Advantages and disadvantages of FDDI. FDDI topologies

In Russia, the process of intensive introduction of new and modernization of existing local area networks (LAN) continues. Increasing sizes of networks, applied software systems, requiring ever-higher information exchange speeds, increasing requirements for reliability and fault tolerance are forcing us to look for an alternative to traditional Ethernet and Arcnet networks. One type of high-speed network is FDDI (Fiber Distributed Data Interface). The article discusses the possibilities of using FDDI in the construction of corporate computer systems.

According to Peripheral Strategies forecasts worldwide by 1997 to local computer networks more than 90% of all personal computers(currently - 30-40%). Network computer systems are becoming an integral means of production of any organization or enterprise. Quick access to information and its reliability increase the likelihood of making the right decisions by staff and, ultimately, the likelihood of winning in the competition. In their managers and information systems firms see the means of strategic superiority over competitors and consider investment in them as a capital investment.

With computers becoming faster and more efficient in processing and transmitting information, there is a real information explosion. LANs are beginning to merge into geographically distributed networks, the number of servers, workstations and peripheral equipment connected to the LAN is increasing.

Today in Russia, the computer networks of many large enterprises and organizations are one or more LANs built on the basis of Arcnet or Ethernet standards. The network operating environment is typically NetWare v3.11 or v3.12 with one or more file servers. These LANs either have no connection to each other at all, or are connected by a cable operating in one of these standards through internal or external NetWare software routers.

Modern operating systems and application software require the transfer of large amounts of information for their work. At the same time, it is required to ensure the transmission of information at ever higher speeds and over ever greater distances. Therefore, sooner or later the performance Ethernet networks and software bridges and routers are no longer meeting the growing needs of users, and they are beginning to consider the possibility of using faster standards in their networks. One of them is FDDI.

How an FDDI Network Works

The FDDI network is a fiber optic token ring with a data transfer rate of 100 Mbps.

The FDDI standard was developed by the X3T9.5 committee of the American National Standards Institute (ANSI). FDDI networks are supported by all leading network equipment manufacturers. The ANSI X3T9.5 committee has now been renamed X3T12.

The use of fiber optics as a propagation medium can significantly expand the cable bandwidth and increase the distance between network devices.

Let's compare the throughput of FDDI and Ethernet networks with multi-user access. The allowable level of utilization of the Ethernet network lies within 35% (3.5 Mbps) of the maximum throughput (10 Mbps), otherwise the probability of collisions becomes not too high and the cable throughput will drop sharply. For FDDI networks, allowable utilization can reach 90-95% (90-95 Mbps). Thus, the throughput of FDDI is approximately 25 times higher.

deterministic nature FDDI protocol(the ability to predict the maximum delay when transmitting a packet over the network and the ability to provide a guaranteed bandwidth for each of the stations) makes it ideal for use in real-time network control systems and in applications that are critical to the time of information transmission (for example, for the transmission of video and audio information) .

FDDI inherited many of its key properties from Token Ring networks (IEEE 802.5 standard). First of all - this ring topology and a marker media access method. Marker - a special signal rotating around the ring. The station that received the token can transmit its data.

However, FDDI also has a number of fundamental differences from Token Ring, which makes it a faster protocol. For example, the data modulation algorithm at the physical layer has been changed. Token Ring uses a Manchester coding scheme that requires doubling the bandwidth of the transmitted signal relative to the transmitted data. FDDI implements a "five out of four" - 4V / 5V coding algorithm that provides the transmission of four information bits by five transmitted bits. When transmitting 100 Mbps of information per second, 125 Mbps are physically transmitted to the network, instead of 200 Mbps, which would be required when using Manchester coding.

Optimized and medium access control (VAC). In Token Ring it is based on a bit basis, while in FDDI it is based on the parallel processing of a group of four or eight transmitted bits. This reduces the hardware performance requirements.

Physically, the FDDI ring is formed by a fiber optic cable with two light-conducting windows. One of them forms the primary ring (primary ring), is the main one and is used for the circulation of data tokens. The second fiber forms a secondary ring (secondary ring), is a backup and in normal mode not used.

Stations connected to the FDDI network fall into two categories.

Class A stations have physical connections to the primary and secondary rings (Dual Attached Station - doubly connected station);

2. Class AND stations are connected only to the primary ring (Single Attached Station - once connected station) and are connected only through special devices called hubs.

On fig. 1 shows an example of connecting a concentrator and stations of classes A and B in a closed loop, through which the marker circulates. On fig. Figure 2 shows a more complex network topology with a branched structure (Ring-of-Trees - a ring of trees) formed by class B stations.

Ports of network devices connected to the FDDI network are classified into 4 categories: A ports, B ports, M ports and S ports. Port A is the port that receives data from the primary ring and sends it to the secondary ring. Port B is the port that receives data from the secondary ring and sends it to the primary ring. M (Master) and S (Slave) ports transmit and receive data from the same ring. The M port is used on the hub to connect the Single Attached Station via the S port.

The X3T9.5 standard has a number of limitations. The total length of a double fiber optic ring is up to 100 km. Up to 500 class A stations can be connected to the ring. The distance between nodes when using a multimode fiber-optic cable is up to 2 km, and when using a single-mode cable, it is mainly determined by the parameters of the fiber and the transceiver equipment (it can reach 60 or more km).

Fault tolerance of FDDI networks

The ANSI X3T9.5 standard regulates 4 basic fault-tolerant properties of FDDI networks:

1. The ring cable system with class A stations is fault-tolerant to a single cable break anywhere in the ring. On fig. Figure 3 shows an example of both primary and secondary fiber breaks in a ring cable. Stations on either side of the cliff reconfigure the token and data path by connecting a secondary fiber optic ring.

2. A power outage, failure of one of the class B stations, or a broken cable from the hub to that station will be detected by the hub and the station will be disconnected from the ring.

3. Two class B stations are connected to two hubs at once. This special kind connection is called Dual Homing and can be used for fault-tolerant (to faults in the hub or in the cable system) connection of class B stations by duplicating the connection to the main ring. In normal mode, data exchange occurs only through one hub. If for any reason the connection is lost, then the exchange will be carried out through the second hub.

4. A power outage or failure of one of the class A stations will not cause the remaining stations connected to the ring to fail, as the light signal will be passively transmitted to the next station via the optical switch (Optical Bypass Switch). The standard allows up to three sequentially located switched off stations.

Optical switches are manufactured by Molex and AMP.

Synchronous and asynchronous transmission

Connecting to the FDDI network, stations can transmit their data to the ring in two modes - synchronous and asynchronous.

Synchronous mode is arranged as follows. During the initialization of the network, the expected token round trip time is determined - TTRT (Target Token Rotation Time). Each station that captures the token is given a guaranteed time to transmit its data to the ring. After this time, the station must complete the transmission and send the token into the ring.

Each station at the time of sending a new token turns on a timer that measures the time interval until the token returns to it - TRT (Token Rotation Timer). If the token returns to the station before the expected TTRT bypass time, then the station may extend the time it takes to send its data to the ring after the end of the synchronous transmission. This is what asynchronous transmission is based on. The additional time interval for transmission by the station will be equal to the difference between the expected and real time bypassing the ring with a marker.

From the algorithm described above, it can be seen that if one or more stations do not have enough data to fully use the time slot for synchronous transmission, then the bandwidth not used by them immediately becomes available for asynchronous transmission by other stations.

cable system

The FDDI PMD (Physical medium-dependent layer) substandard defines a multimode fiber optic cable with a diameter of 62.5/125 µm as the basic cable system. It is allowed to use cables with a different fiber diameter, for example: 50/125 microns. Wavelength - 1300 nm.

The average power of the optical signal at the station input must be at least -31 dBm. With such an input power, the probability of an error per bit when retransmitting data by the station should not exceed 2.5 * 10 -10 . With an increase in the input signal power by 2 dBm, this probability should decrease to 10 -12 .

The standard defines the maximum allowable signal loss level in the cable as 11 dBm.

The FDDI substandard SMF-PMD (Single-mode fiber Physical medium-dependent layer) defines the requirements for the physical layer when using a single-mode fiber optic cable. In this case, a laser LED is usually used as a transmitting element, and the distance between stations can reach 60 or even 100 km.

FDDI modules for single-mode cable are produced, for example, by Cisco Systems for their Cisco routers 7000 and AGS+. Singlemode and multimode cable segments can be interleaved in an FDDI ring. For these Cisco routers, you can select modules with all four port combinations: multimode-multimode, multimode-singlemode, singlemode-multimode, singlemode-singlemode.

Cabletron Systems Inc. releases Dual Attached repeaters - FDR-4000, which allow you to connect a single-mode cable to a class A station with ports designed to work on a multimode cable. These repeaters make it possible to increase the distance between the nodes of the FDDI ring up to 40 km.

The CDDI (Copper Distributed Data Interface) physical layer substandard defines the requirements for the physical layer when using shielded (IBM Type 1) and unshielded (Category 5) twisted pair. This greatly simplifies the process of installing a cable system and reduces the cost of it, network adapters and hub equipment. Distances between stations when using twisted pairs should not exceed 100 km.

Lannet Data Communications Inc. releases FDDI modules for its hubs, which allow you to work either in standard mode, when the secondary ring is used only for failover purposes in case of a cable break, or in advanced mode, when the secondary ring is also used for data transmission. In the second case, the bandwidth of the cable system is expanded to 200 Mbps.

Connecting equipment to the FDDI network

There are two main ways to connect computers to the FDDI network: directly, and also through bridges or routers to networks of other protocols.

Direct connection

This connection method is used, as a rule, to connect files, archiving and other servers, medium and large computers to the FDDI network, that is, key network components that are the main computing centers that provide services to many users and require high I / O speeds over the network .

Workstations can be connected in the same way. However, since network adapters for FDDI are very expensive, this method is used only in cases where a high network exchange rate is a prerequisite for normal operation applications. Examples of such applications: multimedia systems, video and audio transmission.

To connect personal computers to the FDDI network, specialized network adapters are used, which are usually inserted into one of the free slots on the computer. Such adapters are produced by the following companies: 3Com, IBM, Microdyne, Network Peripherials, SysKonnect, etc. There are cards on the market for all common buses - ISA, EISA and Micro Channel; there are adapters for connecting class A or B stations for all types of cable system - fiber optic, shielded and unshielded twisted pairs.

All leading manufacturers of UNIX machines (DEC, Hewlett-Packard, IBM, Sun Microsystems and others) provide interfaces for direct connection to FDDI networks.

Connecting through bridges and routers

Bridges (bridges) and routers (routers) allow you to connect to FDDI networks of other protocols, such as Token Ring and Ethernet. This makes it possible to cost-effectively connect a large number of workstations and other network equipment to FDDI in both new and existing LANs.

Structurally, bridges and routers are manufactured in two versions - in a finished form, which does not allow further hardware growth or reconfiguration (the so-called standalone devices), and in the form of modular hubs.

Examples of standalone devices are Hewlett-Packard's Router BR and Network Peripherals' EIFO Client/Server Switching Hub.

Modular hubs are used in complex large networks as central network devices. The hub is a housing with a power supply and a communication board. Network communication modules are inserted into the slots of the hub. The modular design of the hubs makes it easy to assemble any LAN configuration, combine cable systems of various types and protocols. The remaining free slots can be used for further expansion of the LAN.

Hubs are manufactured by many companies: 3Com, Cabletron, Chipcom, Cisco, Gandalf, Lannet, Proteon, SMC, SynOptics, Wellfleet and others.

The hub is the central node of the LAN. Its failure can bring the entire network to a halt, or, at least, a significant part of it. Therefore, most hub manufacturers take special measures to improve their fault tolerance. Such measures include redundant power supplies in load sharing or hot standby mode, as well as the ability to change or reinstall modules without turning off the power (hot swap).

In order to reduce the cost of the hub, all its modules are powered by common source nutrition. The power elements of the power supply are the most likely cause of its failure. Therefore, redundant power supply significantly extends the service life. uptime. During installation, each of the hub's power supplies can be connected to a separate uninterruptible power supply (UPS) in case of power failures. Each of the UPS is desirable to connect to the hotel power electrical networks from different substations.

The ability to change or reinstall modules (often including power supplies) without turning off the hub allows you to repair or expand the network without stopping service for those users whose network segments are connected to other hub modules.

FDDI-to-Ethernet bridges

Bridges operate on the first two levels of the interaction model open systems- on the physical and channel - and are designed to connect several LANs of single or different physical layer protocols, for example, Ethernet, Token Ring and FDDI.

According to their principle of operation, bridges are divided into two types (Sourece Routing - source routing) require that the sending node of the packet place information about its routing path in it. In other words, each station must have built-in packet routing capabilities. The second type of bridges (Transparent Bridges - transparent bridges) provide transparent communication between stations located in different LANs, and all routing functions are performed only by the bridges themselves. Below, we will discuss only such bridges.

All bridges can add to the table of addresses (Learn addresses), route and filter packets. Smart bridges can also filter packets based on criteria set through the network management system to improve security or performance.

When a data packet arrives on one of the bridge ports, the bridge must either forward it to the port to which the packet's destination host is connected, or simply filter it out if the destination host is on the same port that the packet came from. Filtering avoids unnecessary traffic on other LAN segments.

Each bridge builds an internal table of physical addresses of nodes connected to the network. The filling process is as follows. Each packet has in its header physical addresses origin and destination nodes. Having received a data packet on one of its ports, the bridge works according to the following algorithm. In the first step, the bridge checks to see if its internal table contains the host address of the packet's sender. If not, then the bridge enters it into a table and associates with it the port number on which the packet arrived. The second step checks to see if the address of the destination node is entered in the internal table. If not, the bridge forwards the received packet to all networks connected to all other ports. If the destination host address is found in the internal table, the bridge checks whether the destination host's LAN is connected to the same port that the packet came from or not. If not, then the bridge filters the packet, and if so, then it transmits it only to the port to which the network segment with the destination host is connected.

Three main parameters of the bridge:
- size of the internal address table;
- filtration speed;
- packet routing rate.

The size of the address table characterizes the maximum number of network devices whose traffic can be routed by the bridge. Typical address table sizes range from 500 to 8000. What happens if the number of connected nodes exceeds the address table size? Since most bridges store in it the network addresses of the hosts that last transmitted their packets, the bridge will gradually "forget" the addresses of the hosts as other transmit packets. This may lead to a decrease in the efficiency of the filtering process, but will not cause fundamental problems in the network.

Packet filtering and routing rates characterize the performance of a bridge. If they are below the maximum possible packet rate on the LAN, then the bridge can cause latency and performance degradation. If it is higher, then the cost of the bridge is higher than the minimum required. Let's calculate what the performance of the bridge should be for connecting several Ethernet protocol LANs to FDDI.

Let us calculate the maximum possible intensity of packets in the Ethernet network. Structure Ethernet packets shown in Table 1. The minimum packet length is 72 bytes or 576 bits. The time required to transmit one bit over an Ethernet LAN at 10 Mbps is 0.1 µs. Then the transmission time of the minimum packet length will be 57.6*10 -6 sec. Ethernet standard requires a pause between packets of 9.6 μs. Then the number of packets transmitted in 1 second will be 1/((57.6+9.6)*10 -6 )=14880 packets per second.

If the bridge connects N Ethernet protocol networks to the FDDI network, then, respectively, its filtering and routing rates should be equal to N * 14880 packets per second.

Table 1.
Packet structure in Ethernet networks.

On the FDDI port side, the packet filtering rate should be much higher. In order for the bridge not to degrade network performance, it should be about 500,000 packets per second.

According to the principle of packet transmission, bridges are divided into Encapsulating Bridges and Translational Bridges. Physical layer packets of one LAN are completely transferred to physical layer packets of another LAN. After passing through the second LAN, another similar bridge removes the shell from the intermediate protocol, and the packet continues its movement in its original form.

Such bridges allow two Ethernet LANs to be connected by an FDDI backbone. However, in this case, FDDI will only be used as a transmission medium, and stations connected to Ethernet networks will not "see" stations directly connected to the FDDI network.

Bridges of the second type convert from one physical layer protocol to another. They remove the header and trailing overhead of one protocol and transfer data to another protocol. Such a conversion has a significant advantage: FDDI can be used not only as a transmission medium, but also for direct connection of network equipment, transparently visible to stations connected to Ethernet networks.

Thus, such bridges provide transparency of all networks over network protocols and more. upper levels(TCP/IP, Novell IPX, ISO CLNS, DECnet Phase IV and Phase V, AppleTalk Phase 1 and Phase 2, Banyan VINES, XNS, etc.).

Another one important characteristic bridge - the presence or absence of support for the Spannig Tree Algorithm (STA) IEEE 802.1D. It is also sometimes referred to as the standard transparent bridges(Transparent Bridging Standard - TBS).

On fig. Figure 1 shows a situation where there are two possible paths between LAN1 and LAN2 - via bridge 1 or via bridge 2. Situations similar to these are called active loops. Active loops can cause serious network problems: duplicate packets violate the logic of the network protocols and lead to a decrease in the throughput of the cable system. The STA ensures that all possible paths are blocked except for one. However, in case of problems with the main communication line, one of the backup paths will immediately be set as active.

Intelligent bridges

So far, we have discussed the properties of arbitrary bridges. Intelligent bridges have a number of additional features.

For large computer networks, one of the key problems that determine their effectiveness is to reduce the cost of operation, early diagnosis possible problems, reducing troubleshooting time.

For this, centralized network management systems are used. They usually work on SNMP protocol(Simple Network Management Protocol) and allow the network administrator from his workplace:
- configure hub ports;
- produce a set of statistics and traffic analysis. For example, for each station connected to the network, you can get information about when it last sent packets to the network, the number of packets and bytes received by each station with a LAN different from the one to which it is connected, the number of broadcasts sent (broadcast) packages, etc.;

Install additional filters on hub ports by LAN numbers or by physical addresses of network devices in order to enhance protection against unauthorized access to network resources or to improve the efficiency of individual LAN segments;
- promptly receive messages about all emerging problems in the network and easily localize them;
- carry out diagnostics of concentrator modules;
- view in graphical form an image of the front panels of modules installed in remote hubs, including the current status of indicators (this is possible due to the fact that the software automatically recognizes which module is installed in each particular slot of the hub, and receives information about the current status of all module ports);
- view system log, which automatically records information about all problems with the network, about the time of turning on and off workstations and servers, and about all other important events for the administrator.

These features are common to all intelligent bridges and routers. Some of them (for example, Gandalf's Prism System) also have the following important advanced features:

1. Protocol priorities. According to separate protocols network layer some hubs act as routers. In this case, the setting of priorities of some protocols over others can be supported. For example, you can set TCP/IP to take precedence over all other protocols. This means that TCP/IP packets will be transmitted first (this is useful in case of insufficient cable system bandwidth).

2. Protection against "broadcast storms"(broadcast storm). One of the characteristic malfunctions of network equipment and errors in software- spontaneous generation with a high intensity of broadcast packets, i.e. packets addressed to all other devices connected to the network. The network address of the destination host of such a packet consists of only ones. Having received such a packet on one of its ports, the bridge must address it to all other ports, including the FDDI port. In normal mode, such packets are used by operating systems for service purposes, for example, to send messages about the appearance of a new server on the network. However, with a high intensity of their generation, they will immediately occupy the entire bandwidth. The bridge provides network congestion protection by including a filter on the port from which such packets are received. The filter does not pass broadcast packets and other LANs, thereby protecting the rest of the network from overload and maintaining its performance.

3. Collection of statistics in the "What if?" This option allows you to virtually install filters on bridge ports. In this mode, filtering is not physically performed, but statistics are collected about packets that would be filtered if the filters were actually enabled. This allows the administrator to pre-evaluate the consequences of enabling the filter, thereby reducing the likelihood of errors when incorrectly established conditions filtering and without causing the connected equipment to malfunction.

FDDI Usage Examples

Here are two of the most typical examples of the possible use of FDDI networks.

Client-server applications. FDDI is used to connect equipment that requires a wide bandwidth from a LAN. Usually this file servers NetWare UNIX machines and mainframes. In addition, as noted above, some workstations that require high data exchange rates can also be connected directly to the FDDI network.

User workstations are connected via multiport FDDI-Ethernet bridges. The bridge performs filtering and transmission of packets not only between FDDI and Ethernet, but also between different Ethernet networks. The data packet will only be transmitted to the port where the destination node is located, saving the bandwidth of other LANs. From the side of Ethernet networks, their interaction is equivalent to communication through the backbone (backbone), only in this case it does not physically exist in the form of a distributed cable system, but is entirely concentrated in a multiport bridge (Collapsed Backbone or Backbone-in-a-box).

Advantages and disadvantages of FDDI. FDDI topologies. How FDDI works. Passing the token to FDDI.

FDDI (Fiber Distributed Data Interface) technology is the first technology local networks, in which the data transmission medium is a fiber optic cable. Work on the creation of technologies and devices for the use of fiber-optic channels in local networks began in the 80s, shortly after the start of industrial operation of such channels in territorial networks. The XZT9.5 problem group of the ANSI Institute developed in the period from 1986 to 1988 initial versions FDDI standard, which provides frame transmission at a rate of 100 Mbps over a double fiber-optic ring up to 100 km long.

Advantages.

1. Reliability.

The double ring configuration provides redundancy.

The system is able to cope with single and multiple cliffs by segmenting sections.

2. Fault tolerance.

Dual Homing: Allows for redundant connectivity to the FDDI network in the tree topology. A DAS station can have a dual connection, for this A and B ports are connected to different hubs. If the main port fails, the backup link is activated.

Optical Bypass: This feature ensures that the light signal passes through during power failures of the DAS station. The data simply bypasses the inactive station by going through the optical bypass.

Global Storage: If both logical rings are operational and the system detects a fault in one of the logical rings, the current data is routed without loss over the backup ring.

3.Built-in control.

Each node has a control object, providing big number services.

Thanks to the presence of an extensive MIB, SNMP management is possible.

disadvantages.

The high price is due to expensive transceivers that convert an electrical signal into an optical one and vice versa. Fiber technology: ~700$/port

UTP: ~450$/port

Topology.

· Physical topology

· Double ring without trees

・Double ring with trees

· Logical topology.

split ring

FDDI technology is largely based on Token Ring technology, developing and improving its main ideas. Developer priorities:

Increase the bit rate of data transfer to 100 Mbps;

Increase the fault tolerance of the network due to standard procedures for restoring it after failures of various kinds - damage to cables, incorrect work node, hub, occurrence high level interference on the line, etc.;

Make the most of potential network bandwidth for both asynchronous and synchronous (delay sensitive) traffic.

The FDDI network is built on the basis of two fiber optic rings, which form the main and backup track data transfer between network nodes.

Having two rings is the primary way to increase resiliency in an FDDI network, and nodes that want to take advantage of this increased reliability potential should be connected to both rings. In the normal mode of the network, data passes through all nodes and all sections of the cable of only the primary (Primary) ring, this mode is called the Thru mode, that is, “through”, or “transit”. The secondary ring (Secondary) is not used in this mode.

In the event of some kind of failure where part of the primary ring is unable to transmit data (for example, a cable break or node failure), the primary ring is merged with the secondary, forming a single ring again. This mode of operation of the network is called Wrap, that is, “folding”, or “folding”, rings. The folding operation is performed by means of concentrators and / or network adapters FDDI. To simplify this procedure, data on the primary ring is always transmitted in one direction and on the secondary - in the opposite direction. Therefore, when a common ring is formed from two rings, the transmitters of the stations still remain connected to the receivers of neighboring stations, which makes it possible to correctly transmit and receive information by neighboring stations.

FDDI (Fiber Distributed Data Interface) technology is largely based on Token Ring technology, developing and improving its main ideas. The developers of FDDI technology set themselves the following goals as the highest priority:

· Increase the bit rate of data transfer up to 100 Mb/s;
· Increase network fault tolerance due to standard procedures for restoring it after failures of various kinds - cable damage, incorrect operation of a node, hub, high level of noise on the line, etc.;
· Make the most of potential network bandwidth for both asynchronous and synchronous traffic.

The FDDI network is built on the basis of two fiber optic rings, which form the main and backup data transmission paths between network nodes. The use of two rings is the main way to increase fault tolerance in an FDDI network, and nodes that want to use it must be connected to both rings. In the normal mode of network operation, data passes through all nodes and all sections of the cable of the primary (Primary) ring, so this mode is called Thru mode - "through" or "transit". The secondary ring (Secondary) is not used in this mode.

In the event of some type of failure where part of the primary ring is unable to transmit data (for example, a cable break or node failure), the primary ring is combined with the secondary (figure 2.1), forming a single ring again. This mode of network operation is called Wrap, that is, "folding" or "folding" rings. The folding operation is performed by hubs and/or FDDI network adapters. To simplify this procedure, data on the primary ring is always transmitted counterclockwise, and on the secondary - clockwise. Therefore, when a common ring is formed from two rings, the transmitters of the stations still remain connected to the receivers of neighboring stations, which makes it possible to correctly transmit and receive information by neighboring stations.

The FDDI standards place a lot of emphasis on various procedures to determine if a network has failed and then reconfigure as needed. The FDDI network can fully restore its operability in the event of single failures of its elements. With multiple failures, the network breaks up into several unrelated networks.

Rings in FDDI networks are considered as a common shared data transmission medium, so a special access method is defined for it. This method is very close to the access method of Token Ring networks and is also called the token (or token) ring method - token ring (Figure 2.2, a).

The station can start transmitting its own data frames only if it has received a special frame from the previous station - an access token (Figure 2.2, b). After that, she can transfer her frames, if she has them, for a time called Token Holding Time (THT). After the expiration of the THT time, the station must complete the transmission of its next frame and pass the access token to the next station. If, at the time of accepting the token, the station does not have frames to transmit over the network, then it immediately broadcasts the token of the next station. In an FDDI network, each station has an upstream neighbor and a downstream neighbor determined by its physical links and direction of information transfer.

Each station in the network constantly receives the frames transmitted to it by the previous neighbor and analyzes their destination address. If the destination address does not match its own, then it broadcasts the frame to its subsequent neighbor. This case is shown in the figure (Figure 2.2, c). It should be noted that if the station has captured the token and transmits its own frames, then during this period of time it does not broadcast incoming frames, but removes them from the network.

If the frame address matches the address of the station, then it copies the frame to its internal buffer, checks its correctness (mainly by checksum), passes its data field for further processing to the protocol of the layer above FDDI (for example, IP), and then transmits the original frame over the network of the subsequent station In the frame transmitted to the network, the destination station notes three signs: address recognition, frame copying, and the absence or presence of there are errors in it.

After that, the frame continues to travel through the network, being broadcast by each node. The station, which is the source of the frame for the network, is responsible for removing the frame from the network after it, having made a full turn, reaches it again (Figure 2.2, e). In this case, the source station checks the signs of the frame, whether it reached the destination station and whether it was damaged. The process of restoring information frames is not the responsibility of the FDDI protocol, this should be handled by higher layer protocols.

Distributed Data Interface is the first LAN technology in which the data transmission medium is fiber optic cable. Main characteristics of technology

FDDI technology is largely based on Token Ring technology, developing and improving its main ideas. The developers of FDDI technology set themselves the following goals as the highest priority:

Increase the bit rate of data transfer to 100 Mbps;

Increase the fault tolerance of the network due to standard procedures for restoring it after failures of various kinds - cable damage, incorrect operation of the node, hub, the occurrence of a high level of interference on the line, etc.;

Make the most of potential network bandwidth for both asynchronous and synchronous (delay sensitive) traffic.

The FDDI network is built on the basis of two fiber optic rings, which form the main and backup data transmission paths between network nodes. Having two rings is the primary way to increase resiliency in an FDDI network, and nodes that want to take advantage of this increased reliability potential should be connected to both rings. In the event of some kind of failure where part of the primary ring is unable to transmit data (for example, a cable break or node failure), the primary ring is merged with the secondary, forming a single ring again. This network mode is called Wrap, i.e. "folding" or "folding" the rings. The folding operation is performed by means of hubs and/or FDDI network adapters. To simplify this procedure, data on the primary ring is always transmitted in one direction (in the diagrams, this direction is shown counterclockwise), and on the secondary - in the opposite direction (shown clockwise)

The topology of the FDDI network is a ring, and two multi-directional fiber optic cables are used, which in principle allows the use of full-duplex information transmission at twice the effective speed of 200 Mbps (while each of the two channels operates at a speed of 100 Mbps). A star-ring topology is also used with hubs included in the ring. The formats of the token (Figure 5.15) and packet (Figure 5.16) of the FDDI network are somewhat different from the formats used in the Token-Ring network. The purpose of the fields is as follows.

The preamble is used for synchronization. Initially, it contains 64 bits, but the subscribers through which the packet passes can change its size.

The initial delimiter performs the function of a sign of the beginning of the frame.

Rice. 5.15. FDDI marker format

Destination and source addresses can be 6 bytes (similar to Ethernet and Token-Ring) or 2 bytes.

The data field may be of variable length, but the total packet length must not exceed 4500 bytes.

The checksum field contains the 32-bit cyclic checksum of the packet.

The trailing delimiter defines the end of the frame.

The packet status byte includes an error detection bit, an address recognition bit, and a copy bit (all similar to Token-Ring).

Rice. 5.16. FDDI Packet Format

The format of the FDDI network control byte is as follows (Fig. 5.17):

The class bit of the packet determines whether the packet is synchronous or asynchronous.

The address length bit determines which address (6-byte or 2-byte) is used in this packet.

The frame format field determines whether the frame is a control frame or an information frame.

The frame type field defines what type the given frame belongs to.

Rice. 5.17. Control byte format

In conclusion, we note that despite the obvious advantages of FDDI, this network has not yet become widespread, which is mainly due to high cost its equipment (about a thousand dollars). The main area of application of FDDI now is the basic, backbone (Backbone) networks that combine several networks. FDDI is also used to connect powerful workstations or servers that require high-speed exchange. It is assumed that the Fast Ethernet network can supplant FDDI, however, the advantages of fiber optic cable, marker control method and record allowable size networks currently put FDDI out of the competition. And in cases where hardware cost is critical, a twisted-pair version of FDDI (TPDDI) can be used in non-critical areas. In addition, the cost of FDDI hardware can greatly decrease with an increase in production volume.

21.Network devices: repeater, hub, bridge, switch, router, gateway.Repeaters

Repeater (repeater) - a hardware device that operates at the physical level reference model OSI and provides a connection between two segments of the same computer network.

Repeaters implement one of the simplest forms of internetworking. They simply regenerate, or repeat, data packets between cable segments. Essentially, repeaters physically expand the network. In addition, they provide a high level of fault tolerance by electrically isolating networks, so that a problem that has arisen in one cable segment does not affect other segments. However, together with the packets, they also repeat the interfering signal, without distinguishing between it and the data packets.

Technology Fiber Distributed Data Interface- the first local area network technology that used fiber optic cable as a data transmission medium.

Attempts to use light as a medium carrying information have been made for a long time - back in 1880, Alexander Bell patented a device that transmitted speech over a distance of up to 200 meters using a mirror that vibrated synchronously with sound waves and modulated the reflected light.

Work on the use of light to transmit information intensified in the 1960s in connection with the invention of the laser, which could modulate light at very high frequencies, that is, create a broadband channel for transmitting a large amount of information at high speed. Around the same time, optical fibers appeared that could transmit light in cable systems, similar to the way copper wires transmit electrical signals in traditional cables. However, the light loss in these fibers was too great to be used as an alternative to copper strands. Inexpensive optical fibers providing low light signal power loss and wide bandwidth (up to several GHz) appeared only in the 1970s. In the early 1980s, the industrial installation and operation of fiber optic communication channels for territorial telecommunication systems began.

In the 1980s, work also began on the creation of standard technologies and devices for using fiber optic channels in local networks. Work on the generalization of experience and the development of the first fiber optic standard for local networks were concentrated at the American National Standards Institute - ANSI, within the framework of the X3T9.5 committee created for this purpose.

The initial versions of the various components of the FDDI standard were developed by the X3T9.5 committee in 1986 - 1988, and at the same time the first equipment appeared - network adapters, hubs, bridges and routers that support this standard.

Currently the majority network technologies support fiber optic cables as one of the physical layer options, but FDDI remains the most mature high speed technology, the standards for which have passed the test of time and are well established, so that the equipment various manufacturers shows a good degree of compatibility

Fundamentals of FDDI technology

FDDI technology is largely based on Token Ring technology, developing and improving its main ideas. The developers of FDDI technology set themselves the following goals as the highest priority:

Increase the bit rate of data transfer up to 100 Mb/s;
Increase the fault tolerance of the network due to standard procedures for restoring it after failures of various kinds - cable damage, incorrect operation of the node, hub, the occurrence of a high level of interference on the line, etc.;
Maximize potential network bandwidth for both asynchronous and synchronous traffic.

The FDDI network is built on the basis of two fiber optic rings, which form the main and backup data transmission paths between network nodes. The use of two rings is the main way to increase fault tolerance in an FDDI network, and nodes that want to use it must be connected to both rings. In the normal mode of network operation, data passes through all nodes and all sections of the cable of the Primary ring, therefore this mode is called the mode Thru- "through" or "transit". The secondary ring (Secondary) is not used in this mode.

In the event of some type of failure where part of the primary ring is unable to transmit data (for example, a cable break or node failure), the primary ring is combined with the secondary (figure 2.1), forming a single ring again. This network mode is called Wrap, i.e. "folding" or "folding" rings. The folding operation is performed by hubs and/or FDDI network adapters. To simplify this procedure, data on the primary ring is always transmitted counterclockwise, and on the secondary - clockwise. Therefore, when a common ring is formed from two rings, the transmitters of the stations still remain connected to the receivers of neighboring stations, which makes it possible to correctly transmit and receive information by neighboring stations.

Rice. 2.1. Reconfiguration of FDDI Rings on Failure

The station can start transmitting its own data frames only if it has received a special frame from the previous station - an access token (Figure 2.2, b). After that, she can transmit her frames, if she has them, for a time called the token hold time - Token Holding Time (THT). After the expiration of the THT time, the station must complete the transmission of its next frame and pass the access token to the next station. If, at the time of accepting the token, the station does not have frames to transmit over the network, then it immediately broadcasts the token of the next station. In an FDDI network, each station has an upstream neighbor and a downstream neighbor determined by its physical links and direction of information transfer.

Rice. 2.2. Frame processing by FDDI ring stations

If the frame address matches the address of the station, then it copies the frame to its internal buffer, checks its correctness (mainly by checksum), passes its data field for further processing to the protocol of the level above FDDI (for example, IP), and then transmits the original frame over the network of the subsequent station (Figure 2.2, d). In the frame transmitted to the network, the destination station notes three signs: address recognition, frame copying, and the absence or presence of errors in it.

Figure 2.3 shows the protocol structure of FDDI technology in comparison with the seven-layer OSI model. FDDI defines a physical layer protocol and a media access sublayer (MAC) protocol. link layer. Like many other LAN technologies, FDDI technology uses the 802.2 Data Link Control (LLC) protocol defined in the IEEE 802.2 and ISO 8802.2 standards. FDDI uses the first type of LLC procedures, in which nodes operate in datagram mode - connectionless and without recovering lost or corrupted frames.

Rice. 2.3. Structure of FDDI technology protocols

The physical layer is divided into two sublayers: the media-independent sublayer PHY (Physical), and environment dependent sublayer PMD (Physical Media Dependent). The operation of all levels is controlled by the station control protocol SMT (Station Management).

PMD level provides necessary funds to transfer data from one station to another via fiber optics. Its specification defines:

Optical power requirements and multimode fiber optic cable 62.5/125 µm;
Requirements for optical bypass switches and optical transceivers;
Parameters of optical connectors MIC (Media Interface Connector), their marking;
The wavelength of 1300 nanometers on which the transceivers operate;
Representation of signals in optical fibers according to the NRZI method.

The TP-PMD specification defines the possibility of transmitting data between stations over twisted pair in accordance with the MLT-3 method. The PMD and TP-PMD layer specifications have already been discussed in the Fast Ethernet sections.

PHY level performs encoding and decoding of data circulating between the MAC layer and the PMD layer, and also provides timing of information signals. Its specification defines:

encoding information in accordance with the scheme 4B/5B;
signal timing rules;
requirements for the stability of the clock frequency of 125 MHz;
rules for converting information from parallel to serial form.

MAC level is responsible for network access control, as well as for receiving and processing data frames. It defines the following parameters.