RU2068579C1 - Method user access to data bus in ethernet local network and device which implements said method - Google Patents

Abstract

FIELD: automation and computer engineering. SUBSTANCE: when transmission of message through bus of local Ethernet network is finished, most active users have right to start transmission of message to data bus after delay time which duration is determined by user number and time for signal transmission through data bus. Corresponding device has receiver, transmitter, encoder-decoder, network controller, user connection unit, OR gate, signal generator and dynamic priority unit. EFFECT: increased functional capabilities. 3 cl, 4 dwg

Description

 The invention relates to automation and computer technology, more specifically, to information transfer systems, and in particular to methods and devices for accessing a system subscriber to a local area network Ethernet bus and can be used in devices using a common data bus for a number of subscribers.

 A known method according to US patent N 4628311, in accordance with which each subscriber of the computer network is assigned a unique time interval for transmission delay, and during the operation of the computer network, these unique time intervals are redistributed according to the cyclic law.

 The closest is the method according to US patent N 4063220, which ensures the efficiency of the Ethernet computing network with the coincidence of the time interval of the transmission delay of the message block. This method consists in the fact that the subscribers monitor the state of the data bus and if the bus is free, they start transmitting the message block on the bus after a single time interval for all subscribers necessary to calm the transients in the data bus. If the transmission of message blocks of different subscribers overlap each other, the subscribers interrupt the transmission and repeat it after a time determined by the number of failed transmission attempts. When using this method, as shown by numerous studies and observations in real local Ethernet computing networks with a network flow exceeding about 1/3 of the maximum flow, which is about 3 Mbit / s, it is possible to increase the number of conflicts in the network, up to the clinch at which the throughput of 2 data buses drops to zero.

 A device is known for accessing a subscriber to a data bus of a local area network Ethernet (DP839EB Network Evaluation Board. See Data Communications Local Area Network UARNS Handbook, National Semiconductor, 1990). It contains a data bus and adapters, the number of adapters is equal to the number of subscribers of the computer network, each adapter contains a receiver whose input is connected to the data bus, a transmitter whose output is connected to the data bus, encoder / decoder, network controller and communication unit with the subscriber, input / the output of the communication unit with the subscriber is connected to the subscriber, the first output of the controller is connected to the input of the communication unit with the subscriber, the output of the communication unit with the subscriber is connected to the input of the controller, the second output of the controller is connected to the second input of the encoder / decoder, the output to Der / decoder connected to the input of the transmitter.

 The message block intended by the subscriber for transmission is transmitted through the communication block with the subscriber to the network controller of the adapter and stored in it. The message blocks transmitted over the data bus are received by the receiver, decoded by the encoder / decoder, sent to the network controller, and if it is addressed to the subscriber, stored in the network controller and initiate the transmission of the message block to the subscriber using the communication unit with the subscriber.

 When the message block is located in the network controller of the adapter, and the adapter is waiting for the end of the transmission of the current block of messages on the data bus, after which they begin to transmit a message block through the encoder / decoder to the transmitter input and then to the data bus. If, in this case, the receiver establishes the superposition of the transmissions of two or more message blocks on top of each other, the transmission is stopped and resumed after a random time selected from the interval, the value of which is doubled after each unsuccessful transmission attempt.

 However, when using this device, when the information flow exceeds approximately 3 Mbit / s, there is a danger of a large number of conflicts and even clinching in the network, which reduces the reliability of information transfer and leads to unproductive time losses.

 The purpose of the invention is the creation of such a method of subscriber access to the data bus and device for its implementation, which would improve the efficiency of using the Ethernet data bus by reducing the number of conflicts and increasing the allowable flow of information, which maintains the stability of the system.

 The goal is achieved in that each subscriber has a number in the Ethernet computing network, and after the transmission of the message block of one of the subscribers of the computer network to the other subscriber is granted the right to start transmitting the message block after a time interval, the duration of which is determined by the subscriber’s number and the signal propagation time via the data bus.

 This allows you to drastically reduce the number of conflicts in the computer network and increase its throughput.

 The goal is also achieved by the fact that the subscriber’s access device to the data bus (adapter) contains a receiver, the input of which is connected to the data bus, a transmitter, the output of which is connected to the data bus, an encoder / decoder, a controller and a communication unit with the subscriber, the input-output of the unit is connected connected to the subscriber with the subscriber, the first output of the controller is connected to the input of the communication unit with the subscriber, the output of the communication unit with the subscriber is connected to the input of the controller, the second output of the controller is connected to the second input of the encoder / decoder, the output of the encoder / decoder is connected to the transmitter com, additionally contains a signal driver and an OR element connected in series, the second output of the encoder / decoder is connected to the input of the driver and to the second input of the OR element, the output of the OR element is connected to the second input of the controller.

 This allows you to drastically reduce the number of conflicts in the computer network and increase its throughput. In addition, the device contains a dynamic priority block, the first input of the dynamic priority block is connected to the third output of the encoder / decoder, the second input of the dynamic priority block is connected to the second output of the controller, the output of the dynamic priority block is connected to the second input of the shaper.

 This improves the reliability of the device by eliminating the possibility of blocking the data bus by one or more subscribers.

 FIG. 1 schematically depicts a simplified configuration of a local area network; figure 2 is a block diagram of a device for accessing a subscriber to a computer network bus; figure 3 is a block diagram of another embodiment of a device for accessing a subscriber to a computer network bus; 4 is a block diagram of a dynamic priority block.

The method for accessing subscribers A ₁ , A ₂ , A, A _N to the data bus of the Ethernet computing network is as follows.

The computing network consists of N subscribers A ₁ , A ₂ , A _j , A _N , each of which is connected to the data bus through its adapter D ₁ , D ₂ , D _i , D _N.

Subscribers A ₁ , A ₂ , A _i , A _N can be separate computers exchanging information with each other using devices for accessing subscribers to the data bus D ₁ , D ₂ , D _i , D _N , nodal processors, through which individual local computing networks are combined into a larger network, functional nodes of a computer system, interconnected by a data bus, and so on.

 The data bus can be a coaxial cable, or a twisted pair of conductors, as defined by the IEEE 802.3 standard.

The procedure for access of subscribers A ₁ , A ₂ , A _i , A _N to the data bus is as follows.

When the system is initiated, Priority subscribers (K ≅) A ₁ , A ₂ , A _j , A _{k are} assigned their own number j from the set of numbers (1, K) and there are no two subscribers A ₁ , A ₂ , A _j , A _k , their own whose numbers would match.

 The subscriber’s priority is the higher, the lower his own number. In the future, these numbers can be changed.

All system subscribers exchange messages, passing them to their own Ethernet adapters and further to the Ethernet network data bus
Each time, when a subscriber A _i (1≅ i≅ N) forms a message block for transmission, the subscriber transmits this prepared message block to the Ethernet adapter connected to the subscriber. If the data bus is free at this moment, that is, the message block is not transmitted via the data bus, the adapter starts transmitting the message block received from the subscriber. If the data bus is currently busy sending a message block, the adapter waits for the end of the transmission.

When the data bus is released, that is, at the end of the data transfer from the message block, the adapter D _i sets the logic zero level at the input of determining the bus occupancy of the data of the Ethernet network controller, in response to which the Ethernet controller, after a time determined by its operation algorithm, initiates transmitting a prepared message block to the data bus.

If subscriber A _i is included in the group of priority subscribers A ₁ , A ₂ , A _j , A _k , then when the data bus is released, adapter D _i transmits a signal to the input of the Ethernet network controller, the duration of which is determined by subscriber number A and the signal propagation time via the bus data. This signal can be a meander supplied to the input of the clock pulses of the Ethernet network controller, or a level of a logical unit, which is input to the definition of busyness of the controller data bus.

 The signal duration can be proportional to the double signal propagation time between the ends of the local area network. If the subscribers have information about the position of each subscriber on the data bus, and can determine the point in time when the transmitted message block extends to the boundary of the computer network, the signal duration can be proportional to the signal propagation time between the ends of the computer network.

 If none of the non-priority subscribers has a message block ready for transmission, and one of the priority subscribers starts the transmission, then, since the signal duration is equal to twice the signal propagation time over the data bus, information about the start of transmission will be distributed along the entire data bus, and the remaining priority subscribers they will have time to register this transfer and postpone their transfers until the next bus release. This eliminates the possibility of a conflict of priority subscribers among themselves.

A device for accessing subscribers A ₁ , A ₂ , A _N to the data bus Etnhernet contains a receiver 1, the input of which is connected to the data bus, transmitter 2, encoder / decoder 3, network controller 4, OR element 5, signal conditioner 6 , and the unit of communication with the subscriber 7.

 Data from the output of the receiver 1 is received at the first input of the encoder / decoder 3. From the first output of the encoder / decoder 3, the data intended for transmission via the network data bus B is received at the input of the transmitter 2 and then to the data bus B.

From the second output of the encoder / decoder 3, the data received from the bus In the data of the Ethernet network, is fed to the input of the signal conditioner 4, and to the first input of the OR element 5. The signal from the output of the signal shaper 4 is fed to the second input of the OR element 5. From the output of the OR element 5, the combined signal is fed to the first input of the network controller 6. Message blocks addressed to the subscriber from the first output of the network controller 4 are fed to the input of the communication unit 7 with the subscriber, from the input / output of which they are sent to the subscriber A _i .

 The message blocks intended for transmission via the B data bus are fed to the input-output of the communication unit 7 with the subscriber, and from the output of the communication unit 7 to the subscriber are fed to the second input of the network controller 4. From the second output of the network controller 4, information and clock pulses are transmitted to the second encoder / decoder input 3.

 The device operates as follows.

Subscribers A ₁ , A ₂ , A _{N of} an Ethernet computing network transmit message blocks on data bus B using adapters D ₁ , D ₂ , D _N. To this end, in each subscriber A _i (1≅i≅N), the prepared message block is transmitted through the input / output of the communication unit 7 to the second input of the network controller 4 and stored in it.

 Information about messages transmitted on the B data bus is constantly coming from the output of receiver 1 to the first input of the encoder / decoder 3.

 After the transmission of data message packet data via bus 3 at the first output of the encoder / decoder, a logic zero level is set, which initiates the generation of signal shaper 6 of the signal of the required duration. Data from the first output of the encoder / decoder 3 and the signal at the output of the signal conditioner 6 are combined with an OR element 5 and fed to the first input of the network controller 4.

The message blocks addressed to the subscriber A _i are stored in the network controller 4. In this case, the network controller 4 initiates a request for the communication unit 7 with the subscriber, which transfers the stored information from the input / output of the communication unit 7 with the subscriber to the subscriber A _i .

The information intended for subscriber A _i to be transmitted via data bus B is transmitted from the input / output of the communication unit 7 to the output of the communication unit 7 to the second input of the network controller 4 and is stored in the network controller 4.

Another embodiment of the device for accessing subscribers A ₁ , A ₂ , A _i , A _N to the data bus B of the Ethernet computing network is shown in FIG. 3. The device comprises a receiver 1, a transmitter 2 encoder / decoder, a network controller 4, an OR element 5, a network driver 6 and a communication unit 7 with a subscriber connected to each other and to a data bus 7 in series with each other and connected to the data bus B in the same way as described above in FIG. .2. In addition, each device D _i contains a dynamic priority block 8, the first input of the dynamic priority block 8 is connected to the third output of the encoder / decoder 3, the second input of the dynamic priority block 8 is connected to the second output of the network controller 4, the output of the dynamic priority block 8 is connected to the second Signal conditioner input 6.

Message blocks transmitted by subscribers A ₁ , A ₂ , A _i , A _N on the bus In the data of the Ethernet computing network, come from the third output of the encoder / decoder 3 to the input of the dynamic priority block 8. In block 8 of dynamic priority, the address of the message source is extracted from these message blocks and, in accordance with the service discipline adopted in the system, the number of the subscriber A is changed, according to which the duration of the signal generated by the signal conditioner 6 is set.

For example, after the initial initiation of the system, the subscriber numbers A ₁ , A ₂ , A _i , A _N can be changed according to the cyclic law, as a result of which the subscriber A _j , who last sent the message block via bus B of the Ethernet data network, is set to the lowest priority environment of other subscribers A ₁ , A ₂ , A _i , A _N , which will allow to exclude bus locks In the data of one or a group of the most active subscribers, the total information flow from which exceeds the bandwidth of the data bus B.

 An embodiment of the dynamic priority block 8 is shown in FIG. 4. Dynamic priority block 8 contains a data input 9, a clock signal input 10, an end signal transmission signal input block 11, and a conflict signal input 12, a message source address allocation block 13, a register 14, a memory block 15, a comparator 16, a first AND element 17 , the second element And 18, the counter 19, the block 20 priority conversion, the priority output 21 and the input signal of the own transmission 22, which is the second input of the block 8 dynamic priority.

 Data signals 9, clock signals 10, the end signal of the message block 11 and the conflict signal 12 are received at the first input of the dynamic priority block 8 from the third output of the encoder / decoder 3. Data signals 9 are fed to the first input of the message source address allocation block 13, clock signals 10 are fed to the second input of the message source address allocation block 13. Data from the output of the block 13 allocating the address of the message source is fed to the input of the register 14, from the output of which the stored address goes to the input of the memory block 15.

 The comparator 16 compares the message source address generated at the output of the register 14 and the addresses of the message sources stored in the memory unit 15 and, if this address is not contained in the block 15, the voltage level of the logical unit is set at the output of the comparator 16.

 The output of the comparator 16 is connected to the first input of the first element And 17. At the second input of the first element And 17 receives the signal of the end of the transmission of message 11, the third, inverse, the input of the first element And 17 receives the conflict signal 12 and the fourth, inverse, the input of the first element And 17 a signal of own transmission 22 is received. From the output of the first element And 17, the signal is fed to the first counting input of the counter 19 and to the second input of the recording block 15 of the memory.

 The signal of the end of transmission of the message block 11 is also fed to the first input of the second element And 18, the second, inverse, input of the second element And 18 receives a conflict signal 12 and the third input of the second element And 18 receives the signal of own transmission 22. From the output of the second element And 18 the signal is fed to the second reset input of the counter 19 and to the third reset input of the memory unit 15.

 From the output of the counter 19, the signal goes to the input of the priority conversion unit 20, the output of which 21 is the output of the dynamic priority unit 8.

 Block 8 dynamic priority works as follows.

 Data and clock signals in a serial code from the third output of the encoder / decoder 3 are received at the first input of the dynamic priority block 8 at the first data input 9 and the second clock input 10 of the signal source allocation block 13.

 The message source address allocation unit 13 reads from 113 to 160 bits of each received message block, which, in accordance with the Ethernet exchange protocol, are the message source address, converts them in parallel to a word of 6 bytes in size and puts it in register 14.

 The formation of a new 6-byte word at the output of register 14 initiates the operation of comparator unit 16, which scans the entire contents of memory unit 15 and compares it with the data installed at the first input of comparator 16. If memory unit 15 already contains the word contained in register 14, the output of the comparator 16 set the voltage level of the logical zero, otherwise at the output of the comparator 16 set the voltage level of the logical unit.

If the message source address stored in the register 14 is new and is not contained in the memory block 15, then, in the case of successful completion of the transmission of the message block via the data bus B, if there was no conflict during the transmission and the transfer was not made by the subscriber A _i , and any other subscriber, a pulse is generated at the first counting input of the counter 19 and the contents of the counter 19 are increased by one, and a new address from the output of the register 14 is entered in the memory block 15.

If there was a conflict during the transmission of the message block on the B data bus, then the priority of the subscriber A _i does not change.

If the message block was transmitted via the B data bus by the subscriber A _i , then the output of the second element And 18 forms the voltage level of the logical unit, which sets the counter 19 to zero and clears the contents of the memory block 15.

 The priority conversion unit 20 converts the input code K into the output code L, realizing such a discipline of priority service, which is required from the conditions of the most efficient functioning of the entire system. The number of priority subscribers in the Ethernet network is determined by the number of installed claimed devices. Network subscribers who are not equipped with the claimed device have a higher priority, however, this priority is the same for all subscribers, and if the information flow from these subscribers is high, they can cause conflicts in the process of exchanging messages on the B data bus, reducing the data throughput of the B bus and the whole system.

 If the information flow from any subscriber of the Ethernet network is high enough, this subscriber is equipped with the inventive device and set a lower priority for it. A number of such subscribers supplying the main flow of information do not conflict with low-active subscribers and among themselves, however, depending on the adopted law of code conversion of the code converter unit 20, and thus the characteristics of the entire system will be different from the priority service discipline adopted for them.

 If the conversion law is chosen L K, that is, there is no conversion, then the subscriber who carries out the last transmission of the message block on the data bus B receives the highest priority among priority subscribers. At the same time, the most active priority subscribers will have a preferential right of service, which generally improves system performance. However, in this case, if the total information flow from a number of priority subscribers and from non-priority subscribers exceeds the data throughput of the bus, the system will lose operability.

 If the conversion law is chosen L m K, where m is the number of installed inventive devices in the Ethernet network, then a priority will be cyclically changed for priority subscribers, which reduces the performance of the system as a whole, but guarantees its operability with any information flow from priority subscribers.

 The method and device for accessing a subscriber of an Ethernet computing network to the B data bus can significantly reduce the number of conflicts in an Ethernet network, especially with significant input streams, and dramatically reduce the average and maximum delivery time of a message block.

 This makes it possible to improve the performance of already installed Ethernet networks.

 In Ethernet networks, it is not necessary to install a subscriber access device to the data bus to each Ethernet adapter in order to feel a significant improvement in the network operating mode. If there are several highly active subscribers in the network, for example, servers, they should be equipped with the proposed devices. In this mode, the average message delivery time on the LAN is reduced several times, and the maximum message delivery time is reduced tens of times.

Claims (3)

 1. A way for a subscriber to access a data bus in an Ethernet computing network, which consists in monitoring the status of the data bus, and if the data bus is free, subscribers having a message block start transmitting them via the data bus, if the data bus is busy, then subscribers with a message block wait for the end of transmission on the data bus and begin transmitting the message block on the data bus after a time interval, if several subscribers register a simultaneous transmission of the message block, all subscribers stop the transmission and they resume the attempt to transmit a message block after a time interval, determined by the number of unsuccessful attempts to transmit a message block, characterized in that each subscriber of the computer network is assigned a number and subscribers who have a message block begin transmission after a time interval, the duration of which is determined by the subscriber number and the signal propagation time by data bus.  2. A device for implementing a method of accessing a subscriber to a data bus in an Ethernet computing network, comprising a receiver, a transmitter, an encoder / decoder, a network controller, and a communication unit with a subscriber, the input of the receiver being the input of the device for connecting to the data bus, the output of the receiver is connected to the first the input of the encoder / decoder, the first output of which is connected to the input of the transmitter, the output of which is the output of the device for connecting to the data bus, the first output of the network controller is connected to the second input of the encoder / decoder, the second the course of the network controller is connected to the input of the communication unit with the subscriber, the output of which is connected to the first input of the network controller, the input / output of the communication unit with the subscriber is the input / output of the device for connecting to the subscriber, characterized in that the signal driver and the OR element are introduced, the second the output of the encoder / decoder is connected to the first inputs of the OR element and the signal conditioner, the output of which is connected to the second input of the OR element, the output of which is connected to the second input of the network controller.  3. The device according to claim 2, characterized in that the dynamic priority block is introduced, the third output of the encoder / decoder connected to the first input of the dynamic priority block, the second input of which is connected to the first output of the network controller, the output of the dynamic priority block is connected to the second input of the driver signals

Images