connecting hubs through a standard MDI-X port, a non-standard crossover cable must be used. Therefore, some manufacturers provide the hub with a dedicated MDI port that does not have crossover pairs. Thus, two hubs can be connected with an ordinary non-crossover cable if this is done through the MDI-X port of one hub and the MDI port of the second. More often than not, a single hub port can function as both an MDI-X port and an MDI port, depending on the position of the pushbutton switch.
A multiport Ethernet repeater hub can be treated differently when using the 4-hub rule. In most models, all ports are connected to a single repeater block, and when a signal passes between two repeater ports, the repeater block introduces a delay only once. Therefore, such a hub should be considered as one repeater with the restrictions imposed by the 4-hub rule. But there are other models of repeaters, in which several ports have their own repetition block.
In this case, each repetition block should be considered a separate repeater and counted separately in the 4-hub rule.
Some differences may be shown by models of hubs operating on a single-mode fiber optic cable. The range of a cable segment supported by an FDDI hub on such a cable can vary significantly depending on the power of the laser emitter - from 10 to 40 km.
However, if the existing differences in the performance of the main function of the concentrators are not so great, then they are much larger than the spread in the possibilities for the implementation of additional functions by the concentrators. Disable ports.
Very useful in network operation is the ability of a hub to disable malfunctioning ports, thereby isolating the rest of the network from problems that have arisen in the node. This feature is called auto-segmentation. For the FDDI hub, this function is the main one for many error situations, as it is defined in the protocol. At the same time, for an Ethernet or Token Ring hub, the auto-segmentation function is optional for many situations, since the standard does not describe the hub's response to this situation. The main reason for port disablement in Ethernet and Fast Ethernet standards is the lack of response to the link test pulse train sent to all ports every 16 ms. In this case, the failed port is placed in the "disabled" state, but link test pulses will continue to be sent to the port so that when the device is restored, work with it will be continued automatically.
Consider the situations in which Ethernet and Fast Ethernet hubs disable a port:
o Frame-level errors. If the rate of passing frames with errors through the port exceeds the specified threshold, then the port is disabled, and then, if there are no errors within the specified time, it is enabled again. Such errors can be: incorrect checksum, incorrect frame length (greater than 1518 bytes or less than 64 bytes), unformatted frame header.
o Multiple collisions. If the hub detects that the same port was the source of the collision 60 times in a row, then the port is disabled. After a while, the port will be enabled again.
o Long transfer (jabber). Like a network adapter, a hub controls how long it takes for one frame to pass through a port. If this time exceeds the maximum length frame transmission time by 3 times, then the port is disabled.
Backup link support
Since the use of redundant links in hubs is defined only in the FDDI standard, for other standards, hub developers support this feature with their own proprietary solutions. For example, Ethernet/Fast Ethernet hubs can only form hierarchical links without loops. Therefore, redundant links should always connect disabled ports so as not to violate the logic of the network.
Usually, when configuring a hub, the administrator must determine which ports are the main ones, and which ones are reserved in relation to them (Fig. 16). If for any reason the port is disabled (auto-segmentation mechanism is triggered), the hub makes its backup port active.
Rice. sixteen.
Rice. 16. Redundant links between Ethernet hubs.
When considering some hub models, the question arises - why does this model have such a large number of ports, for example 192 or 240? Does it make sense to share a 10 or 16 Mbps medium between so many stations? Perhaps ten or fifteen years ago the answer might have been yes in some cases, such as those networks where computers used the network only to send small mail messages or to rewrite a small text file.
Today, there are very few such networks left, and even 5 computers can fully load an Ethernet or Token Ring segment, and in some cases, a Fast Ethernet segment. Why, then, do you need a hub with a large number of ports, if they are practically impossible to use due to bandwidth limitations per station? The answer is that such hubs have several unconnected internal buses that are designed to create multiple shared environments.
For example, the hub shown in Fig. 17 has three internal Ethernet buses. If, for example, such a hub has 72 ports, then each of these ports can be connected to any of the three internal buses. In the figure, the first two computers are connected to Ethernet bus 3, and the third and fourth computers are connected to Ethernet bus 1. The first two computers form one shared segment, and the third and fourth computers form another shared segment.
Rice. 17. Multi-segment hub.
Computers connected to different segments cannot communicate with each other through the hub, since the buses inside the hub are not connected in any way. Multi-segment hubs are needed to create separable segments, the composition of which can be easily changed. Most multi-segment hubs, such as Nortel Networks' System 5000 or 3Com's PortSwitch Hub, allow the operation of connecting a port to one of the internal buses in a purely software fashion, such as through local configuration through the console port.
As a result, a network administrator can attach user computers to any ports on the hub, and then use the hub's configuration program to control the composition of each segment. If segment 1 becomes overloaded tomorrow, then its computers can be distributed among the remaining segments of the hub.
The ability of a multi-segment hub to programmatically change port connections to internal buses is called configuration switching.
ATTENTION
Configuration switching has nothing to do with the frame switching that bridges and switches perform. Multi-segment hubs are the programmable backbone of large networks. To connect the segments to each other, devices of a different type are needed - bridges / switches or routers. Such a gateway device must connect to multiple ports of a multi-segment hub connected to different internal buses, and transfer frames or packets between segments in the same way as if they were formed by separate hub devices.
For large networks, a multi-segment hub plays the role of an intelligent cross-connect cabinet, which performs a new connection not by mechanically moving the cable plug to a new port, but by programmatically changing the internal configuration of the device. Hub management via SNMP.
As you can see from the description of additional features, many of them require configuration of the hub. This configuration can be done locally via the RS-232C interface available on any hub that has a control unit. In addition to configuring in a large network, the function of monitoring the status of the hub is very useful: whether it is operational, what state its ports are in.
1. Types of computer networks. Types, main components of the LAN.
Types of computer networks:
Computer network (computer network, data transmission network)- a communication system between two or more computers. Various physical phenomena can be used to transmit information, as a rule, various types of electrical signals or electromagnetic radiation. Types of computer networks: Personal network (English Personal Network) is a network built "around" a person. These networks are designed to unite all personal electronic devices of the user (telephones, personal digital assistants, smartphones, laptops, headsets, etc.). The standard for such networks is currently Bluetooth. LAN- serves to combine computers located at a small distance from each other. Such a network usually does not go beyond one room. Urban area network(English MAN - Metropolitan Area Network) covers several buildings within the same city or the city as a whole. Corporate network- a set of LAN, powerful computers and terminal systems that use a common information highway for exchange. National Network- a network that unites computers within one state (National LambdaRail, GEANT) Global Computing Network- a data transmission network designed to serve a significant territory using public communication lines.
Types: By type of functional interaction: Peer-to-peer - the most simple and intended for small slave groups. With their help, users of several computers can use shared drives, printers, and other devices, send messages to each other, and perform other collective operations. Here, any computer can act as both a server and a client. Such a network is cheap and easy to maintain, but cannot provide information protection for large network sizes). Multi-rank (they use dedicated servers for storing shared data and programs for using shared access resources. Such a network has good opportunities for expansion, high performance and reliability, but requires constant qualified maintenance). By type of network topology: Tire, Star, Ring, Lattice. Mixed topology. By network OS: Windows, UNIX, Mixed.
Types, main components of the LAN:
Slave station- computer, intended for local network. A network adapter is a special board that allows the computer to interact with other devices on this network. It provides a physical connection between the m / y network devices via a network cable. Server- some service device, the cat in the LAN acts as a control center and data concentrator. It is a combination of hardware and software that is used to manage shared network resources.
3. Network topology. Network standards (types of networks) Communication medium (network cable).
Network topology(from Greek τόπος, place) - description of the network configuration, layout and connection of network devices.
The network topology can be:
physical- describes the real location and connections between network nodes.
logical- describes the movement of the signal within the framework of the physical topology.
There are many ways to connect network devices, of which five basic topologies can be distinguished: bus, ring, star, mesh topology and lattice. The remaining methods are combinations of the basic ones. In general, such topologies are called mixed or hybrid topologies, but some of them have their own names, such as "Tree".
Ring- the basic topology of a computer network in which workstations are connected in series to each other, forming a closed network. The ring does not use a competitive method of sending data, a computer on the network receives data from a neighbor and redirects them further if they are not addressed to him. To determine to whom data can be transferred, a token is usually used. Data goes around in circles, only in one direction.
Advantages: Easy to install; Almost complete absence of additional equipment; The possibility of stable operation without a significant drop in data transfer rate during heavy network load, since the use of a marker eliminates the possibility of collisions.
Disadvantages: Failure of one workstation, and other problems (cable break), affect the performance of the entire network; Difficulty in configuring and customizing; Difficulty in troubleshooting;
Tire, is a common cable (called a bus or backbone) to which all workstations are connected. There are terminators at the ends of the cable to prevent signal reflection.
A message sent by a workstation propagates to all computers on the network. Each machine checks - to whom the message is addressed, and if it is, then it processes it. In order to exclude the simultaneous sending of data, either a “carrier” signal is used, or one of the computers is the main one and “gives the floor” to the other stations. Advantages Short network setup time; Cheap (requires less cable and network devices); Easy to set up; The failure of a workstation does not affect the operation of the network;
Disadvantages Any problems in the network, such as a cable break, failure of the terminator completely destroy the operation of the entire network; Complex localization of faults; With the addition of new workstations, network performance drops.
Star- the basic topology of a computer network in which all computers in the network are connected to a central node (usually a network hub), forming a physical network segment. Such a network segment can function both separately and as part of a complex network topology (usually a "tree").
The workstation to which data needs to be sent sends them to the hub, which determines the addressee and gives him the information. At a given time, only one machine on the network can send data, if two packets arrive at the hub at the same time, both packets are not received and senders will need to wait a random amount of time to resume data transmission.
Advantages: the failure of one workstation does not affect the operation of the entire network as a whole; good network scalability; easy troubleshooting and breaks in the network; high network performance (subject to proper design); flexible administration options.
Disadvantages failure of the central hub will result in the inoperability of the network (or network segment) as a whole; networking often requires more cable than most other topologies; the finite number of workstations in a network (or network segment) is limited by the number of ports in the central hub.
Mesh topology(in English mesh) - connects each network workstation with all other workstations of the same network. Topology refers to fully connected, unlike others - not fully connected.
The sender of the message connects in turn with the network nodes until it finds the right one, which will receive data packets from it.
Comparison with other topologies
Advantages reliability, if the cable breaks at the computer, there are enough connection paths on the network.
Disadvantages high installation cost; complexity of setup and operation;
In wired networks, this topology is rarely used, because it becomes too expensive due to excessive cable consumption. However, in wireless technologies, networks based on mesh technology are becoming more common, as the cost of network media does not increase and the reliability of the network comes to the fore.
Lattice- a concept from the theory of organization of computer networks. This is a topology in which the nodes form a regular multidimensional lattice. In this case, each edge of the lattice is parallel to its axis and connects two adjacent nodes along this axis. A one-dimensional "lattice" is a chain that connects two external nodes (having only one neighbor) through a certain number of internal nodes (which have two neighbors - left and right). By connecting both external nodes, a "ring" topology is obtained. Two- and three-dimensional lattices are used in the architecture of supercomputers.
Advantages: high reliability. Disadvantages: difficult to implement.
The computer acts as a physical medium for signal transmission
Network cable.Coaxial- comp. from a copper core, insulation, its surrounding, copper braid and outer sheath. May have an additional layer of foil. Thin coax cable - flexible, with a diameter of approximately 0.5 cm, is capable of transmitting signals at a distance of up to 185 m without noticeable distortion. Capable of transmitting data at a rate of 10 Mbps, allows you to implement the bus and ring topology. Thick coax cable - about 1 cm in diameter, the copper core is thicker than that of a thin one. It transmits signals at a distance of 500 m. To connect to it, a special device is used - a transceiver, the cat is equipped with a special connector. twisted pair- two insulated copper wires twisted around each other. Twisting wires allows you to get rid of electrical interference induced by neighboring pairs and other sources. STP (shielded twisted pair) and UTP (unshielded twisted pair) - allows you to transmit a signal up to 100 m. There are 5 categories of UTP: 1) a traditional telephone cable for transmitting analog 2) 4 twisted pair cable capable of transmitting signals at 4Mbps 3) 4 twisted pair cable capable of transmitting signals at 10Mbps 4) 16 Mbps 5) 100-1000 Mbps c (The higher the pair category, the shorter the laying steps). An RJ-45 connector is used to connect the twisted pair to the network. Use in star topology. fiber optic- data is transmitted through optical fibers in the form of modulated light pulses. It is a reliable and secure transmission method, since no electrical signals are transmitted, therefore, the fiber optic cable cannot be opened and data intercepted. Fiber optic lines are designed to move large amounts of data at high speeds. The signal in them practically does not fade and is not distorted. It consists of a thin glass cylinder, called a core, covered with a layer of glass (shell) with a distortion factor different from that of the core. Sometimes optical fiber is made of plastic. Each fiber transmits signals in only one direction, so the cable consists of 2 fibers with separate connectors (transmit and receive). singlemode and multimode– for communication over short distances, because it is easier to install. Optical fiber is used for laying information highways, corporate networks, for data transmission over significant distances. (2 kilometers full duplex over multimode optical fiber and up to 32 kilometers over single mode).
Wireless LAN (WLAN) - wireless local area network. Wi-Fi is one of the options for Wireless LAN. Allows you to deploy a network without laying a cable, can reduce the cost of deploying and expanding the network. Standards 802.11a/b/g speeds from 11 to 53 Mbps. WiMAX is a broadband radio communication protocol (Worldwide Interoperability for Microwave Access) developed by a consortium (English WiMAX Forum). . Unlike WiFi networks (IEEE 802.11x), where access to an access point is granted to clients randomly, in WiMAX, each client is given a clearly regulated period of time. In addition, WiMAX supports mesh topology.
Computer network (CS) - a set of computers and terminals connected via communication channels into a single system that meets the requirements of distributed data processing.
In general, under telecommunications network (TS) understand a system consisting of objects that perform the functions of generating, transforming, storing and consuming a product, called points (nodes) of the network, and transmission lines (communications, communications, connections) that transfer the product between points.
Depending on the type of product - information, energy, mass - information, energy and material networks are distinguished, respectively.
Information network (IS) - a communication network in which the product of generating, processing, storing and using information is information. Traditionally, telephone networks are used to transmit sound information, television is used for images, and telegraph (teletype) is used for text. Nowadays information is becoming more and more widespread. integrated service networks, allowing to transmit sound, image and data in a single communication channel.
Computing network (CN)- information network, which includes computing equipment. The components of a computer network can be computers and peripheral devices that are sources and receivers of data transmitted over the network.
Aircraft are classified according to a number of criteria.
- 1. Depending on the distance between network nodes, aircraft can be divided into three classes:
- · local(LAN, LAN - Local Area Network) - covering a limited area (usually within the remoteness of stations no more than a few tens or hundreds of meters from each other, less often 1 ... 2 km);
- · corporate (enterprise scale)- a set of interconnected LANs covering the territory where one enterprise or institution is located in one or more closely spaced buildings;
- · territorial- covering a significant geographical area; Among territorial networks, one can single out regional networks (MAN - Metropolitan Area Network) and global networks (WAN - Wide Area Network), having, respectively, regional or global scales.
Particularly distinguish the global network of the Internet.
2. An important feature of the classification of computer networks is their topology, which determines the geometric arrangement of the main resources of the computer network and the connections between them.
Depending on the topology of node connections, there are bus (backbone), ring, star, hierarchical, and arbitrary networks.
Among the LANs, the most common are:
- · bus- a local network in which communication between any two stations is established through one common path and the data transmitted by any station simultaneously become available to all other stations connected to the same data transmission medium;
- · ring- the nodes are connected by a ring data transmission line (only two lines go to each node). Data, passing through the ring, alternately becomes available to all network nodes;
- · stellar (star)- there is a central node from which data transmission lines diverge to each of the other nodes.
The topological structure of the network has a significant impact on its bandwidth, network resilience to equipment failures, logical capabilities and network cost.
Computer networks and telecommunications of the XXI century
Introduction
2.1 Types of LAN architectures
2.3 Access methods in computer networks
3. Local networks for scientific purposes
4. Telecommunications
List of used literature
Introduction
A computer network is an association of several computers for the joint solution of information, computing, educational and other problems.
One of the first tasks that arose during the development of computer technology, which required the creation of a network of at least two computers, was to provide many times more reliability than one machine could give at that time when controlling a critical process in real time. Thus, during the launch of a spacecraft, the required rate of reaction to external events exceeds human capabilities, and the failure of the control computer threatens with irreparable consequences. In the simplest scheme, the work of this computer is duplicated by the second one, and if the active machine fails, the contents of its processor and RAM are very quickly transferred to the second one, which picks up control (in real systems, of course, everything happens much more complicated).
Computer networks have given rise to significantly new information processing technologies - network technologies. In the simplest case, network technologies allow the sharing of resources - mass storage devices, printing devices, Internet access, databases and data banks. The most modern and promising approaches to networks are associated with the use of a collective division of labor in joint work with information - the development of various documents and projects, the management of an institution or enterprise, etc.
Computer networks and network technologies for information processing have become the basis for building modern information systems. The computer should now be considered not as a separate processing device, but as a "window" into computer networks, a means of communication with network resources and other network users.
1. Computer network hardware
Local area networks (LAN computers) unite a relatively small number of computers (usually from 10 to 100, although occasionally there are much more) within the same room (training computer class), building or institution (for example, a university). The traditional name - local area network (LAN) - is rather a tribute to those times when networks were mainly used to solve computing problems; Today, in 99% of cases, we are talking exclusively about the exchange of information in the form of texts, graphic and video images, and numerical arrays. The usefulness of drugs is explained by the fact that from 60% to 90% of the information necessary for an institution circulates within it, without needing to go outside.
The creation of automated enterprise management systems (ACS) had a great influence on the development of drugs. ACS includes several automated workstations (AWP), measuring complexes, control points. Another important field of activity in which drugs have proved their effectiveness is the creation of classes of educational computer technology (KUVT).
Due to the relatively short lengths of communication lines (as a rule, no more than 300 meters), information can be transmitted via LAN in digital form at a high transmission rate. At long distances, this method of transmission is unacceptable due to the inevitable attenuation of high-frequency signals, in these cases it is necessary to resort to additional technical (digital-to-analog conversions) and software (error correction protocols, etc.) solutions.
A characteristic feature of the LAN is the presence of a high-speed communication channel connecting all subscribers for transmitting information in digital form. There are wired and wireless channels. Each of them is characterized by certain values of parameters that are essential from the point of view of LAN organization:
1. data transfer rate;
2. maximum line length;
3. noise immunity;
4. mechanical strength;
5. convenience and ease of installation;
6. cost.
Currently, four types of network cables are commonly used:
1. coaxial cable;
2. unprotected twisted pair;
3. protected twisted pair;
4. fiber optic cable.
The first three types of cables transmit an electrical signal over copper conductors. Fiber optic cables transmit light over glass fiber.
Most networks allow multiple cabling options.
Coaxial cables consist of two conductors surrounded by insulating layers. The first layer of insulation surrounds the central copper wire. This layer is braided on the outside with an external shielding conductor. The most common coaxial cables are thick and thin "Ethernet" cables. This design provides good noise immunity and low signal attenuation over distances.
There are thick (about 10 mm in diameter) and thin (about 4 mm) coaxial cables. With advantages in noise immunity, strength, length, a thick coaxial cable is more expensive and more difficult to install (it is more difficult to pull it through cable channels) than a thin one. Until recently, a thin coaxial cable has been a reasonable compromise between the main parameters of LAN communication lines and is most often used to organize large LANs of enterprises and institutions. However, thicker, more expensive cables provide better data transmission over longer distances and are less susceptible to electromagnetic interference.
Twisted pairs are two wires twisted together at six turns per inch to provide EMI shielding and electrical resistance matching. Another name commonly used for such wire is "IBM type-3". In the US, such cables are laid during the construction of buildings to provide telephone communications. However, the use of a telephone wire, especially when it is already placed in a building, can create big problems. First, unprotected twisted pairs are susceptible to electromagnetic interference, such as the electrical noise generated by fluorescent lights and moving elevators. Interference can also be created by signals transmitted in a closed loop in telephone lines running along the LAN cable. In addition, poor quality twisted pairs may have a variable number of turns per inch, which distorts the calculated electrical resistance.
It is also important to note that telephone wires are not always laid in a straight line. A cable connecting two adjacent rooms can actually bypass half of the building. Underestimating the cable length in this case may result in it actually exceeding the maximum allowable length.
Protected twisted pairs are similar to unprotected twisted pairs, except that they use thicker wires and are protected from the external impact of the insulator neck. The most common type of cable used in local area networks, "IBM type-1" is a protected cable with two twisted pairs of continuous wire. In new buildings, type-2 cable may be the best option, as it includes, in addition to the data line, four unprotected pairs of continuous wire for carrying telephone conversations. Thus, "type-2" allows you to use one cable to transmit both telephone conversations and data over a local network.
The protection and tight twists per inch make shielded twisted pair cable a reliable cable connection alternative. However, this reliability comes at a cost.
Fiber optic cables transmit data in the form of pulses of light to glass "wires". Most LAN systems currently support fiber optic cabling. Fiber optic cable has significant advantages over any copper cable options. Fiber optic cables provide the highest transmission speed; they are more reliable, as they are not subject to packet loss due to electromagnetic interference. Optical cable is very thin and flexible, making it easier to transport than heavier copper cable. Most importantly, however, optical cable alone has the bandwidth needed for faster networks in the future.
So far, the price of fiber-optic cable is much higher than copper. Compared to copper cable, the installation of an optical cable is more labor intensive, but its ends must be carefully polished and aligned to ensure a reliable connection. However, now there is a transition to fiber optic lines, which are absolutely not subject to interference and are out of competition in terms of bandwidth. The cost of such lines is steadily decreasing, and the technological difficulties of splicing optical fibers are being successfully overcome.
Wireless communication on radio waves can be used to organize networks within large premises such as hangars or pavilions, where the use of conventional communication lines is difficult or impractical. In addition, wireless lines can connect remote segments of local networks at distances of 3 - 5 km (with a wave channel antenna) and 25 km (with a directional parabolic antenna) under the condition of direct visibility. Organization of a wireless network is significantly more expensive than a conventional one.
For the organization of training LANs, twisted pair is most often used, as the cheapest, since the requirements for data transfer speed and line length are not critical.
Network adapters (or, as they are sometimes called, NICs) are required to connect computers using LAN links. The most famous are: adapters of the following three types:
1. ArcNet; 2. Token Ring; 3. Ethernet.
2. LAN configuration and organization of information exchange
2.1 Types of LAN architectures
In the simplest networks with a small number of computers, they can be completely equal; the network in this case provides data transfer from any computer to any other for collective work on information. Such a network is called peer-to-peer.
However, in large networks with a large number of computers, it turns out to be appropriate to allocate one (or several) powerful computers to serve the needs of the network (storage and transmission of data, printing to a network printer). These dedicated computers are called servers; they run on a network operating system. A high-performance computer with large RAM and a hard drive (or even several hard drives) of high capacity is usually used as a server. The keyboard and display for the network server are not required, since they are used very rarely (for setting up the network OS).
All other computers are called workstations. Workstations may not have hard drives or even drives at all. Such workstations are called diskless. The initial loading of the OS on diskless workstations occurs over a local network using RAM chips specially installed on the network adapters of workstations that store the boot program.
LANs, depending on the purpose and technical solutions, can have different configurations (or, as they say, architecture or topology).
In a ring LAN, information is transmitted over a closed channel. Each subscriber is directly connected to two nearest neighbors, although in principle it is able to communicate with any subscriber in the network.
In the star-shaped (radial) LAN, there is a central control computer in the center, which sequentially communicates with subscribers and connects them with each other.
In a bus configuration, computers are connected to a common channel (bus) through which they can exchange messages.
In a tree-like one, there is a "master" computer, to which computers of the next level are subordinate, and so on.
In addition, configurations without a distinct nature of the connections are possible; the limit is a fully meshed configuration, where every computer on the network is directly connected to every other computer.
In large LANs of enterprises and institutions, bus (neck) topology is most often used, corresponding to the architecture of many administrative buildings with long corridors and employees' offices along them. For training purposes in KUVT, ring and star-shaped drugs are most often used.
In any physical configuration, support for access from one computer to another, the presence or absence of a dedicated computer (in the KUVT it is called "teacher", and the rest - "student"), is performed by a program - a network operating system, which in relation to the OS of individual computers is superstructure. For modern highly developed OS of personal computers, the presence of network capabilities is quite characteristic (for example, OS / 2, WINDOWS 95-98).
2.2 Network communication components
The process of data transmission over the network is determined by six components:
1. source computer;
2. protocol block;
3. transmitter;
4. physical cable network;
5. receiver;
6. destination computer.
The source computer can be a workstation, a file server, a gateway, or any computer connected to the network. The protocol block consists of a chipset and a software driver for the network interface card. The protocol block is responsible for the logic of transmission over the network. The transmitter sends an electrical signal through a physical topology. The receiver recognizes and receives the signal transmitted over the network and sends it to be converted into a protocol block. The data transfer cycle begins with the source computer transferring the initial data to the protocol block. The protocol block organizes the data into a transmission packet containing the corresponding request to the servers, information on processing the request (including, if necessary, the address of the recipient) and the initial data for transmission. The packet is then sent to the transmitter to be converted into a network signal. The packet propagates along the network cable until it reaches the receiver, where it is recoded into data. Here, control passes to the protocol block, which checks the data for failure, sends a "receipt" about the receipt of the packet to the source, reformulates the packets and transfers them to the destination computer.
The exchange of information using computer networks is called computer telecommunications.(CT). It differs from transmission by mail, telegraph, using radio communication in that the processing and creation of information is carried out in the process of transmission. CT makes it possible to create information systems for collective use that exchange information both between several computers, the user and a remote computer, and between users through a computer.
CT is being implemented in local area networks (LAN) at the level of an enterprise, organization, at the regional (territorial) level (corporate, city networks, etc.) and globally at the national and international level.
Computer telecommunications are direct communication lines of computers, various communication systems and communication equipment: telephone, radio, fiber optic and space (satellite). CT makes it possible to quickly exchange information, including the ability to work in real time.
Communication can be established between two stand-alone PCs and with a remote subscriber - another PC or fax (modem connection). For the first type of communication, the software supports file exchange between PCs via cable via serial ports. To support PC modem communication, more complex software is required, but the possibilities of such communication are much higher - voice information and high-speed digital information (ISDN technology) are simultaneously transmitted over the same telephone line.
Computer (computing, information) networks based on CT and mass distribution PCs enable PC users connected to communication lines and having the necessary devices (modem, fax modem, network card) and telecommunications software to send e-mail messages, participate in teleconferences, conduct banking and trading operations , receive information from banks, databases and knowledge bases, etc.
Initially, the CUs had serial, circular(1970s), a star-shaped or backbone structure (topology) of subscriber communications. For example, Xerox's ETHERNET CS had a backbone structure with a bidirectional communication line.
Regional network is formed by linking local CS into a single network of one or another topology. In turn, the union of regional networks gives a global network. The connection of the CS is carried out using special devices, powerful computers or PCs and complex technical systems - telephone networks, satellite and fiber-optic and other communication systems. Identical networks are connected using a bridge - this is the simplest connection. Gateway-based networks communicate when address translation of destinations and data reformatting is required. Communication CS through the repeater implements the accumulation of data.
Communication between the CS and the PC is carried out through dedicated and wireless lines. Offices, hotels, other institutions and private homes are equipped with a LAN to connect to the global network from any room.
Data transmission to the CS is based on two methods- circuit switching and packet switching. Channel switching is carried out for the duration of a communication session (for example, telephone communication). The communication line remains busy for the duration of the message transmission. Data is transmitted in small frames with error checking in each frame. There are message-switched CSs that block not the entire transmission path, as in circuit switching, but only a part between the nearest repeaters.
Circuit switching is used when high reliability, high noise immunity and confidentiality of communications are required (for example, between government agencies, heads of state, in the military sphere, etc.).
In packet switching, messages are divided into packets of a fixed length (128 bytes, etc.), are provided with markers with the sender and recipient addresses and the packet number, and are sent over the network as independent messages. The packets belonging to various messages accumulated in the buffer of the communication node are transmitted to the neighboring communication node. At the destination, the interface processor combines the packets into a single message and delivers them to the destination.
The method of switching packets and transmitting them along different paths improves reliability and reduces message transmission time, providing higher throughput, in particular for short messages, which effectively supports the real-time conversational mode that is popular in today's world.
In the initial period of the creation of the CU (1970s), their differences made it difficult to integrate into global networks. But as a result of the development of the CS, a hierarchical approach to organizing networks has been formed, embodied in the standard open systems communication model (OSI-architecture) of the International Standards Organization (ISO).
The section "Computer telecommunications" is focused on the basic level recommended by the school curriculum, but easily develops into one or two elective courses ("Computer networks", "Website building") with the involvement of additional material and the expansion of the set of workshops and projects. These extensions are contained in the "Getting Online" tutorial referenced above.
