JPH04340829A - Congestion control method for in-network device - Google Patents

Abstract

PURPOSE:To release a congesting state without deteriorating a communicating efficiency. CONSTITUTION:Plural devices 11, 12,..., connected with a network 1 of a taken passing system, alternately operate a data communication by the medium control sublayer of the data link layer of a communication control protocol. The priority of the data communication of the specific devices 11, 13,... among those plural devices 11, 12,..., is high, and that of the other devices 12, 14,... is low. At the time of detecting the congesting state of the devices 11, 13,... whose priority is high, the priority in the frame of the medium control sublayer is set to be high, so that the data communication can be operated only by the devices 11, 13,... whose priority is high. Thus, the communication by the devices 12, 14,... whose priority is low can be controlled, and the congesting state can be released.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a congestion control method for a plurality of devices connected to a token passing network.

0002

[Prior Art] A plurality of devices connected to a token passing network access communication media such as magnetic disk devices and exchange data between each device in accordance with the ISO 8802-5 standard. Lower Layer Implementation Regulations S012 (V1.0) Information Processing Interoperability Technology Association, Token Ring Acc
ess Method ANSI/IEEE”). In such data exchange, the congestion state of each device is generally controlled by a protocol at a layer higher than the logical link sublayer of the OSI communication layer.

FIG. 2 is a diagram showing the communication layer of OSI. As is well known, the communication function of an open system in OSI (Open Systems Interconnection) is modeled in seven layers as shown in the figure. That is, these layers include an application layer 21, a presentation layer 22, a session layer 23, a transport layer 24, and a network layer 2.
5, a data link layer 26, and a physical layer 27. A detailed description of each communication layer will be omitted.

Of these communication layers, the data link layer includes a logical link sublayer 28 and a medium control sublayer (MAC sublayer).
It consists of 29. The logical link sublayer 28 is a communication layer that transmits and receives logical data. The medium control sublayer 29 is a communication layer that allows direct access to storage media such as magnetic disks.

[0005] In end systems such as server systems that establish connections and communicate with an unspecified number of parties, data communication is performed using a logical link sublayer protocol. Therefore, in server systems and the like, even when controlling congestion, control is performed in the logical link sublayer. Here, congestion refers to a request for communication exceeding the communication capacity of a server system, etc. The communication capacity is determined by the number of communication buffers prepared in each end system such as a server system.

FIG. 3 is a conceptual diagram of communication buffer management between communication layers. In the logical link sublayer, the communication buffer prepared in the medium control sublayer is divided into appropriate management units, and buffer management is performed for each management unit. Furthermore, in the network layer, the communication buffer management unit in the logical link layer is divided into appropriate management units, and buffer management is performed for each management unit.

FIG. 4 is an explanatory diagram of communication buffer management in the logical link sublayer. In the illustrated example, the total number of communication buffers in the media control sublayer is nine. These communication buffers are managed in units of three in the logical link sublayer. Therefore, the total number of connections in the logical link sublayer is 3.
Become an individual.

FIG. 5 is an explanatory diagram of congestion control in the logical link sublayer. In the illustrated example, an end system 50 establishes connections with end systems 51, 52, and 53 through logical links, and communicates with them. As shown in FIG. 4, the total number of connections in the logical link sublayer is three. Therefore, a connection cannot be established with the fourth end system 54, resulting in a congestion state. Congestion control in the logical link sublayer has conventionally been performed using the following method.

The first method is to control the number of connections by keeping the maximum number of buffers required for each connection constant. That is, the method is shown in FIG. In other words, the number of connections on the right side of the following equation (1) is limited so that it does not exceed the total number of buffers on the left side. (Total number of buffers) ≧ (Maximum number of buffers required for one connection) × (Number of connections) … (1)
In the example of FIG. 4, the total number of buffers is nine, the maximum number of buffers required for one connection is three, and the number of connections is limited to three or less. This basically prevents congestion from occurring on the established connection.

The second method is to control the maximum number of buffers required for each connection. That is, the following formula (2
) controls the maximum number of buffers required for each connection on the right side so that the total for all connections does not exceed the total number of buffers on the left side. (Total number of buffers) ≧ (Maximum number of buffers required for each connection) for all connections... (2)

The second method will be explained with reference to FIG. 4. The maximum number of buffers required for each connection is not limited to three, but is changed to two or one depending on usage conditions, etc. . As a result, even if the total number of communication buffers is nine, for example, the maximum number of buffers for three connections can be set to two, and the maximum number of buffers for one connection can be set to one, increasing the total number of connections to four. As a result, a connection can also be established with the fourth end system 54 shown in FIG. Therefore, it is possible to basically prevent congestion from occurring while improving the usage efficiency of communication buffers.

0012

[Problems to be Solved by the Invention] However, the above-mentioned conventional technology has the following problems. That is, in the first method, the usage efficiency of communication buffers per connection is low, and even if there are communication buffers that are not used for data communication, it is not possible to increase the number of connections by more than a certain number (for example, 3). There wasn't. Therefore, there was a problem in that congestion was likely to occur. Also, in the second method,
Since the maximum number of buffers was made variable depending on usage conditions, etc., there was a problem in that the structure of the software for variable control became complicated. The present invention has been made in view of the above points, and it is an object of the present invention to provide a congestion control method that uses buffers efficiently and has a simple software structure.

[0013]

[Means for Solving the Problems] The congestion control method for devices in a network of the present invention connects a plurality of devices to a network using a token passing system in which only devices with transmission rights can transmit data, and The device is
Data communication is performed mutually using the medium control sublayer of the data link layer of the communication control protocol, and certain devices among the plurality of devices are given high priority for data communication, and For other devices, the data communication priority is set low, and when a congestion state of the high priority device is detected, the priority provided in the priority field in the frame of the medium control sublayer is set high. Accordingly, data communication is performed only between the high-priority devices, and after the high-priority devices are no longer in a congested state, the priority in the frame of the medium control sublayer is lowered. It is.

[0014]

In the congestion control method for devices in a network according to the present invention, a plurality of devices connected to a token passing network perform data communication with each other using the medium control sublayer of the data link layer of the communication control protocol. Here, data communication priority is set high for some specific devices among these multiple devices. Furthermore, for other devices, the data communication priority is lowered. When a congestion state of a high-priority device is detected, data communication is performed only between the high-priority devices by setting a high priority in the priority field of the frame of the medium control sublayer. Let's do it. This restricts communication by low priority devices and releases the congestion state. After the high-priority device is no longer in a congested state, the priority in the medium control sublayer frame is returned to a low state, and communication is performed between all devices.

[0015]

Embodiments Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a system using the method of the present invention. The illustrated system consists of a plurality of devices 11, 12, . . . connected to a token-passing network 1. In the token passing network 1, only specific devices can transmit data in order to prevent communication conflicts. That is, the token passing method is a method in which data collisions do not occur on the transmission path of the network by mutually exchanging transmission rights called tokens. The tokens are circulated within the network 1 one or more predetermined times.

The plurality of devices 11, 12, . . . are terminal devices in a narrow sense such as a display and keyboard used by an operator, a file server such as a magnetic disk device, a printer server, and the like. That is, these multiple devices 11, 1
2,... constitute a so-called workstation.

The devices 11, 12, . . . mutually perform data communication via the network 1 using the medium control sublayer of the data link layer of the communication control protocol. The communication control protocol is, for example, based on OSI (Open Systems Interconnection). The communication function of an open system is
As mentioned above, it is modeled in seven layers. One of these layers is the data link layer. The data link layer consists of a logical link sublayer and a media control sublayer. The logical link sublayer is a communication layer that transmits and receives logical data. The media control sublayer is a communication layer that allows access to storage media such as magnetic disks.

Among the plurality of devices 11, 12, . . . , some of the devices 11, 13, . . . shown in the figure are given high priority during data communication. In addition, other devices 12
, 14, . . . are given low priority. For example, the device 11 is a server system such as a magnetic disk device. High priority devices 11, 13,
... sets a high priority in the priority field in the frame of the medium control sublayer when a congestion state is detected during data communication. The priority field in the frame of the medium control sublayer is provided for arbitrary use by the user.

In the present invention, the priority set in the priority field of the frame to be transmitted is determined according to the congestion state of the server system as follows (a) and (b).
) The settings are controlled as shown in . Here, the congestion state means that the communication buffer capacity of the server system is, for example, 80%.
This means that the tank is almost full.

(a) If the server system is not in a congested state, the priority is set low. This results in
All end systems are equally entitled to transmit. (b) When the server system is in a congested state, the priority is set high. This allows only devices that can transmit with high priority, including the server system, to obtain the right to transmit. Therefore, transmission is prohibited for general end systems. As a result, transmission from the server system is prioritized, and the congestion state is relieved. Hereinafter, the congestion control method of the present invention will be sequentially explained with reference to the flowchart of FIG. 6 and the operation example of FIG. 7.

FIG. 6 is a flowchart showing the congestion control operation procedure of the server system. This figure shows the internal processing of the server system, and the processing consists of the following three
It is broadly divided into (1) Processing corresponding to the event (reception processing, transmission completion processing)
(Steps S1 to S4) (2) Congestion determination processing (Step S5) (3) Transmission processing (Steps S6 to S11)

[0022]
Among these, the present invention has a feature in the transmission processing (steps S6 to S11) after the congestion determination processing, and if it is determined that the congestion state is present, the priority value set in the priority field of the frame is increased. setting (step S6) and transmitting data (steps S7 to S9). At this time, if there is data to be transmitted, that data is transmitted, but if there is not, dummy data is generated (step S8).
, send this data.

[0023] The reason why dummy data is transmitted when there is no data to be transmitted is that even if the priority of the frame in the medium control sublayer is set high, if no data is actually transmitted, other communication layers This is because there is a risk that the priority of frames in the medium control sublayer may be lowered due to control of the media control sublayer. Note that communication protocols are often arranged so that the priority in a frame cannot be increased unless there is data to be transmitted.

[0024] By transmitting such high-priority data, transmission by other end systems is suppressed, and the congestion state is relieved. Next, an operational example of congestion control will be described with reference to FIG.

FIG. 7 is a diagram showing an example of the congestion control operation of the server system. In the operation example shown in this figure, the initial state is determined as follows. end system a
, b, and c all have multiple pieces of data to send to the server system, and upon receiving a low-priority free token, immediately send the data. It is assumed that the server system S is not currently in a congested state, but will be in a congested state when it receives one more frame. Next, the operation in this initial state will be explained with reference to FIG.

At time T1, the server system S is in a non-congested state, and the end system a transmits data addressed to the server system S. At this time, the priority P of the token is L (low), and the reserve priority R of the token is also L. Reserve priority R is used when priority P cannot be changed.
This is information set for making reservations for changes. on the other hand,
At this time, the server system S is in a data reception waiting state. At time T2, server system S receives data from end system a. At this time, the server system S becomes congested, and the server system S
requests data transmission with high priority.

At time T3, end system a removes the transmitted data and transmits a low priority free token. This low priority free token is received by end system b. Time point T4
Then, end system b, which has received the free token from end system a, transmits data. At time T5, the server system S receives data from the end system b and sets the reserve priority R of the token reserve field in the frame to H (high).

At time T6, end system b removes the transmitted data and transmits a high priority free token. At time T7, end system c is unable to transmit data due to the high priority free token. Therefore, the server system S obtains the transmission right and transmits the data. This performs congestion control. At time T8, server system S
removes the transmitted data and continues transmitting data with high priority unless the congestion condition is relieved.

At time T9, if the server system S is in a non-congested state, it transmits a high priority free token. At time T10, end system b changes the priority of free tokens from high to low. Time T1
1, end system c sends data upon receiving a low priority free token. This results in
Communication takes place between all end systems a, b, c.

In the above-described embodiments, a ring type network has been described as a token passing type network, but the present invention is not limited to this, and it goes without saying that a bus type network or the like may be used.

[0031]

As explained above, according to the congestion control method for devices in a network of the present invention, the congestion of devices such as server systems that establish connections with an unspecified number of parties can be controlled by controlling the logical link sublayer for each connection. Instead of controlling using the above protocol, the media control sublayer controls each device, resulting in the following effects. That is,
Since it is not necessary to set the maximum number of buffers for each connection, the structure of software that executes congestion control can be simplified. Also, when exchanging data with multiple connections at the same time, the number of buffers used in each connection can be increased or decreased depending on the traffic (data communication amount) of each connection, improving buffer usage efficiency. .

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a system using the method of the present invention.

FIG. 2 is a diagram showing the communication layer of OSI.

FIG. 3 is a conceptual diagram of communication buffer management between communication layers.

FIG. 4 is an explanatory diagram of communication buffer management in the logical link sublayer.

FIG. 5 is an explanatory diagram of congestion control in the logical link sublayer.

FIG. 6 is a flowchart showing a congestion control operation procedure of the server system.

FIG. 7 is a diagram illustrating an example of congestion control operation of the server system.

[Explanation of symbols]

1 Token passing network 11, 1
3. Devices with high data communication priority 12, 14,... Devices with low data communication priority

Claims (1)

[Claims] Claim 1: A plurality of devices are connected to a token-passing network in which only devices with transmission rights can transmit data, and the plurality of devices are connected to a media control sublayer of a data link layer of a communication control protocol. Data communication is performed with each other by setting a high priority for data communication for certain devices among the plurality of devices, and a higher priority for data communication for other devices among the plurality of devices. If the congestion state of the high-priority device is detected, the priority provided in the priority field in the frame of the medium control sublayer is set to high, so that the high-priority devices can communicate with each other. 1. A congestion control method for devices within a network, characterized in that the priority in the frame of the medium control sublayer is lowered after the device with the higher priority performs data communication and the device with the higher priority is no longer in a congested state.