DE3508417A1 - Method for optically transmitting bit-serial message information, particularly of voice and video signals, via networks configured in the form of a ring - Google Patents

Abstract

The invention relates to a method for optically transmitting bit-serial message information, particularly of voice and video signals, via local area networks equipped with coupling devices and configured in the form of a ring, to which a number of stations processing different types of communication and adapted to the network structure of these local area networks can be connected in arbitrary combination with one another, each station exhibiting transmitting and receiving devices for transmitting the message information and the order of access of the individual stations to the local area network being controllable by a synchronisation clock and the transmitting and receiving device of the stations being connectable to a passive star coupler, constructed as transmissive mixer, in such a manner that the length of the transmitting line is equal to the length of the receiving line of the respective station. Within the network configured in the form of a ring, the synchronisation clock is generated by a source address information formed by each station 1 to N, which, in the case of a failure of one or a number of stations 1 to N is in each case transmitted updated by means of a source address correction device. <IMAGE>

Description

Method for the optical transmission of bit-serial

News information, in particular voice and video signals, via ring-shaped configured networks The invention relates to a method for optical transmission of bit-serial message information, in particular of Voice and video signals via ring-shaped configured coupling devices equipped local networks, one to the transmission speed and adapted to the network structure of these local networks number of different Communication types (voice, data, image, text) processing stations (end devices) sina can be connected to each other in any combination, with each station transmitting and receiving means for transmitting the message information and the stations are arranged within the local area networks that the Message information between the stations can be exchanged as desired, with the order in which the individual stations access the local network a synchronization clock can be controlled, and the transmitting and receiving devices of the stations to a passive star coupler designed as a straight-through mixer can be switched on so that the length of the transmission line is equal to the length of the reception line corresponds to the respective station.

The essential requirements for such a defined system with a view to efficient transmission of real-time news information over local area networks (LAN) are a high utilization of the network capacity and a limited network delay. Fiber optic topologies Networks, which meet these requirements are currently bus and ring network structures, use the synchronous or quasi-synchronous access protocols. So is for example from IEEE Journal on Selected Areas in Communications, Vol. Sac-1, April 1983, pp. 493 - 499 a so-called "D network" is known, in which the transmission sides of all Stations combined and connected to a common input of a star coupler are, while the receiving sides of the stations are fed separately to the star coupler are. In this network, a so-called locomotive generator is provided, the Send and receive in the same way as the individual stations with the Star coupler is connected. This generator is essentially used for synchronization the access of the individual stations to the network and then always sends a joint Network access signal to be received by all stations, if from the receiver the generator recognizes the end of a message information sequence. As soon multiple stations want to send message information at the same time Although all stations receive the network access signal from the generator, only one can single station transmit its own message information over the network, since all other subsequent stations are affected by the ongoing transmission of the Message information automatically locks itself on the sending side.

This star-coupled D network, in which the stations passively with are connected to the network, physically represents an open T-Bus network, in which the reduction in the network access signal caused by the signal distribution only allows a maximum of six stations.

To increase the number of stations within this D network, an active network in the form of an open ring is known from the same report. The stations are connected to the D network in such a way that the receiving sides of the generator and the stations with a common return line are connected, while the transmitter side of the generator is connected to another receiving input the first station, the transmission side of the first station with a further reception input of the second station etc. and finally the transmission side of the nth station with the common return line is connected. This means that both the network access signal as well as the message information itself sent from station to station, there must be received and sent again.

Both systems described - cas passive, open T-bus network and the active, open Ring-D network - are, however, in terms of operational safety severely restricted because, for example, if the central generator fails, the Message information can no longer be transmitted between the stations. In addition, failures of individual active ones already result in the active D network Stations that can arise, for example, from a cable break, for interruption the entire transmission of message information within such a network.

Furthermore, from the Siemens research and development report, Volume 10 (1981), No. 1, published by Springer Verlag 1981, a comparative consideration known between T-bus and star-bus structured networks, from which it can be seen that the number of stations within such networks assuming the transfer of the same amount of data at the same transfer rate through a star bus configuration can be enlarged considerably.

This is essentially due to the different network delays within these various networks; because while in the T-bus network the stations are connected one after the other in series, the stations in the star bus network are through the star coupler is practically connected in parallel.

This means that the network delay in the T-bus network is proportional the number of stations increases, during the network delays in the star coupling network only proportional to the logarithm of the number of stations increases.

The object on which the invention is based is to provide a method for the networks defined at the beginning, in which the number of stations does not primarily depend on the signal distribution losses of the system and the operational reliability such local networks, especially with regard to the failure of individual Stations, is vastly improved. According to the invention, this is achieved through the combination of the characteristic features 1.1 to 1.3 achieved.

As essential for the invention is to be seen that with the through the original address information of the respective stations formed the synchronization clock, which all stations in the local network always receive at the same time, accordingly evaluate and update forwarding, the bus access to the network synchronous and takes place without collision. Since the authorization of access is not from central facilities controlled but formed by the original addresses of the individual stations themselves is, the operational reliability of this transmission system is compared to the known Systems significantly improved. The original address correction device ensures also that in the event of a fault, defective stations are practically bridged without that the operability of the remaining, non-defective stations within the local network can be influenced. The authorization to access the local In this case, the network is converted into the reached the original address corresponding to the defective station, which corrected Source address then from the next functional station as access authorization is recognized. This station then sends its own original address information possibly with a message packet and thus switches the access authorization Further.

Only through the defined network delays within the passive, Star-configured networks, it is possible with a larger number of Stations reliably transmit news information in real-time.

The invention is explained in more detail by several figures, the Figure 1 shows a passive, open T-bus network, Figures 2 and 3 the application according to the application Explain the method in more detail and FIG. 4 shows the network structure of several, coupled represents local networks.

In the figure 1 a passive, open T-bus network is shown, in which the stations 1 to N with their separate receiving lines R to corresponding Connections of the star coupler SK are connected, while the transmission lines T and the receiving lines S for monitoring the ring line and evaluating the synchronization signal combined into a common optical transmission path and used as an input for the star coupler SK is used. There is also a so-called locomotive generator LG shown, with its receive and transmit line in the same way as stations 1 to N are connected to the star coupler SK. The messaging between stations 1 to N takes place in such a way that the generator LG has its Transmission line always sends out an access authorization for the network when the receiving line of the generator LG no longer receives a signal. The access authorization is recognized for example by station 1 via the receiving line S and the to Transmission of upcoming message information via the transmission line T is sent. The station N, which has also received the access authorization signal, provides with the corresponding receiving line that there is already a message information transmission on the network, namely, the message information content from the station 1 is transmitted. the Station N waits to send the message information until the message information station 1 is transmitted via the star coupler SK, where the End of this message information transmission by the receiving line of the station N is recognized. Unless further news information from others, no stations shown are transmitted over the network, the station begins N in turn with the sending of the message information in the same way via the star coupler SK not only to the actual receiving station, but is transmitted simultaneously to all stations 1 to N and to the generator LG. Depending on the status of the individual stations 1 to N, this process is repeated continuously, the generator LG giving the individual stations 1 to N access to the network with controls the synchronization clock.

In the figure 2, the network configuration is shown that the the process according to the application is based. Here are the individual stations 1 to N with their transmission lines S1 to SN and receiving lines El to EN like this connected to the passive star coupler PSK that from each station 1 to N the sent Receive message information from every other station 1 to N in a targeted manner can be. The passive star coupler PSK is equipped as a straight-through mixer, in which each input signal on all receiving lines El to EN of the stations 1 to N is evenly distributed.

FIG. 3 shows a timing diagram in which the transmission of the message information from stations 1 to N via the network in the appropriate time sequence is shown. For the individual stations 1, 2 to N there is in each case the Data packet delay Tp and the station locking time for the end of the data packet td. Furthermore, the time segments T ... - (1 to N) represent the respective cycle time the message information of the individual stations 1 to N, which is from the respective Send line via the passive star coupler to the corresponding receive line of this Stations 1 to N run.

FIG. 4 shows the structure of several via so-called gateways coupled local area networks, in which the message information from the individual stations S ...

of the first local network LAN 1 depending on the connection request to the local Networks LAN 2 or LAN 3 and the stations S ... located there.

The method according to the application describes a synchronous and therefore collision-free access method which enables the transmission of message information, in particular voice information, via a fiber-optic, passive star bus. This procedure is not only based on a firmly defined network delay, but also on high network efficiency. To clarify the synchronous access, it is assumed that N stations are connected to the network according to FIG. 2 and are connected to one another via the passive star coupler PSK. Furthermore, it is a prerequisite that each message information including the transmission within the information packet has at least introductory bits, source address bits, destination address bits, data bits and possibly check bits for testing certain error states of the data packets. For the sake of simplicity, it is also assumed that the control of the bus, which is passed on from station to station by evaluating the original address bits, takes place in the order T1, T2 ..., TN-1, TN, T1 ... It is also assumed that station 1 has received authorization to access the bus and that the sender of station 1 transmits the message information as an information packet over the bus. Since all stations 1 to N are linked to the network via the passive star oppler, the data packet is received by all stations 1 to N - including the transmission stations - so that each station 1 to N can check the original address bits. This operation continues as long as message information is transmitted over the network. Thus, all stations 1 to N know the origin address of the packet information before the rest of the message information is received. In this way, for example, station 2 can recognize, before it has stopped receiving the data packet from station 1, that it will be the next station that will have access to and control of the bus. The other stations must postpone their data transmission accordingly. Station 2 now expects the end of the data packet sent by station 1 with its receiving line. As soon as the end of the data packet is recognized after td seconds after arriving at the receiver of station 2, the bus is out of order and station 2 can start transmitting its packet information via the bus. After a running time (propagation delay) of the beginning of the packet information from station 2 arrives at the receiver of station 3. It follows that the transit time Ti (i = 1, 2, 3 ... N) is the cycle time of the message information for the corresponding station i and thus corresponds to the transmission time between the transmitter and the receiver of station i. As soon as the original address bits have been received by station 2, station 3 recognizes that it is the next station to get access to the bus. This process described is repeated indefinitely.

Should a station have control and therefore access to the local Network does not contain any message information, the network function will by sending out so-called empty packet information (frame bits without data bits) maintained.

So soon a transmission error within the original address bits present, or a station fails, or the optical conductor to the station is defective the self-blocking of the network is avoided as follows. After, for example Station 1 has received the end of its own data packet, the station monitors 1 the bus for a period of time, for example tl. Tl is greater than that greatest Circulation time of stations 1 to N. If the network is working properly, it should During this time, the data packet from station 2 is received by the receiver at station 1 will. If this is not the case, the following occurs. Station 1 poses first a transmission error in the original address bits by sending out another Package stuck on the bus. If there is no further package in station 1, an empty packet is sent out. This procedure is provided as the probability of transmission errors within the original address bits of two consecutive following packages is very low. The probability that a packet will have a transmission error in the original address bits is obtained by multiplying the Quantities Ps x Pe, where Ps is the number of source address bits and Pe is the error rate of the transmission system. The probability of error for two consecutive Packets assuming statistically independent events results in (Ps x Pe). Assuming that Ps = 32 and Pe = 10 ~ 9, the probability of error is approx. 10 -15 or with a packet length of 1 ms on average one error in 31,000 years.

Should again not receive a packet in the following time period t1 after the station 1 has received the end of the second data packet, so it can be assumed that between the star coupler and station 2 there is a break in the optical There is a transmission link or station 2 is defective. Station 1 tried hard the bus control - access to the network - to station 3 Transfer bypassing the faulty station 2. Station 1 will then forwarded an empty package on the bus. This packet contains the original address bits those that correspond to station 2. Since the original address bits sent out by station 1 if they do not correspond to those of station 1, this packet cannot contain any further information bits and is therefore sent out as an empty packet. This makes it possible that the Station 3, though the station 2 has failed, access to the Bus receives.

The next time station 1 gets access to the network, If the station 1 sends out the present packet again, the time period tl waits after receiving the end of this packet and immediately bypasses the station 2 by sending another empty packet, which in turn contains the original address bits the station 2 contains. With this check is the faulty state of the station 2 established and bridged.

If station 3 is also separated from the bus, it will Procedure by station 1 including checking for the original address transmission errors Repeatedly and bridged the faulty station 3 in the same way. To this In this way, faulty stations can be switched off one after the other without this entire network becomes inoperable.

This routine cannot be used in active T-bus networks, neither in the well-known D network, since here the stations are one behind the other switched and, in the event of an error, all subsequent stations from the transmission station are separated.

This is followed by the efficiency and limit values of this passive star bus network in terms of network delay calculated assuming no stations fail due to errors. The parameters are in relation to a fully charged bus defined, d. This means that every station connected to the network always receives a packet that must be transmitted over the bus. The state when interrupted Transmission paths or faulty stations are then checked.

When the bus is fully charged, the passive star coupler, the common The information packages are in function for all connected stations see the time, as shown in FIG. In Figure 2 is the value Tp is the packet transmission time, which is defined as Tp = po, with C the transmission data rate from the bus in where C is the bits / s and P is the number of bits in the data packet.

For the network structured in Figure 2 is the network efficiency defined by the ratio of the sum of the transmission time of N consecutive Packets at the total time required to transmit the N packets and results accordingly to E = N x Tp / N Tp + td) + T (1); where T = T1 + T2 ... + TN the entire Network transit time and td is the time it takes for a station to finish of a data packet after it has received the T own receiver has reached. For Tp << td + N we get E = 1. Since Tp = C, it follows the efficiency approximately equal to the delay for a successful station, provided that the value is P >> C (td + T) (2). The network delay of a packet is defined as the period of time when the packet is the beginning of the from the station sent message and then successfully transmitted via the bus will.

For any existing system that uses the described access protocol is required, there is an upper limit for the network delay. Provided, that the bus is fully charged and the packet to be sent from the authorized station has reached the beginning, this authorized station starts the transmission of the corresponding package via the bus. For the network structured in FIG. 2 the upper limit of the delay or the bond is given by: D = N (Tp + td) + T (3).

The condition of a broken cable or a failed station reduces this effective number of stations corresponding. this is directly related to the operation related to the bridging of the failed or the stations adjacent to the cable break together. Reducing the number of stations is therefore a function of the network transmission rate and the magnitude the message information to be transmitted tl. The greater tl and the higher the bit transmission rate, the greater the loss of the effective number of stations within the network.

In the following invoices, the malfunction remains - disturbed Transmission paths or failed stations - not taken into account.

Equation (3) is used to calculate the number of voice channels and the resulting number of stations that a network can serve. If the corresponding values are rearranged in equation (3), the order of magnitude of is, the number of stations is obtained as N = C (0 - T) P1 + P2 + C x td (4) 'where P1 + P2 = P (5). Here, the value P1 includes all higher-order bits, such as introduction, address and check bits, while P2 represents the number of information bits in the message information. The value of P1 should be kept as small as possible and is usually between 100 and 200. In equation (4), the dependence of T on N can be neglected if the value D> T, or within equation (3) Value P2 »CT is. This is always the case when the transmission time of the information part of the packet is very much greater than the average of the station throughput time. This reduces equation (4) to: N (7). CD Pl + P2 + C x td For a given bus transmission rate of C and the detection time for the end of the data packet by station td, equation (7) shows that there the number of stations N is given by optimizing the values of D and P2 .

Each analog speech signal is transmitted through a vocoder at a bit rate Digitized by Cv. The output bits are grouped in a packet that is then sent via the local network is transmitted to a receive vocoder.

To get an interactive, i. i.e., a mutually consecutive To achieve voice transmission, the end-to-end delay for each information bit must be that is, the time from the generation of the bit by the original vocoder to the time the receipt of this bit by the receive vocoder. With these Three types of delays can be detected in the transmission system: the Packet formation delay, the network delay, and the signal propagation delay between the originating and receiving vocoders. The sum of the delays must not exceed the maximum value of the delay requirement of the transmitted voice signal.

If D vmax is the maximum allowable delay for each bit of the digitized speech signal, tpl the longest signal propagation delay between two vocoders each (originating and receiving vocoders of a station) and TF the packet form time, i.e. the packet integration time for a P2 vocoder, then Tf = Cv (8).

An acceptable voice quality is achieved when the package is straight shaped is transmitted over the bus of the local area network before the shaping of the next package is reached. If a packet is formed every Tf seconds, then the network delay must not exceed the value of Tf. This is guaranteed if the upper limit of the network delay D does not reach the value Tf. Therefore must be Df 7f (9). According to the above, the sum may be all Delays - packet formation delay, network delay, and propagation delay between the originating and receiving stations - does not reach the value of Dvmax. From this it follows: Tf + Dc Dv (10a) where Dv = Dv max - tpl (1Db).

The delay from the formation of the bit by the originating vocoder It must therefore be tied up until it is transmitted over the bus by the originating station be.

The maximum number of voice channels, and thus the number of stations, can be determined by maximizing equation (7) taking into account P2 and D (P1, C and td = const.) And the binding conditions given by equation (9) and ( 10a), can be obtained. From equation (8) to (10a) the permissible values of P2 and D are given by: It is important that in these equations the permissible evaluations of P2 and D are taken into account as separate variables and not as permissible combinations of P2 and D which follow from the inequality of equations (9) and (10a). If the value of P2 assumes the extreme value of 0 and (Dv x Cv), it follows from equations (7) to (10a) that the value N is equal to zero in each case. However, since the value N 0 is defined, there is a global maximum value of N and accordingly also optimal values of P2 and D. From equation (7), the value N is maximized for each fixed value of P2 if the value D corresponds to Condition of equation (llb) is maximized. The maximum permissible value of D for each fixed value of P2 can therefore be obtained from equations (8) to (10a).

There D max = P2 D max = Dv - P2 Substituting D max for D in equation (7), two values for N are obtained as a function of P2. Here each term corresponds to one of the two possible values of P2 given by equations (12a) and (12b). Examination of these equations shows that steadily increasing and for steadily decreases. It follows that the value for N is maximized when: P2 = P2 opt. = Dv x Cv (13a) 2 and D = D'max = Tf = P2 opt. = Dv Cv (13b).

With equation (7), the maximum number of voice channels is, and thus the maximum number of stations that this passive star coupler of the local Network, given by: N max = Dv x Cv (14), x max = Dv x Cv + 2 (P1 + C x td) provided that (Equation 13 inserted into Equation 6) 2T x C (15) is.

Dv 2 N max x Cv Furthermore, N max is maximized as a function of P1 and td when P1 = td = 0, where N * max N max C 0 = C Pl * max # N max pl = 0 Cv (16) td = 0 is.

It can be seen that with the value Pl = td = 0 the value of N * max does not represent a practical situation, but has to be assumed as a representative of the ideal case. Rather, the ratio of the values of N max to N * max is given by inserting equation (3) into equation (1), where the values of D = D'max and are assumed, from which it follows that is.

As long as the condition from equation (6) or (15) and the deviation from the equation (2) holds, it holds that the value E) goes 1.

In connection with the value of Cv = 64 k bit / s for voice transmission P2 is opt. " P1 as long as the value of Dv is »5 ms. It does if the assumption for the value of P1 remains between 100 and 200. on the other hand applies if the condition from equation (6) resp.

(15) no longer exists that again an approximate solution of N max is achieved by equation (4) assuming T = N x ropes, where T ave is the total station cycle time. The comparable conditions for the Values of D and P2 are in turn given by equations (13a) and (13b). The resulting result for the maximum number of stations for N max is this same as in the equation (14), with the value td being replaced by (td + Tave). The ratio of N max to N * max is in turn expressed by equation (17).

It is important that the ratio of N max to N * max can increase, since the values of D max and P2 opt. regardless of the network transmission rate C are.

For a given type of vocoder, N max / N * max is only a function of the efficiency E (see equation 17), which in turn is a function of Tp and thus of C. The previous discussion N max N * max, in which equation (6) or (15) is inequality, applies to low values of C. The upper limit of C, in which N max N * max can be found from equation (15). Under the condition that Cv = 64 kbit / s, Dv max = 300 ms and the average station cycle time Tave = 10 Ps, then N max N * max, provided that C «is 960 Mbit / s.

For a simple local network, such as that shown in Figure 2 is the maximum signal propagation delay between each The originating and receiving stations are neglected, so that the value Dv is Dv max.

The ratio of N max to N * max as a function of C is in Table for which the above values for Cv, Dv, Tave and for P1 = 100 and td = 0.5 ps are taken into account. The corresponding network efficiency E is the same approximately the ratio N max / N * max. The values of N max are also in listed in this table. It should be noted that the value of N max accordingly is to be reduced if a cable break or one or more faulty stations must be included in the consideration.

Are the geographically isolated local networks across be connected to so-called gateways, then the stations do not correspond longer with the value of Nmax within a local network. This is possible, for example from FIG. 4, in which the stations of the networks LAN 2 and LAN 3 have the central network LAN can be connected to each other. Should an information package for example from the network LAN 2 to the network LAN 3, so it has to pass through the central network LAN 1. The information package contains thereby three network delays in the propagation path between the Originating and receiving station. These network delays can be withdrawn be grouped together in the above analysis (see especially equations 9 to 13) with all stations and gateways connected to the corresponding networks LAN 1 to LAN 3 are connected. This results from the individual values of Tave and finally for the value of N max for the three networks LAN 1 to LAN 3. The number of stations including the gateways that are connected in one of these networks LAN 1 to LAN 3 are limited by the condition that the total number of stations in the three networks LAN 1 to LAN 3 must not exceed the value of N max. aside from that applies that connection combinations, for example from network LAN 2 to any a different, non-central network, different values for Tave and accordingly Retighten N max.

It follows that the maximum number of Stations not only depends on how many stations already have other networks are connected, but also from the value N max in connection with all possible Combinations of origin and destination that comprise these networks.

In addition, it must be taken into account that overarching connections of Networks over distances greater than 1000 km or more have a maximum propagation delay of tp, which are not neglected in connection with the value Dv max can. Since Dv max is fixed, the reduction results from the equation (1Db), where the value N max (equation 14) and more importantly the transmission bit rate for N max- + N N * max (equation 15) are also reduced.

For example, if the value Dv max = 300 ms and tpl = 100 ms (corresponding to approximately over 2 x 10 4 km) then Dv = 200 ms, so that N max-, N N * max goes, provided that the transmission bit rate C is approx. 640 Mbit / s. It should be noted that the value of N max indicates the number of all stations, the are connected to all networks, i.e. with the propagation paths between the Origin and receiving stations of all networks.

Then the start process for the passive star-configured local network described in accordance with FIG. 2, in which the previously discussed Access method is applied. After a station is supplied with voltage, this monitors the bus for a certain time ts. Here at is the time period ts selected so that all other stations are also functional. Receives this station sends an information packet within this time period ts, so is the connection switched on successfully. If no packet is received from this station, set this station precedes the so-called "live state" and occupies the bus accordingly itself. The transmission of the bit-serial message information via the local The network is thus started by this station, which initially picks up an empty packet transfers the bus. Further functions are already described in advance. Important is, that the stations must first be switched on and it is still a prerequisite that the stations are always in the switched-on state as long as the stations do not fail due to malfunctions. After a station has successfully accessed the local network is connected, the bus can be monitored within the period ts not applicable. This ensures that in the event of a cable break the appropriate station the bus by sending an empty packet on the bus within these ts seconds after determining the end of a packet information received from the receiver of this station confused.

This synchronous access protocol is for the transmission of bit serial News information, in particular real-time data via, fiber-optic, passive Star coupler particularly suitable.

C (Mbit / s) Nmax Nmax N * max 1 0.989 15 2 0.988 30 5 0.984 76 10 0.979 153 20 0.969 303 50 0.939 733 100 0.893 1400 200 0.814 2500 500 0.642 5000 1000 0.475 7400 2000 0.313 9800 5000 0.154 12 100 1 claim 4 figures - Blank page -

Claims (1)

Claim 1. A method for the optical transmission of bit-serial News information, in particular voice and video signals, in the form of a ring configured local networks equipped with coupling devices to which one to the transmission speed and the network structure of these local networks adapted number of different types of communication (voice, data, image, Text) processing stations (end devices) can be connected to each other in any combination are, with each station transmitting and receiving devices for transmitting the message information and the stations are arranged within the local networks in such a way that that the message information can be exchanged as required between the stations, whereby the order of access of the individual stations to the local network can be controlled by a synchronization clock, and the transmitting and receiving device of the stations to a passive star coupler designed as a straight-through mixer can be switched on so that the length of the transmission line is equal to the length of the reception line corresponds to the station in question, but the combination the characteristics 1.1 the synchronization clock is through a decentralized evaluation of the News information by all of the stations receiving the news information (1 to N) determines the authorization of access to the local network of a first terminal (1) to the next terminal (2 ... N) can be transmitted in rotation 1.2 the synchronization clock is through an original address information of the respective stations (1 to N) formed, with the original address information after the reception by the respective station (1 to N) recognized by this and updated is emittable, 1.3 the end devices (1 to N) each contain one Source address correction device, which if the receipt of the source address fails a subsequent authorized access station (1 to N) after repeating its own Source address the source address of the failed authorized access station (1 to N) transmits.