JPH03250937A - Lan control system - Google Patents

Abstract

PURPOSE:To improve the performance of a network by providing priority on each terminal equipment or each data package and decreasing the number of times of the retransmission at a driver section. CONSTITUTION:A packet control section 3 receives the information as for the number of times of collision from a collision read section 4 and a transmission packet generated by a controller 2 and discriminates whether or not a priority packet (broadcast type) is to be sent by the quantity of the increase of the number of times in collision. When a priority packet is sent from a hardware 5 by the instruction of the packet control section 3, till other PC receiving the priority packet receives a releasing packet or a transmission requirement with higher priority than the received priority received from a host, the own transmission is suspended. Since the traffic is tentatively decreased by such an operation, the transmission according to the priority is ensured.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a LAN control system, and in particular, even when traffic increases, collisions can be prevented and retransmissions can be reduced.
The present invention relates to a LAN control method that can improve network performance.

[Conventional technology]

Traditionally, CSMA/CD (carrier
sense multiple access/c
collision detection) method
A N (local area network)
) are widely used. In the C8MA/CD method, it is detected whether data is flowing on the network, and if there is space, data transmission is started. In this method, two nodes may start transmitting at the same time, so the node receives the transmitted data, checks whether it collides with other transmitted data, and if a collision is detected, (For LAN systems, see "Nikkei Electronics" 1985).
．． 4.22 pp. 186-239).

A conventional LAN system, for example, as shown in FIG. 4(a), includes an external storage device 41 storing shared resources, a server (SRV, ) 42, and a plurality of personal computers (PCs, to PCs) 43 to 45. It is constructed by connecting in series. Here, when the resources of the server 42 are shared by a plurality of personal computers 43 to 45, the server 42 and each of the personal computers 42 to 45 are linked by creating a virtual line.

Furthermore, as shown in FIG. 4(b), there is also a configuration in which a plurality of computers 46 to 48 are interconnected via a HUB 40. Currently, personal computer 46 is connected to LAN.
When participating in a network, first, by transmitting a unique network address to each computer 46 to 48 in a broadcast type, the broadcast type network address is stored in the network address storage table in the HUB 40. The personal computer 46 establishes a virtual line with a personal computer 47 that has performed the same operation, and sends and receives data.

However, since each personal computer mutually sends and receives packets at appropriate time intervals, a timer interrupt is required for each terminal, and if a system timer is used, timer interrupt conflicts with other applications may occur. was.

In addition, in the case of a CSMA/CD type LAN, as network traffic increases, collisions increase, which makes transmission difficult, and retransmissions increase, resulting in degraded performance.

[Problem to be Solved by the Invention] As described above, in a LAN system in which lower layers are controlled using the CSMA/CD method, as line traffic increases, collisions occur frequently and the number of retries due to collisions is reached. In this case, the packet cannot be sent.

In that case, software or firmware (media access control section (MAC) of the driver section)
) will return an error, and if collisions occur frequently, the number of collisions will increase. In this way, when an error that could not be sent to the driver is returned,
It takes a long time to request retransmission.

Therefore, if such a situation occurs frequently, performance will deteriorate significantly.

An object of the present invention is to provide a LAN control method that can solve these conventional problems and improve network performance by reducing retransmissions in the driver section even when traffic increases. be.

[Problems to be solved] In order to achieve the above object, the LAN control method of the present invention has the following features:
A CS M that performs mutual communication by reading the transmission data written by the network program of any terminal, generating transmission packets by the controller, and sending the transmission packets to other terminals via the cable by the hardware.
In an A/CD type LAN system, each terminal has a counter that obtains the increase in collisions from a network program and counts the number of collisions, and receives information from the counter as network traffic increases.
Packet control means is provided which generates a priority packet that determines the priority of the packet and instructs hardware to transmit it, and after the packet control means generates the priority packet and transmits the packet, the controller generates the priority packet. It is characterized by transmitting data packets.

[Function] In the present invention, by transmitting packets with priority information from each terminal on the LAN, a state in which it is easy to transmit is set. A terminal that receives a priority packet suspends transmission until it receives a release packet or receives a request to transmit a higher priority packet. There are two ways to prioritize the data: one to give priority to the data itself, and the other to give priority to the terminal itself. This results in
It is possible to reduce retransmissions in the driver part and improve network performance.

〔Example〕

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a configuration diagram inside a terminal showing an embodiment of the present invention.

In FIG. 1, 1 is a network program that executes LAN system protocols and communications under the control of O8, 2 is a media access controller that creates transmission packets and stores them in memory, and 3 is a newly provided program according to the present invention. 4 is a collision read unit that reads whether or not a packet has collided with a packet transmitted from another terminal; 5 is hardware such as an encoder/decoder, a pulse transformer, and a transceiver.

In FIG. 1, if a packet collides with a packet sent from another terminal, an error is returned to the software when the number of retries is reached, so the collision read unit 4
The number of collisions indicated by a LAN processor or the like is periodically read using a timer interrupt or the like. L.A.
After writing the data to be transmitted into the memory, the N processor sends the address of the transmission data and the 6th transmission command to the controller (MAC) 2. The controller 2 reads necessary data from the memory, adds a preamble, destination/own address, CRC, etc., and generates a packet. The generated packets are subjected to parallel-to-serial conversion within the hardware S and encoded by an encoder/decoder. The transceiver in the hardware S converts the encoded code (media) into a specified voltage using a built-in driver, and then sends it out to the coaxial cable. In the case of reception, the opposite operation is performed. That is, when the transceiver receives a packet input from the coaxial cable, the encoder/decoder in the hardware 5 converts the code received from the transceiver into an NRZ code and sends it to the controller 2.

Controller 2 disassembles the packet created by serial-parallel conversion, checks whether the received packet is addressed to itself,
If the packet is addressed to itself, the data is written into memory.

Controller 2 not only generates/disassembles packets, but also performs network control functions (in this case, retransmission of packets, etc.)
I also have In other words, the encoder/decoder and transceiver perform operations corresponding to the physical layer (first layer), the controller 2 performs operations corresponding to the data link layer (second layer), and the network layer (third layer) and transport The network program 1 performs operations corresponding to the layer (4th layer), and executes operations corresponding to the session layer (5th layer), presentation layer (6th layer), and application layer (7th layer).
The program on the host computer performs the operation corresponding to the layer).

The packet control unit 3 of the present invention receives information on the number of collisions from the collision read unit 4, receives the transmission packet generated by the controller 2, and transmits a priority packet (broadcast type) according to the amount of increase in the number of collisions. Determine whether or not.

Here, there are two ways of assigning priorities:

(a) Priority order based on data: The upper layer of the protocol determines the priority order based on the type of data, and notifies the packet control unit 3 of the priority order.

(b) Priority order for each personal computer (PC): A network administrator assigns a priority order to each PC in advance based on the frequency of use, etc.

In the case of (a) above, when a priority packet is transmitted from the hardware 5 according to an instruction from the packet control unit 3,
Other PCs that have received this priority packet stop their own transmission until they receive a release packet or a transmission request with a higher priority than the received priority from the upper layer.

In the case of (b) above, when the hardware 5 transmits the priority packet] and the priority packet according to an instruction from the control unit 3, other PCs that receive this priority packet will determine whether their own power is the priority packet or not. If it is low, it stops transmitting until it receives a release packet. Also, if you have a higher priority,
You can continue sending as before.

These operations can temporarily reduce the amount of traffic, so it is possible to ensure transmission based on priority.

FIG. 2 is a format diagram of a priority bucket used in the present invention.

As shown in FIG. 2, the priority packet includes a destination address whose contents are all F, a source address, a 2-byte type field, a logical link control (LLC), and a priority identifier. Priority number 1
and a 4-byte PC3 (check pit). The destination address and source address are 12 bytes, and the type field is 2 bytes.

FIG. 3 is an operational flowchart of the present invention.

The operation of this embodiment will be explained by giving an example.

Now, assume that the collision lead interval is once every 0.1 seconds. Further, it is assumed that the amount of increase in the number of collisions for determining the transmission of a priority packet is 100, and that the priority is divided into three levels, with 3 being a higher priority than 1.

FIG. 3 shows the operation flow in case (a) above. PCI, PC2, PC3 exist, PCI is PC
When transmitting data to the packet controller 2, the packet controller 3 operates according to the following procedure.

First, when there is a transmission request to PC2 (step 101)
, it is determined whether the increase in collision is 100 or more (step 102), and if it is 100 or more, the other PC
In order to suppress the transmission of packets, priority [2] packets are transmitted as broadcast type packets (step 1o3). Thereafter, the data packet is transmitted to the PC 2 (step 104). After the transmission is completed, a cancellation packet is transmitted (step 105). Note that when the collision increase amount is not 100 or more (step 102), a data packet is transmitted (step 104).

PC2 that received the priority packet has a transmission request (priority [1]) to PC3 at that time (step 111).
), and PC3 sends a transmission request to PC2 (priority [3
]) (step 121), PC2 cannot transmit, and PC3 can transmit (step 122).

After receiving the release packet, PC2 transmits its own data packet (step 112).

Next, in the case of (b) above, PCI, PC2゜PC3
If the priorities are 2, 1, and 3, respectively, that is, PC3 is the highest, PCI is next, and PC2 is the lowest, the processing is exactly the same as in case (a) above.

In this way, in this embodiment, the collision read section 4 is provided with a counter for acquiring the amount of increase in collision.
By providing the packet control unit 3 that transmits priority packets, priority packets are transmitted as network traffic increases, so transmission from other terminals can be suppressed and transmission becomes easier.

[Effects of the Invention] As explained above, according to the present invention, by setting priorities for each terminal or each data packet, it is possible to create a state in which transmission is easy even when traffic increases. Network performance can be improved by reducing retransmissions on the driver part.

[Brief explanation of drawings]

FIG. 1 is a configuration diagram of a control portion of a LAN terminal showing an embodiment of the present invention, FIG. 2 is a format diagram of a priority packet used in the present invention, and FIG. 3 is an operation flowchart of each terminal according to the present invention. FIG. 4 is a LAN connection configuration diagram. 1: Network program module, 2 Media access controller (MAC), 3 Packet control unit,
4: Collision lead section, 5 hardware, 41: External storage device, 42. Servers, 43 to 50: Personal computers forming terminals, 40: HUB.

Claims (1)

[Claims] (1) CSMA, in which the controller reads the transmission data written by the network program of any terminal, generates a transmission packet, and the hardware sends the transmission packet to another terminal via a cable, thereby performing mutual communication. In a /CD type LAN system, each terminal has a counter that obtains the increase in collisions from the network program and counts the number of collisions, and a counter that receives information from the counter and counts the number of packets as network traffic increases. Packet control means is provided for generating a priority packet for determining a priority and instructing the hardware to transmit the packet, and after the packet control means generates the priority packet and transmits the packet, the controller A LAN control method characterized by transmitting data packets generated by.