JPH02162937A - Priority control system in medium access of lan - Google Patents

Abstract

PURPOSE:To attain the setting of priority of an optional number and to simplify a constitution and an operation by assigning a time slot group composed of plural consecutive time slots to priority and varying a time slot number included in one time slot group. CONSTITUTION:A time slot management node 3 managing a time slot is provided. The management node 3 forms a time slot group composed of plural consecutive time slots and priority number is given to each time slot group, the utilizing rate of the time slot inculded in each time slot group is measured and the time slot number included in the time slot group is adjusted based on the result of measurement. In the case of active state, the transmission of an idle time slot is awaited for a node. Time slots form all time slot groups and nodes recognize the priority of each time slot. For example, the idle time slot of priority j(1<=j<=n) is acquired and so long as the priority of the preserved information is (j), the node uses an idle time slot to send the information.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to medium access in local area networks (LANs) that allow multiple nodes to communicate with each other using one transmission medium. It concerns a method, and more specifically, when communicating a mixture of information that cannot tolerate a large transmission delay (such as audio and video) and information that can tolerate a relatively large transmission delay (such as data), the former is used to communicate the latter. The present invention relates to a method for accessing a medium by giving priority to

[Conventional technology]

Traditionally, local area networks (LANs)
It was often used for data communication between computers and between computers and terminals. Therefore, only data needs to be considered as the information to be communicated, and very strict conditions are not imposed on transmission delays associated with communication. However, even in data communication, the smaller the transmission delay, the better the service quality, so many studies and proposals have been made regarding delay reduction.

However, in recent years, there has been a demand for multimedia LANs. Multimedia means not only the data that has traditionally been handled between computers, but also audio and video transmission over a single LAN, making it compatible with all types of media communications. A new problem that arises as LANs become multimedia is how to transport information that requires immediacy, such as audio and video, with minimal delay time. Information that requires immediacy, such as audio and images, is meaningless if it arrives at the communication partner late, so strict delay characteristics are required.

On the other hand, with regard to data, it is preferable that the transmission delay be small, but in terms of immediacy, the requirements are smaller than for audio and images. Considering the above, there should be a difference in the handling within the LAN between information that requires immediacy and information that does not require immediacy. In other words, there is a necessity to treat the former with priority over the latter. This is called priority control, and it is becoming an essential function in multimedia LANs.

Orwell ring is a typical example of priority control.
(Orwelling). One reference for this example: R.M. Falconer, J.L.
Adams, co-author, “Overnian Protocol for
An Integrated Services Local
'Network', British Telecom Technology Journal, October 1985 (R,M,
Falconer and J.L.Ada+ms,”
Orwell: a protocol for an
integrated 5 services locala
lrletwork”, Br, Telecom Te
chnol, J., vol. 3. no, 4゜Oct
, 1985).

FIG. 9 is a diagram showing the hardware configuration of the Or%Iell ring, where 1 is a transmission node and 2 is a transmission medium. 1 and 2 are connected as shown to form a ring. There are an integer number of time slots (hereinafter referred to as time slots) on the ring, which circulate in the direction of the arrow.

The operation is as follows. Nodes that have information to send are waiting for free time slots to be sent to them from upstream. When a vacant time slot is acquired, the destination address, source address, and information to be transmitted are inserted into the time slot and transmitted to the next node. All nodes monitor the time slots sent from upstream. If the destination address of the sent time slot matches its own address, it understands that the time slot information is addressed to its own node and takes in the information. Then, the time slot is left vacant and transmitted to the next node.

In this way, communication using time slots is established. The medium access method in the Orwellian ring is basically a throttling method represented by the Cambridge ring. However, in the Cambridge Ring, the node that sends out information in a time slot goes around the ring and returns the time slot to empty, whereas in the Orwell Ring, the node that sends out the information in a time slot goes around the ring and returns the time slot to empty. The difference is that the node returns the time slot to an empty state.

Next, medium access priority control in the Orwellian ring will be explained. Each node is assigned a local block limit. There are two types of local block upper limit values: a high priority local block upper limit value and a low priority local block upper limit value.

The high priority local block upper limit value DH1 (i indicates the node number) and the low priority local block upper limit value DL are as follows:
There is a relationship of DHi>DLl. Each node also has high priority and low priority local block counters.

A node with high priority (or low priority) information can use an empty slot to transmit information, but 1
Each time a block is sent, the high priority (or low priority) local block counter is incremented. Then, the high priority counter value is DFI, and the low priority counter value is DL□
Once that is reached, the node enters a pause state and stops transmitting. When all nodes enter the pause state, all nodes are reset, the counter becomes O, and transmission can be resumed. Even if all nodes do not become paused, a reset will occur if all nodes become either paused or idle (the local block counter is below the upper limit, but there is no information to send). .

Whether all nodes have become paused or idle can be checked in a trial slot. In trial slot, trial bit=1. Vacant indication = empty, destination address = own address. A node in an idle or paused state sets its try bit to 1 whenever it has a free slot. Since the trial slot is free, other nodes can use it if they have information to send. If the trial slot returns empty, the node that sent the trial slot determines that all other nodes are idle or paused. Then, it sends a reset bit to notify all nodes of the reset.

When traffic is relatively slow, there are many idle nodes, so resets occur frequently. Therefore, information to be transmitted can be transmitted smoothly regardless of its priority level. However, when traffic becomes busy, the reset interval becomes longer. In the most crowded conditions,
A reset is not applied until the block Σ(DH engineering+1+l DL,) completes communication (N is the number of nodes). In this case, there is a low priority. D
Since H>DLL, the proportion of priority information can be said to be extremely small. That is, high-priority information is handled preferentially over low-priority information. The above is an outline of priority control in Orwell's ring.

[Problem to be solved by the invention]

The problems with the priority control in the conventional Orwellian ring described above are as follows.

■Only two priorities can be set.

■Node operation is complicated.

The first problem is that DH,>DLL is set. Now, suppose that a new intermediate priority is created to correspond to the three priorities, and its local block upper limit value DMi is D Hi > D Mi > D L
Suppose that it is set as i. If you do this, DHi
has priority over DMi, and DMi has priority over DL, so three priorities can be achieved. However, the size of each block is a problem. 1 block to 1
If it is configured with time slots, then DL is 1
Block.

That is, even for one time slot, DH1 is not reset until at least 100 blocks (that is, 100 time slots) have been cycled, which means that it takes a very long time.

The second problem will be explained. A state transition diagram of the nodes of the Orwell ring is shown in FIG.

As you can see from this diagram, states, actions, and judgments intersect in a complex way. One problem with such complexity is that it leads to heavy hardware/software implementation. As a result, not only is economic efficiency lost, but processing delays also increase. The second problem is the effect of malfunction or failure. For example, if the local block counter fails, a situation occurs in which slots are sent out endlessly and no reset occurs. The problem is that failure of one node can affect the entire network.

An object of the present invention is to solve such conventional problems and to provide an LA in which node operation is simple and priority control can be performed effectively.
The object of the present invention is to provide a priority control method for N medium access.

[Means to solve the problem]

In order to achieve the above object, the priority control method for medium access in a LAN of the present invention is such that a fixed length time slot circulates on a ring network consisting of a transmission medium connecting a plurality of nodes in a ring shape. The communication system that transmits information between the nodes using slots as transfer units has a time slot management node that manages the time slots, and the time slot management node is configured to manage time slots consisting of a plurality of consecutive time slots. Configure slot groups, assign a priority number to each time slot group, measure the utilization rate of the time slots included in each time slot group, and based on the measurement results, select the time slots included in the time slot group. A node that has information to send will adjust the number of timeslots and determine if the time slot sent from the prefix node is free and the priority number assigned to the time slot group to which that time slot falls When the priority number of the information to be transmitted matches the priority number of the information to be transmitted, the destination address, the source address, and the information to be transmitted are inserted into that time slot and transmitted, and the plurality of nodes monitor all the time slots and transmit the information to their own node. It is characterized by the fact that it captures and receives information addressed to it.

Further, in the priority control method for medium access of the LAN, the time slot management node measures the utilization rate of all time slots, and configures a time slot group only when the utilization rate is equal to or higher than a predetermined value. A priority number is assigned to each time slot group, and when the utilization rate is less than a predetermined value, the time slot group is not formed.

[Effect]

In the present invention, a time slot group consisting of a plurality of consecutive time slots is assigned to one priority level, another time slot group is assigned to a different priority level, and the time slots included in one time slot group are used. The number is variable.

That is, in the present invention, time slot groups are assigned for each priority, so when a node that has information to transmit receives an empty slot, it inserts information of the corresponding priority into that slot along with an address. and send.

All nodes monitor all time slots and capture and receive information destined for themselves. The time slot management node manages the number of time slots for each priority and measures the utilization rate (throughput rate). If the utilization rate of the high priority time slot group exceeds a certain value and the low priority time slot group is not zero, the time slot management node increases the number of high priority time slots.

Reduce the number of low priority time slots. This results in
The configuration is simple, effective priority control operation is possible, and an economical communication system that is resistant to failures can be realized.

〔Example〕

Hereinafter, one embodiment of the present invention will be described in detail with reference to one drawing.

FIG. 2 is a diagram showing the hardware configuration of a ring network to which the present invention is applied, in which 1 is a transmission node, 2 is a diagram showing a hardware configuration of a ring network to which the present invention is applied;
is a transmission medium, and 3 is a time slot management node. As explained in FIG. 9, time slots circulate on the ring an integral number of times. A node with sending information is
It captures an empty time slot, adds the destination address and source address to the information, inserts it into the time slot, and sends it out. On the other hand, all nodes monitor the time slots sent from upstream, and if the time slot is addressed to their own node, they take in the information contained therein.
Furthermore, if the time slot was sent by its own node, it is returned to an empty state and released. It basically operates in the same way as a Cambridge ring.

FIG. 1 is a diagram for explaining a priority control system in medium access showing an embodiment of the present invention.

(a) shows a state transition diagram (medium access method) at the time of information transmission in the node 1 shown in FIG. 2, and (b) shows the operation in the time slot management node 3 shown in FIG.

As shown in FIG. 1(a), when information to be transmitted occurs in the idle state, the state changes to active. In the active state, a node is waiting for free time slots to be sent to it. The time slots all form time slot groups, and each time slot group is assigned a priority order, so that a node can recognize the priority of each time slot. Assume now that an empty time slot of priority j (1≦j≦n) can be acquired and that the priority of the held information is j. Only in this case can a node use an empty slot to transmit information. moreover.

Remains active when there is information to send;
Otherwise, it transitions to an idle state. In this way, as is clear from the state transition diagram during Orwell ring transmission explained in FIG. 10, this embodiment can be greatly simplified.

As shown in FIG. 1(b), the time slot management node 3 manages the time slot utilization rate (throughput rate).
is measured (101). The measurement method will be described later. After understanding the utilization rate of each time slot group, it is necessary to refer to the table (created on the internal storage device of the time slot management node 3) (102) and change the configuration of the time slot group using the method described below. Determine whether or not there is (10
3). If it is not necessary, move on to the measurement operation of the tie 11 slot utilization rate again6. If necessary, display the change in the configuration of the time slot group, that is, the change in the number of time slots included in one time slot group ( 104).

Thereafter, the operation moves to measuring the time slot utilization rate again.

Next, the priority control mechanism will be explained.

FIG. 3 is a diagram for explaining the concept of the priority control mechanism in the embodiment of the present invention. Each of the figures (1) to (9) represents the size of a high-priority time slot group, the size of a low-priority time slot group, and their utilization rates within one frame. Note that one frame includes a fixed number of time slots. In Figure 3, the high priority time slots are about 5 of the total.
5%, and the low priority time slot group occupies about 45% of the total. Suppose there is 1 in 1 frame.
If 00 time slots are included, 55 time slots are assigned to high priority and 45 time slots are assigned to low priority. There is a clear demarcation between both priorities. How to establish this boundary will be described later.

The method for determining the time slot utilization rate (X) will be explained.

The time slot management node 3 knows the number of high priority time slots and their areas, and the number of low priority time slots and their areas. and.

All time slots passing through the time slot management node 3 are monitored, and the number of high priority time slots in use and the number of low priority time slots in use are counted. The counted value is divided by the number of time slots in the already set time slot group to obtain the time slot utilization rate (X). Note that if the value of X varies significantly from frame to frame, it is effective to use the average value of X as the utilization rate.

Figure 3 shows a group of high priority time slots (hereinafter referred to as H
) and low-priority time slots (hereinafter referred to as
(abbreviated as L) is shown qualitatively. (1
) indicates a case where both the utilization rates (X) of H and L are high. In this case, the time slot management node 3 increases the number of time slots for H and decreases the number of time slots for L by changing the boundary between H and L. (2)
, Perform the same operation in case (3), (
In 4), (5), and (6), the time slot utilization rate of H is moderate, so no operation is performed.For (7), the time slot utilization rate of H is low and the time slot utilization rate of L is low. Since the slot utilization rate is high, the number of time slots is decreased and the number of time slots of L is increased. (8), (
Regarding 9), the time slot utilization rate of H is low, but the time slot utilization rate of L is low.
Since the time slot utilization rate of 1 or more is below medium, the summary of 1 or more in which no operation is performed is as follows. When H is insufficient (utilization rate is high), L is eroded. If H has a medium utilization rate, the current status is maintained regardless of L. L erodes H only when L is insufficient and H has sufficient margin.

By increasing or decreasing the number of time slots in this way, the transmission of high priority information can be carried out more smoothly than the transmission of low priority information.

FIG. 4 summarizes methods for increasing and decreasing the number of time slots for each case of high priority time slot utilization rate and low priority time slot utilization rate. Here, set the time slot utilization rate to "high", "medium",
It is divided into three levels, ``low'' and ``high'' indicates that there is a lot of traffic with the corresponding priority and the access delay is large, and ``medium'' indicates that there is not a lot of traffic with the corresponding priority and at least "Low" indicates a case where the traffic of the relevant priority is small and the access delay does not increase significantly even if the number of time slots is reduced. ``High,''``medium,'' and ``low'' timeslot utilization cannot be expressed as a percentage.

It depends on the case. However, if the scale of the system is determined, each range can be roughly determined.

An example is shown below.

FIG. 5 is an example showing the relationship between time slot utilization rate and transmission delay. These are the results obtained by simulation under the following conditions.

Number of bits on the ring = 6000 Number of slots on the ring = 30 Number of bits in one slot = 200 Number of bits in one packet = 2000 As can be seen from Figure 5, the time slot utilization rate is 80.
%, the transmission delay increases rapidly.

Therefore, in this network, a usage rate of 80% or more may be considered "high" or "low".
and "moderate" are separated by 50%, for example. This depends on the range of increase/decrease in the number of time slots, but the range of increase/decrease is 10
This is because if it is set as %, neither side will be greatly affected by the change.

The above figure 4 shows the case where there are two priorities, but the 6th
The figure shows a case where there are three priorities. Figure 6 (a
) is the increase/decrease in time slots when the high priority time slot utilization rate is high, Figure 6(b) is the increase/decrease in time slots when the high priority time slot utilization rate is medium, and Figure 6(c) is the increase/decrease in time slots when the high priority time slot utilization rate is medium. ) shows the increase/decrease in time slots when the utilization rate of high priority time slots is low. By increasing or decreasing the number of time slots in this way, the size of the time slot group can be optimized, and the transmission of high-priority information can be performed more smoothly than the transmission of low-priority information.

FIG. 7 is a diagram showing a method for distinguishing between time slot groups of a plurality of priorities. (a) is the first method.

(b) shows the second method, and (c) shows the third method. In both cases, two priority orders are used as examples.

The first method is a configuration in which the range of the high-priority time slot group and the range of the low-priority time slot group are clearly indicated at the beginning of the frame (the hatched part). By referring to the hatched area, each node can determine that #1 to 4m is the range of high priority time slots, and #(m+1) to #n is the range of low priority time slots. Therefore, when it has high-priority information, it accesses the former, and when it has low-priority information, it accesses the latter for communication. The second method is
The start point (hatched part) of a high-priority time slot group and the start point (hatched part) of a low-priority time slot group are clearly marked as the beginning of the respective priority time slot groups. This is the configuration. Similarly in this case.

Each node can know the range of its respective priorities. The third method is a configuration in which the priority of each time slot is clearly indicated at the beginning (hatched part) of each time slot. In this case as well, each node can know the range of its respective priorities.

FIG. 8 is another configuration example of the present invention, and is a state transition diagram of a system in which a mode in which time slot groups are configured for each priority and a mode in which they are not configured are switched depending on the level of traffic.

The system starts in time slot group unconfigured mode. In this mode, time slot groups are not configured for each priority as described above. In other words, information of all priorities can be communicated without discrimination. In this state, the time slot management node 3 determines the time slot utilization rate (
x) is measured regularly. Then, unless X exceeds a certain value (α: 50%, for example) (x<α), the system remains in the time slot group unconfigured mode. However, if X exceeds α (x>α), the system transitions to the time slot group configuration mode and performs the operation described in FIG. is set by the time slot management node based on the time slot utilization rates collected so far.

〔Effect of the invention〕

As described above, according to the present invention, an arbitrary number of priorities can be set, and the configuration and operation of nodes can be simplified. Furthermore, since the time slot management node knows the time slot utilization rate for each priority, this information can be transmitted to each node to promote or restrict transmission from each node.

[Brief explanation of the drawing]

FIG. 1 is a diagram for explaining a priority control method showing an embodiment of the present invention, FIG. 2 is a diagram showing an example of the hardware configuration of a ring network to which the present invention is applied, and FIG. FIG. 4 is a diagram showing the concept of the priority IIJ control mechanism; FIG. 4 is a diagram showing a method for increasing/decreasing the number of time slots for each case of high priority time slot utilization rate and low priority time slot utilization rate; FIG. 5 is a diagram showing the relationship between time slot utilization rate and transmission delay, FIG. 6 is a diagram showing how to increase or decrease the number of time slots for each case of time slot utilization rate when the priority is 3, and FIG. FIG. 8 is a state transition diagram of a system showing another configuration example of the present invention; FIG. 9 is a diagram showing the hardware configuration of an Orwell ring;
FIG. 10 is a state transition diagram during information communication in the nodes of the Orwellian ring. 1: Node, 2: Transmission medium, 3: Time slot management node. (a) Figure (Part 1)

Claims (2)

[Claims] (1) In a communication system in which fixed-length time slots circulate on a ring-shaped network consisting of a transmission medium that connects multiple nodes in a ring, and information is transmitted between the nodes using time slots as transfer units. , a time slot management node that manages the time slots, the time slot management node forming a time slot group consisting of a plurality of consecutive time slots,
A priority number is assigned to each time slot group, the utilization rate of the time slots included in each time slot group is measured, and based on the measurement results, the number of time slots included in the time slot group is adjusted and transmitted. A node that has information to be transmitted must ensure that the time slot sent from the prefix node is free, and that the priority number assigned to the time slot group to which the time slot belongs is the priority number of the information to be transmitted. When a match occurs, the destination address, source address, and information to be sent are inserted into that time slot and transmitted, and the multiple nodes monitor all time slots and take in the information addressed to their own node. A priority control method in LAN medium access characterized by receiving. (2) The time slot management node measures the utilization rate of all time slots, and only when the utilization rate is greater than or equal to a predetermined value, forms a time slot group and assigns a priority number to each time slot group. 2. The priority control system for medium access in a LAN according to claim 1, wherein when the usage rate is less than a predetermined value, no time slot group is formed.