JP3394476B2 - Route selection method and device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a route selection method on a network, and more particularly to a route selection method capable of facilitating service provision on the network based on information on a moving average of traffic volume. The present invention relates to a selection method and device.

[0002]

2. Description of the Related Art In a service for distributing a large amount of data using a network, a considerable load is imposed on the network in terms of resources. In such a large amount of data distribution service, it is important that reliable delivery is more important than immediate delivery. Conventionally, in such a distribution service, after receiving a request,
The request information is temporarily stored in a secondary storage device such as a hard disk, and a route is selected to deliver the request by selecting a route in a time zone such as midnight when the network load is likely to be relatively low. .

[0003]

However, the conventional route selection system has the following problems. That is,
In general, the load on the network fluctuates depending on the time and the time for each route, etc., and the load is not necessarily low even in the midnight hours. Depending on the load situation, there is a shortage of network resources. Depending on the circumstances, it may not be possible to deliver even at midnight.

Further, when a large amount of data is distributed via the selected predetermined route, the load on the network associated with the route is not always low. Therefore, by distributing a large amount of data, the network associated with the route. However, there is a problem in that the resources of the above will be occupied and distribution for other requests may become impossible.

(Object) It is an object of the present invention to provide a route selection method and device capable of selecting a route in which network resources can be secured in response to a request for a distribution service.

Another object of the present invention is to provide a route selecting method and apparatus for predicting the traffic volume of each route and selecting an optimal route with less traffic.

[0007]

The present invention provides a service by predicting the traffic volume of each route from two moving averages of a long-term moving average and a short-term moving average based on measured values and their data. A method and apparatus for selecting the most suitable route for

The route selection method of the present invention is a route selection method in a network for transmitting data via a plurality of nodes, collecting the traffic volume of each route for each node, and collecting the data from a predetermined node. At the time of transmission, the change in the traffic volume of each route is predicted from two moving averages, the long-term moving average and the short-term moving average calculated from the collected traffic volume, and the data is based on the traffic volume and the prediction result of the change in the traffic volume. It is characterized in that the route to the destination most suitable for transmission is selected. Also, in the route selection method in a network that transmits data via multiple nodes, the traffic volume of each route for each node is collected and calculated from the collected traffic volume when transmitting data from a predetermined node. The change in the traffic volume of each route is predicted from the two moving averages of the long-term moving average and the short-term moving average, and the traffic amount based on the average value of the traffic amount and the standard deviation value is predicted, and based on the both prediction results. The feature is that a route to a destination with less traffic is selected. Further, in each of the route selecting methods, the predetermined node stores transmission data in a storage device in response to a data transmission request, and starts selecting the route in a time zone when traffic is expected to be low. You can

Further, the route selection device of the present invention is a route selection device in a network for transmitting data between a plurality of nodes, and collects and stores the traffic volume of each route for each node on the network. A data collection unit and a predictive traffic calculation unit that calculates two moving averages of a long-term moving average and a short-term moving average from the collected traffic amount, and calculates a traffic amount predicted from the traffic amount and the two moving averages. And a route selection request accepting unit that selects a route with less traffic to a data destination based on the predicted traffic volume according to a request from a predetermined node. The predicted traffic calculation unit may calculate the predicted traffic amount from the traffic amount based on the average value and the standard deviation value of the traffic amount and the prediction of the change in the traffic amount by the two moving averages. Characterize.

More specifically, the route selection device of the present invention is a data collection unit (200 in FIG. 2) that collects and stores route measurement traffic data from each node on the network. Using the long-term moving average and short-term moving average calculated based on the data obtained, the predicted traffic calculation unit (210 in FIG. 2) that calculates the predicted traffic volume, and the selection of the route with less traffic from the node providing the service A route selection request receiving unit (22 in FIG. 2) that receives the request.
0) and (Fig. 2).

(Operation) Predicted traffic of each route calculated from the average value and standard deviation value of the measured traffic amount assuming that the collected traffic amount for each route of each node follows a normal distribution. Volume, and two moving averages of a long-term moving average and a short-term moving average are calculated from the measured traffic amount to correct the measured traffic amount or the predicted traffic amount, and based on the corrected measured traffic amount or the predicted traffic amount. Select the most suitable route for the service.

A route selection request source, that is, a service providing node that is a node that provides a network service,
By using the route selected by the method,
It is possible to minimize problems such as resource shortage due to the influence of network traffic, suppress burst or local traffic increase in the network, and facilitate service provision.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION (Description of Configuration) An embodiment of a route selection method and apparatus using a moving average according to the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing an embodiment of the present invention. This embodiment has a function of selecting a route by traffic prediction, and adopts a network configuration suitable for distribution of a large amount of data.

In the present embodiment, it is composed of a network configuration node 100 which constitutes a network and a network management node 110 which manages the traffic condition of the network. The network configuration node 100 is composed of a plurality of nodes A to H and route connections (routes) a1, a2, to g1 that connect between the respective nodes, and in this example, a service provided by the node A. The providing node 120. Further, the network management node 110 receives the traffic data of the measured traffic volume from each of the nodes A to H, predicts the traffic volume based on the collected traffic data, and informs the service providing node 120 of the reservation information of the route. It has a traffic reservation function to provide and notify.

FIG. 2 is a diagram showing a detailed configuration of the network management node 110 having the traffic prediction function of this embodiment. The network management node 110 receives a report of traffic data from each of the nodes A to H, collects and updates data, and a predicted traffic calculation unit 210 that performs a prediction calculation of a predicted traffic amount from the collected traffic data. , A route selection request receiving unit 220 having a function of selecting an optimal route and notifying reservation information of the route in response to the route selection request based on the predicted traffic volume.
It consists of and. The outline of the configuration of each part is as follows.

The data collecting unit 200 includes a collected data receiving unit 201, a data updating unit 202, and a data extracting unit 20.
3 and 3. The predicted traffic calculation unit 210 includes a data extraction unit 211, a normal distribution group / center value calculation unit 212, a moving average calculation unit 213, and a correction value calculation unit 2.
14 and a predicted value calculation unit 215. The route selection request receiving unit 220 includes a request receiving unit 221 and a route indexing unit 2
22, a predicted value request unit 223, and a reservation status confirmation unit 224
And an optimal route selection unit 225 and a reservation update unit 226.

(Description of Operation) Next, the operation of the present embodiment will be described in detail with reference to FIGS.

Each of the nodes A to H constituting the network in the network configuration node 100 receives the traffic data for each of the routes a1, a2, to g1 every hour,
Report to the data collection unit 200 of the network management node 110.

In the data collection unit 200 of the network management node 110, the collected data reception unit 201 receives the traffic data for each route from each node, and the data update unit 202 stores the traffic data for each route of each node. A database consisting of data and the average value and standard deviation of traffic calculated based on this data is created, and updated and stored as actual measurement data.

FIG. 3 is a diagram showing a data structure of the actual measurement data and the like stored as a database. The data structure in which the traffic data, the average value, and the standard deviation value for each route for each node are recorded on an hourly basis for each day of the week is adopted.

The service providing node 120 is connected to a plurality of terminals (not shown), etc., and always accepts the user's request for the distribution service from the terminals etc. When the request for the distribution service is accepted, the network management node 1
It requests the route selection requesting unit 220 of 10 to select a route that is predicted to have the lowest traffic suitable for transmitting the distribution service information.

In the network management node 110, the request acceptance unit 221 of the route selection request acceptance unit 220 accepts the route selection request from the user, and the route indexing unit 222 becomes the delivery destination node of the requested delivery data. The predicted value requesting unit 223 requests all the related routes and requests the predicted traffic calculating unit 210 to calculate the predicted traffic amount data.

In the predicted traffic calculation unit 210, the data extraction unit 211 obtains the service information from all the route related to the request, that is, the service providing node 120 among the collected route data shown in FIG. The actual data of all the routes 130 to the destination node (for example, the node H) is taken out.

Normal distribution group and center value calculation unit 21
In No. 2, the measured data of the traffic volume at the time of reception extracted by the data extraction unit 211, the average value and the standard deviation value are set and recorded, and the fluctuation range of the traffic is considered as a normal distribution, and the center value and the group number are calculated. To record.

For example, a 24-division group of normal distribution and its center value for the actual data at the time of request acceptance are obtained. The center value and the group number are obtained by the following method. That is, the fluctuation of traffic on a certain route is considered as a normal distribution, and this is modeled to calculate the predicted traffic amount. The normal distribution has an average value T0 (MV) as shown in FIG.
And standard deviation (σ), it can be created by calculating a numerical value group based on random numbers. As a general formula, the center value of each group is, in the case of 24 divisions, Tn = T0 + 0.5 (n-12.25) × σ, where T0 (MV) = 135.09 (for example, the average value of traffic) σ = 8.46 For example, a group with group number n = 6 (group 6)
Is T6 = 135.09 + 0.5 (6-12.25) × 8.
46 = 108.6525. Therefore, in the group 6, the center value is 108.6525, and the range of the upper limit value and the lower limit value with respect to the center value can be calculated as the center value ± 1.0575.

Next, the correction value for the center value is
It is calculated from the long-term moving average and short-term moving average of past data, and corrected by this.

First, the moving average calculation unit 213 determines the long-term moving average, which is the average value of data for 24 hours in one hour units from 24 hours before the same time one week before the request is received, and when the request is received. The short-term moving average, which is the average value of the data for 3 hours in 1-hour units from 3 hours before, is calculated and stored together with the data. The correction value calculation unit 214 calculates and stores a half of the difference between the long-term moving average and the short-term moving average as a correction value.

From the calculated center value and the correction value, the prediction value calculation unit 215 adds the correction value to the center value for the actual data at the time of request acceptance to calculate the corrected predicted traffic value. Further, from the standard deviation value for the measured traffic value at the time of request reception, the predicted upper limit value and the predicted lower limit value of the corrected predicted traffic value are obtained and predicted as the traffic fluctuation range after the request reception, and the traffic volume prediction information is used. The prediction upper limit value and the prediction lower limit value are stored as certain prediction values.

FIG. 5 is a diagram showing a data structure in which the above data is stored as long-term short-term moving average calculation data for each route. The data structure mainly includes a center value, a group number, a short-term and long-term moving average, a correction value, a prediction upper limit value, a prediction lower limit value, and the like. It can be said that the predicted traffic value tends to increase when the long-term moving average is higher than the short-term moving average, and tends to decrease when the long-term moving average is lower than the short-term moving average.

In the route selection request receiving unit 220 of the network management node 110, when the predicted values of all the related routes have been obtained by the predicted traffic calculation unit 210, the reservation status confirmation unit 225 requests the predicted value request unit 223. Through FIG.
Enter and confirm the information shown in, and select the optimum route selection unit 225.
Is the reservation route if the route already notified as the optimum route in response to the similar request from the service providing node 120,
When the route is selected, the priority to be selected is lowered, the load is not concentrated on the same route, and the routes that are predicted to have less traffic are combined to determine the route. It is preferable to specifically determine the route based on the predicted value. For example, a route with a small prediction lower limit value and a route with a small difference between the prediction upper limit value and the prediction lower limit value are selected and the reserved route is selected. And The determined reservation route is recorded in the reservation updating unit 226, and is notified from the reservation updating unit 226 to the service providing node 120 via the request receiving unit 221 as the optimum route.

FIG. 6 is a diagram showing a data structure of a reserved route recorded as route reservation status data in the reservation updating unit 226. In the figure, as the routes from the service providing node 120 to the destination node, the routes a1, b1, ... G1 are for the request at 1 o'clock, and the routes a1, b2 are for the request at 23:00. It is shown that ... Is selected and recorded as the reservation route and is notified to the service providing node 120.

The service providing node 120 establishes a connection to the route consisting of the route notified from the network management node 110 and sends the distribution data information to a predetermined destination node. If the service providing node 120 cannot establish a connection to the node, the requested distribution data information is stored in the secondary storage device, and the request is issued to the user who issued the request. Notify the acceptance and end the processing after acceptance. afterwards,
Retry is performed at a fixed timing to restart the process. If a connection cannot be established even after repeating such a retry operation a certain number of times, the network management node 110 is notified that the notified route cannot be used, and the route selection is requested again. If the distribution data information cannot be sent to the destination node even after the route selection request is made a specified number of times, the delivery failure is notified to the terminal designated by the user's request.

The service providing node 120 requests the network management node 110 to release the reserved route when the connection of the route (route) determined by the network management node 110 is established and the data has been sent to the destination node. . In addition, the user is requested to notify the designated terminal of the completion of delivery. Network management node 1
When the route requested to be released is the reserved route, 10 releases the route.

(Other Embodiments) In the above embodiments, the long-term moving average and the short-term moving average are calculated by averaging over 24 hours and 3 hours, respectively, but this period is one month and one day, respectively. be able to. Also, the number of divisions of the normal distribution and the like can be changed as appropriate. Also in this embodiment, the operation at each node is the same. That is, each node 100 of the network reports the traffic data for each route 130 to the data collection unit 200 of the network management node 110 every hour. The data collection unit 200 of the network management node stores the reported traffic data of each route of each node, the actual measurement data, the average value, and the standard deviation in the same manner as in FIG.

The predicted traffic calculation section 210 of the network management node 110 extracts the collected data of all the routes based on the collected route data, and receives the request acceptance request 1
Calculate the 1-month moving average one month ago and the 1-day moving average. Next, a group of 30 normal distributions with respect to the actual value at the time of request reception and its center value are obtained. 1 month moving average, 1
One half of the difference of the daily moving average is used as the correction value. From these values, the daily fluctuation range of predicted traffic is obtained. When the predicted traffics of all the routes are obtained, the route selection request acceptance unit 220 of the network management node 110 sets 1
Notify each node of the preferred route of day.

Each node 100 preferentially selects the route notified from the network management node 110. When a route failure occurs, the network management node 11
0 newly calculates the priority selection route and notifies each node 100.

In the above embodiments, each data is described as an example of being stored in a table format. However, such storage means is not necessarily required, and the data shown in FIG. 3, FIG. 5 and FIG. It goes without saying that the minimum information among them can be stored in each unit. Further, in the above embodiment, the average value and the standard deviation value are calculated from the measured traffic amount and the traffic amount is predicted to follow the normal distribution. This can be done by adding a correction based on the moving average and short-term moving average.

[0039]

According to the present invention, there are two movements, that is, the measured traffic volume for each route of each node, or the traffic volume based on its normal distribution, and the long-term moving average and the short-term moving average calculated from the measured traffic volume. Since it is configured to predict the traffic volume of each route by averaging, it is possible to predict the more accurate traffic volume for each route,
It is possible to select a route that is predicted to have low traffic and avoid concentrating traffic on a specific route, so even when providing a service that places a considerable load on the network, lack of resources, etc. It is possible to minimize the impact on other services due to the inability to provide the service due to, or occupying resources, suppress the burst or local traffic increase of the network, and provide stable service, It is possible to optimally control the traffic volume of each route of the network.

Further, even if each node accepts a service request, the node stores the information in the storage device so as to start the transmission of information in a time zone when the traffic is expected to be low, so that the traffic is high. Even if a service request is accepted at the time, it is not necessary to immediately send information, and the network is not burdened.

Further, in order to predict the traffic at the specified time or time zone, the service is realized by selecting the route with low traffic not only in the specific time zone but also in any time zone, so that the information is delivered. It is possible to expand the range of services without having to limit the traffic to only the nighttime zone where the traffic is expected to be low.

[Brief description of drawings]

FIG. 1 is a diagram showing a network configuration of an embodiment of a route selection method and device of the present invention.

FIG. 2 is a diagram showing a functional block configuration of the present embodiment.

FIG. 3 is a diagram showing a data structure of traffic data according to the present embodiment.

FIG. 4 is a diagram showing a normal distribution of traffic.

FIG. 5 is a diagram showing a data structure of long-term short-term moving average calculation data according to the present embodiment.

FIG. 6 is a diagram showing a data configuration of route reservation status data according to the present embodiment.

[Explanation of symbols]

100 network configuration nodes 110 network management node 120 service providing nodes 130 ways 200 Data Collection Department 201 Collected data reception section 202 Data update unit 203, 211 Data extraction unit 210 Predicted Traffic Calculation Unit 212 Normal Distribution Group and Center Value Calculation Unit 213 Moving average calculator 214 Correction value calculation unit 215 Prediction Value Calculation Unit 220 Route selection request reception section 221 Request reception unit 222 Directional indexing section 223 Prediction value request unit 224 Reservation status confirmation section 225 Optimal route selector 226 Reservation Update Department

Claims (5)

(57) [Claims] 1. A route selection method in a network for transmitting data via a plurality of nodes, wherein the traffic amount of each route for each node is collected, and the traffic amount collected when transmitting data from a predetermined node. Based on the difference between two moving averages of the long-term moving average and the short-term moving average calculated from the above, the change in the traffic volume of each route is predicted, and the transmission destination with less traffic is calculated based on the prediction result of the traffic volume and the change in the traffic volume. A route selection method characterized by selecting routes up to. 2. A route selection method in a network for transmitting data via a plurality of nodes, wherein the traffic amount of each route for each node is collected, and the traffic amount collected when transmitting data from a predetermined node. While predicting the change in the traffic volume of each route based on the difference between the two moving averages of the long-term moving average and the short-term moving average calculated from, predicting the traffic volume based on the average value and standard deviation of the traffic amount, A route selection method comprising selecting a route to a transmission destination with less traffic based on both the prediction results. 3. The predetermined node stores transmission data in a storage device in response to a data transmission request, and starts selecting the route in a time zone in which traffic is expected to be low. The route selection method according to Item 1 or 2. 4. A route selection device in a network for transmitting data between a plurality of nodes, and a data collection unit that collects and stores the traffic volume of each route for each node on the network, and the collected traffic. Two moving averages, a long-term moving average and a short-term moving average, are calculated from the traffic volume, and the traffic volume and the two moving averages are calculated .
A predicted traffic calculation unit that calculates a predicted traffic amount based on the difference, and a route selection request that selects a route with less traffic to the data destination based on the predicted traffic amount requested by a predetermined node. A route selection device comprising: a reception unit. 5. The predicted traffic calculation unit calculates the predicted traffic based on a traffic amount based on an average value and a standard deviation value of the traffic amount and a change in the traffic amount based on a difference between the two moving averages. The route selection device according to claim 4, wherein the amount is calculated.