JP4709263B2 - Network bandwidth calculation method, system and program - Google Patents

Description

  The present invention relates to a technique for calculating a bandwidth necessary for keeping a call loss rate below a certain level in a network environment where different types of applications require different bandwidths, and in particular, a mixture of applications requiring two or more different bandwidths. However, the present invention relates to a technique suitable for efficiently calculating a bandwidth necessary for satisfying a call loss rate equal to or less than a specified value in a network that performs admission control for each call occurrence (accommodates multi-traffic).

  As a technique for calculating a frequency bandwidth in a network, for example, “Erlang-B formula” is known. This “Erlang-B formula” is used to calculate the number of lines necessary to satisfy the call loss rate below the specified value when the required bandwidth of the application is one type, and when all calls use the same bandwidth. Applicable.

  For example, in a rejection system with K lines, if the traffic density is ρ, the call loss rate B (K) representing the ratio of calls rejected is given by the following equation (1). It is called “Erlang-B formula”. In Equation 1, the required bandwidth can be calculated by obtaining the number K of lines necessary to achieve a call loss rate below the specified value and multiplying the obtained number K of lines by the bandwidth per line.

  Further, a “multiple call required bandwidth calculation method” that can be applied to an application that requires two or more different bandwidths by extending the “Erlang-B formula” is also known.

In this “multiple call required bandwidth calculation method”, the required bandwidth is calculated by giving the call volume for each bandwidth type. Therefore, the number of completed calls for each bandwidth type (capture the bandwidth when a new call is accepted). Calls) and the number of incomplete calls (calls whose bandwidth could not be captured when a new call was accepted) and average bandwidth used (obtained at any of a plurality of times T _n within the observation period T) it is necessary to average value of the used bandwidth value) and keep the average number of connections (and store any mean value of the connection number of calls determined in a plurality of time T _n) a plurality of information such as in the observation period T retained.

For example, consider that N applications are added to a network in a rejection system with K lines. Considering the greatest common divisor of the bandwidth required by each application as the bandwidth of the basic call, and assuming that the bandwidth required by each application is a _i times the bandwidth of the basic call, the call loss rate B _n (K) of application n is when the traffic intensity application i and Er _i, given by: expression 2, this is referred to as "multi-dimensional Erlang-B formula." This “multidimensional Erlang-B formula” is also described in Non-Patent Document 1, for example.

In this way, the call loss rate B _n (K) of the application n is expressed by the above equation 2, and the number of lines K required to achieve a call loss rate equal to or less than a predetermined value determined for each application is calculated. it can. Then, the bandwidth required for the network can be obtained by multiplying the calculated number K of lines by the bandwidth of the basic call.

  Conventionally, in a bandwidth design of a telephone network represented by a PSTN (Public Switched Telephone Network) network, it is necessary to consider only a telephone call, and therefore, a necessary bandwidth has been calculated by the Erlang-B formula.

  However, due to the recent progress of multimedia services, improvement of network technology, etc., it is common for various applications to be mixed in the network, and multiple applications with different bandwidths exist in a single network (network). There are many more.

  In such a case, it is not possible to apply a conventional required bandwidth calculation method based on the Erlang-B formula, which can be applied when all calls use the same bandwidth.

  Further, in the required bandwidth calculation method based on the multidimensional Erlang-B formula for calculating the required bandwidth for multi-source calls having different bandwidth requirements, the number of completed calls and the number of incomplete calls, and the average used bandwidth and average for each bandwidth type Since it is necessary to store and hold information such as the number of connections, adding services increases the load on network operations.

  In addition, since it is necessary to hold traffic information for each band type, the number of band types increases and the required bandwidth becomes continuous. It may be an obstacle to adding new services.

  As described above, conventionally, there is a “multidimensional Erlang-B method” as a technique in which the “Erlang-B method” as a necessary bandwidth calculation technique in the case of one type of application is extended to a plurality of types of applications. In the “multidimensional Erlang-B system”, it is necessary to store and hold call information for all types of applications, and as a result, the amount of information to be stored becomes enormous as the bandwidth type increases.

Hideaki Takagi, Hitoshi Yoshino, Naoto Matoba, Mitsuhiro Higashi, “Calculation of required frequency bandwidth in next-generation mobile communication systems,” IEICE Transactions B Vol. J89-B No. 2 pp. 135-142, 2006.

  The problem to be solved is that, in the technique of calculating the necessary bandwidth using the conventional multidimensional Erlang-B method, if there are applications that require different bandwidths in the same network, Since it is necessary to hold traffic information for each band, the amount of information to be stored becomes enormous as the type of band increases.

  The object of the present invention is to solve these problems of the prior art, enable bandwidth calculation with a small amount of information, and increase the amount of information used for bandwidth calculation even when the types of required bandwidth increase. It is possible to greatly reduce the operation cost of the network.

In order to achieve the above object, in the present invention, when there is an application requesting a plurality of bands in the network, when calculating a band necessary for achieving a call loss rate below a certain rule, traffic information for each band is obtained. The total number C of calls (completed calls) that could be band-captured and the total number C ′ of calls (band-completed calls) that could not be band-captured during a certain observation period T without being used. The total value B of the required bandwidth of each call of the total number C that could be obtained, the total value B ′ of the required bandwidth of each call of the total number C ′ that could not be acquired, the number of connected calls L _C , and the used bandwidth value L _BW Or the total number C of completed calls, the total number C ′ of incomplete calls, the number of connected calls L _C , the used bandwidth value L _BW , and the total value C of the requested bandwidth of each completed call. Or the total number C of completed calls and the total number C 'of incomplete calls Connection call number L _C, calculates a necessary bandwidth using four information used bandwidth value L _BW, among a plurality of types of applications, similar to that only two types are present, the average amount of information about the call The bandwidth required to make the call loss rate below the specified value is calculated using

  According to the present invention, since only the maximum six pieces of information are used for calculating the necessary bandwidth, it is possible to keep the cost associated with the bandwidth calculation constant even when the type of required bandwidth increases. The operating cost can be greatly reduced.

  The best mode for carrying out the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of a network bandwidth calculation system according to the present invention, and FIG. 2 is an explanatory diagram showing the concept of bandwidth amount and overflow bandwidth amount used in the network bandwidth calculation system in FIG. FIG. 4 is an explanatory diagram showing a comparative example of the amount of information required in the network bandwidth calculation system in FIG. 1 and the amount of information required in the prior art, and FIG. 4 is a diagram according to the present invention by the network bandwidth calculation system in FIG. FIG. 5 is a flowchart showing a second processing operation example according to the present invention by the network bandwidth calculation system in FIG. 1, and FIG. 6 is a flowchart showing the present invention by the network bandwidth calculation system in FIG. FIG. 7 is a flowchart showing the third processing operation example, and FIG. 7 shows another configuration example of the network bandwidth calculation system according to the present invention. Block diagram, FIG. 8 is a flowchart, Figure 9 showing a processing operation example according to the present invention by the network bandwidth calculation system in FIG. 7 is an explanatory diagram showing an effect example by the network bandwidth calculation system in FIG.

  1 includes a CPU (Central Processing Unit) and a main memory, a display device such as a CRT or LCD, an input device such as a keyboard or a mouse, an HDD, or the like. It consists of a computer configuration equipped with an external storage device, and after installing a program or data recorded in a storage medium such as a CD-ROM via an optical disk drive device or the like in the external storage device, this external storage device is transferred to the main memory. A function for executing each processing according to the present invention is implemented by processing by the reading CPU.

In other words, the network bandwidth calculation system 1 includes a call control unit 2, a call count / bandwidth calculation unit 3, a connected call count / use bandwidth calculation unit 4, and first and second application selections as means for executing computer processing by a program. 5, number of completed calls N ₁ , N ₂ , number of incomplete calls N ′ ₁ , N ′ ₂ calculation unit 6, hold time h ₁ , h ₂ calculation unit 7, call volume Er ₁ , Er ₂ calculation unit 8, necessary A band calculation unit 9 is provided.

  With such a configuration, when there is an application that requests a plurality of bands in the network 10, the network band calculation system 1 determines the traffic required for each band when calculating the band required to achieve a call loss rate below a certain standard. Without using the information, as in the following first to third embodiments, the bandwidth required to make the call loss rate equal to or less than the specified value is efficiently calculated.

For example, in the first embodiment, in a certain observation period T, the number of calls that have been able to capture a band when a new call is accepted (the number of completed calls) C, and the band is captured when a new call is accepted. The total number of calls that could not be completed (the number of incomplete calls) C ′, and a value (bandwidth amount) B obtained by integrating the requested bands of each completed call of the total number C that was able to capture the bandwidth when a new call was accepted , A value obtained by accumulating the requested bandwidth of each incomplete call (total overflow amount) B ′ of the total number C ′ that could not capture the bandwidth at the time of accepting a new call, and the number of calls connected to the network 10 (number of connections call) and L _C, values of bandwidth being used by the network 10 using six information (used bandwidth) L _BW, further, among the plurality of types of applications, and that only two are present Approximately use only average amount of information about the call Te to calculate the bandwidth required to below a specified blocking probability.

In the second embodiment, the number C of completed calls, the number of incomplete calls C ′, the number of connected calls L _C , the used bandwidth value L _BW , and the total value C of the requested bandwidth of each completed call of the total number C are five. Using information, it approximates that there are only two types of applications among multiple types of applications, and uses only the average amount of information about the call to calculate the bandwidth required to make the call loss rate below the specified level .

In the third embodiment, two types of applications are used among four types of applications by using four pieces of information, that is, the number of completed calls C, the number of incomplete calls C ′, the number of connected calls L _C , and the used bandwidth value L _BW. Only the average amount of information related to the call is used, and the bandwidth required to make the call loss rate below the specified value is calculated.

First, <First Example> will be described. Incidentally, including this first embodiment, the following, in the second and third embodiments, as connection call number L _C, any average connection number of calls determined in a plurality of time T _n in the observation period T The average number of connected calls that is a value is used, and the average used band value that is an average value of the used band values obtained at any of a plurality of times T _n within the observation period T is used as the used band value L _BW .

  In the network 10, when applications that require two or more different bands coexist, the band information is lost because it is not possible to know which band has been accepted only by the number of completed calls or the number of incomplete calls. The necessary bandwidth cannot be calculated.

  Therefore, in this example, a bandwidth amount B and an overflow bandwidth amount B ′ are newly introduced as indicating the bandwidth information of the call. This makes it possible to use the number of calls and the bandwidth information without leaking them.

  The concept of the bandwidth B and the overflow bandwidth B 'used in this example is shown in FIG. The bandwidth amount B is a value obtained by integrating the requested bandwidths of the completed calls of the total number C that can capture the bandwidth at the time of accepting a new call, and the overflow bandwidth amount B ′ captures the bandwidth at the time of accepting a new call. This is a value obtained by integrating the requested bandwidths of each incomplete call of the total number C ′ that could not be obtained. In this way, in this example, by introducing the concept of bandwidth amount B and overflow bandwidth amount B ′, it is possible to quantitatively represent the bandwidth shortage of the network that cannot be seen from the number of lost calls. become.

In the first embodiment, the above-described bandwidth B and overflow bandwidth B ′, the number of completed calls C and the number of incomplete calls C ′ that have been generally acquired in the network from the past, and the average number of connected calls L _{C. From} the average used bandwidth value L _BW , a bandwidth necessary for making a call loss rate equal to or lower than a predetermined specified value is calculated.

However, the recording interval information is T, the completed call number C and not completed call number C 'is the number that have occurred within the period T, the mean used bandwidth value L _BW and average connection call number L _C is for each period T It is a measured value.

  The feature of the first embodiment is that when there are applications that require two or more different bands on the network, it is considered that there are only applications that require two bands, and the necessary band is calculated. is there.

  As a method of selecting the two bands, it is necessary to use a virtual band value obtained by dividing the band amount B by the number C of completed calls, that is, a value larger and smaller than the average band. For example, as an example of how to select two bands, the maximum band and the minimum band of an application that joins the network may be used.

The bandwidths of the two applications are set as BW ₁ and BW ₂ . An application with a bandwidth of BW ₁ is called a first application, an application with BW ₂ is called a second application, the number of completed calls of the first application is N ₁ , the number of incomplete calls is N ′ ₁ , The number of completed calls is N ₂ and the number of incomplete calls is N ′ ₂ , and the number of completed calls N ₁ and N ₂ and the number of incomplete calls N ′ ₁ and N ′ ₂ are obtained by the following relational expression:

In the equation (3), the completed numbers N ₁ and N ₂ and the incomplete numbers N ′ ₁ and N ′ ₂ are unknown numbers (values to be obtained), and the bandwidths BW ₁ and BW ₂ of the first and second applications are set. The value, the bandwidth B, the overflow bandwidth B ′, the number of completed calls C, and the number of uncompleted C ′ are observed values.

For example, it is assumed that there is an application requesting five types of bands of 64 kbps, 1 Mbps, 5 Mbps, 10 Mbps, and 100 Mbps. At this time, only the values of the minimum bandwidth and the maximum bandwidth of 64 kbps and 100 Mbps are used, and BW ₁ = 64 kbps and BW ₂ = 100 Mbps may be used. Here, since information regarding three of 1 Mbps, 5 Mbps, and 10 Mbps is not used but only the statistics as a whole is used, the amount of information to be held is small compared to the multidimensional Erlang-B formula.

Next, h _{1 represents} the hold time of the first application (the time that the first application occupies the band), and h _{1 represents} the hold time of the second application (the time that the second application occupies the band). _{2 and} the hold times h ₁ and h ₂ of the first and second applications are obtained by the following relational expression 4.

In the above formula 4, h ₁ and h ₂ (holding times of the first and second applications) are unknown numbers (values to be obtained), and L _C (average number of connected calls) and L _BW (average use band value) Is an observed value, and T (period) is a set value.

From the N ₁ , N ₂ , N ′ ₁ , N ′ ₂ , h ₁ , and h ₂ obtained by the above equations 3 and 4, the call volumes Er ₁ and Er ₂ of the first application and the second application are obtained. From the definition of call volume, the following formula 5 is used.

Then, using the call volumes Er ₁ and Er ₂ of the first and second applications calculated in this way, the Erlang-B equation represented by the following equation 6 or the multidimensional Erlang-B equation represented by the following equation 7 is used. The number of lines K ₁ and K ₂ required to set the call loss rates B ₁ (K) and B ₂ (K) for the first and second applications calculated by either of them to a predetermined value or less are calculated. Using the calculated number of lines K ₁ , K ₂ and the bandwidths BW ₁ , BW _{2 of} the first and second applications, the bandwidth necessary for making the call loss rate in the network 10 equal to or less than a predetermined value BW _des is calculated.

As a procedure for calculating the bandwidth BW _des , the number of lines K ₁ and K ₂ calculated by the Erlang-B equation and the bandwidths BW ₁ and BW _{2 of} the first and second applications are used. 1, every second application, call loss probability was calculated bandwidth _{BW des1, BW DES2} required to be less than the predetermined value, to calculate the total value of each band _{BW des1, BW DES2,} band BW Step calculated as _des, or the larger the number of lines of the multi-dimensional Erlang-B line number _K 1 calculated by equation, _{K 2,} first, the band _BW 1, BW ₂ of the second application One of the procedures for calculating a value obtained by multiplying the greatest common divisor as the band BW _des is used.

  Next, <Second Example> will be described. The feature of the second embodiment is that it is a bandwidth calculation technique that does not use the overflow bandwidth amount B 'used in the first embodiment. In the second embodiment, compared to the first embodiment, the amount of information that needs to be retained is smaller by the amount that the overflow bandwidth amount B ′ is not used, but the calculation accuracy is higher than that of the first embodiment. Is also inferior.

Similar to the first embodiment, symbols are placed, and the necessary bandwidth is calculated assuming that there is only an application that requires two bandwidths. First, the number of completed calls and the number of incomplete calls (N ₁ , N ₂ , N ′ ₁ , N ′ ₂ ) of the first and second applications are obtained by the following relational expression (8).

Then, the holding time of the first application h _1, the holding time of the second application h ₂ Distant, by the following Equation 9 relationship, first, holding time h ₁ of the second _application, h ₂ is determined.

Then, the number of completed calls (N ₁ , N ₂ ) and number of incomplete calls (N ′ ₁ , N ′ ₂ ) of the first and second applications obtained by the above formulas 8 and 9 and the holding time h ₁ , From h ₂ , the call volumes Er ₁ and Er ₂ of the first application and the second application are obtained from the definition of the call volume according to the following equation (10).

Using the call volumes Er ₁ and Er ₂ of the first and second applications obtained in this way, either the Erlang-B equation shown by the above equation 6 or the multidimensional Erlang-B equation shown by the above equation 7 is used. The number of lines K ₁ , K ₂ necessary to make the call loss rates B ₁ (K), B ₂ (K) for the first and second applications calculated by the formula below a predetermined value are calculated. The bandwidth BW _des required to set the call loss rate in the network 10 to a predetermined value or less using the number of lines K ₁ and K ₂ and the bandwidths BW ₁ and BW _{2 of} the first and second applications. Is calculated.

As a procedure for calculating the bandwidth BW _des , the first and second application bandwidths K ₁ and K ₂ calculated by the Erlang-B equation and the bandwidths BW ₁ and BW _{2 of} the first and second applications are used. For each second application, the bandwidths BW _des1 and BW _des2 required to make the call loss rate equal to or lower than the predetermined value are calculated, and the total value of the calculated bandwidths BW _des1 and BW _des2 is defined as the bandwidth BW _des. Maximum pledge between the calculation procedure or the larger number of lines K ₁ and K ₂ calculated by the multidimensional Erlang-B equation and the bandwidths BW ₁ and BW ₂ of the _first and _second applications One of the procedures for calculating a value obtained by multiplying the number as the band BW _des is used.

  Next, a <third embodiment> will be described.

  The feature of the third embodiment is that the bandwidth calculation technique that does not use the bandwidth B used in the second embodiment, that is, both the overflow bandwidth B ′ and the bandwidth B used in the first embodiment. The bandwidth calculation technique is not used. In the third embodiment, the amount of information that needs to be retained is smaller than that in the second embodiment, because the bandwidth amount B is not used, but the calculation accuracy is the first and second embodiments. Inferior to examples.

  Also in the third embodiment, symbols are placed in the same manner as in the first embodiment, and the necessary bandwidth is calculated assuming that there is only an application that requires two bandwidths.

First, the number of completed calls and the number of incomplete calls (N ₁ , N ₂ , N ′ ₁ , N ′ ₂ ) of the first and second applications are obtained by the following relational expression (11).

Next, the hold time of the first application is set to h ₁ , the hold time of the second application is set to h ₂ , and the hold times h ₁ , h of the _first application and the second application are expressed by the following equation (12). ₂ is determined.

Based on the number of completed calls N ₁ and N ₂ , the number of incomplete calls N ′ ₁ and N ′ ₂ and the holding times h ₁ and h _{2 of} the first and second applications obtained in this way, The call volumes Er ₁ and Er 2 of the application ₂ are obtained by the following formula 13 from the call volume definition.

Then, using the call volumes Er ₁ and Er ₂ of the first and second applications, it is calculated by either the above-described Erlang-B equation represented by Equation 6 or the above-described multidimensional Erlang-B equation represented by Equation 7. The number of lines K ₁ , K ₂ required to set the call loss rates B ₁ (K), B ₂ (K) for the first and second applications to a predetermined value or less are calculated, and the calculated number of lines Using K ₁ , K ₂ and the bandwidths BW ₁ , BW _{2 of} the first and second applications, a bandwidth BW _des required to make the call loss rate in the network 10 equal to or less than a predetermined value is calculated. .

As a procedure for calculating the bandwidth BW _des , the first and second application bandwidths K ₁ and K ₂ calculated by the Erlang-B equation and the bandwidths BW ₁ and BW _{2 of} the first and second applications are used. For each second application, the bandwidths BW _des1 and BW _des2 required to make the call loss rate equal to or lower than the predetermined value are calculated, and the total value of the calculated bandwidths BW _des1 and BW _des2 is defined as the bandwidth BW _des. Maximum pledge between the calculation procedure or the larger number of lines K ₁ and K ₂ calculated by the multidimensional Erlang-B equation and the bandwidths BW ₁ and BW ₂ of the _first and _second applications One of the procedures for calculating a value obtained by multiplying the number as the band BW _des is used.

  The amount of information necessary for calculating the bandwidth required to satisfy the call loss rate below a certain specified value by the processing described in the first to third embodiments, and the required bandwidth for conventional multi-way calls. A comparison with the amount of information is shown in FIG.

  In FIG. 3, the horizontal axis indicates the number of band types, and the vertical axis indicates the amount of information required. As shown in FIG. 3, in the conventional multi-band call bandwidth calculation technique, as the number of bandwidth types increases, the amount of information required for bandwidth calculation increases linearly. However, the first to third embodiments ( In the figure, “Examples 1, 2, and 3”), it is understood that the amount of information necessary for bandwidth calculation does not increase even if the number of types of bandwidth increases.

  Also, depending on the type of application, it is assumed that the required bandwidth may change continuously on a call basis, so in the conventional technology, it is necessary to record information for all calls and use it to calculate the required bandwidth. Then, the necessary amount of information is constant regardless of the band type, and even in such a case, the superiority of this example can be confirmed.

  Next, the processing procedure in the first to third embodiments will be described with reference to FIGS.

  In the first embodiment, as shown in FIG. 4, in a network in which applications having different request bandwidths are mixed and admission control is performed every time a call from the application occurs, a call loss rate indicating the ratio of calls to be discarded is set in advance. When calculating a bandwidth required to be equal to or less than a predetermined value by programmed computer processing, the network bandwidth calculation system 1 first determines the call control unit 2 in advance by the call number / bandwidth calculation unit 3. The total number C of completed calls that have been able to capture a band when a new call is received within the observed period T and the total number C ′ of incomplete calls that are unable to acquire a band when a new call is received are counted. And it memorize | stores in the memory | storage devices 3a and 3b (step S401).

Then, the connection call number and use bandwidth calculation unit 4, any number of calls connected to the network 10 at time T ₁ L _C and used bandwidth value L _BW and storage device 4a seeking in the observation period T (Step S402).

  In addition, the call count / bandwidth calculation unit 3 can capture the total value B of the required bandwidth of each completed call of the total number C that can be acquired within the observation period T, and can acquire the band within the observation period T. The total value B ′ of the requested bandwidth of each incomplete call of the total number C ′ that has not been calculated is calculated and stored in the storage devices 3a and 3b (step S403).

  Further, the first and second application selection unit 5 selects the first and second applications from the applications mixed in the network 10 (step S404).

Then, the number of completed calls N ₁ , N ₂ and the number of incomplete calls N ′ ₁ , N ′ _{2 are} calculated by the first application band BW ₁ and the second bandwidth selected by the first and second application selection unit 5. And the total number of calls C and C ′ obtained by the call number / bandwidth calculation unit 3 and the integrated values B and B ′ using the above formula 3 to obtain the first application's bandwidth BW ₂ . completed call number _{N 1} and 'a completed call number _{N 2} of the _first and second application not completed call number N' uncompleted calls the number N for calculating the ₂ (step S405).

Furthermore, the number of completed calls N ₁ and N ₂ , the number of incomplete calls N ′ ₁ and the number of incomplete calls N ₁ calculated by the N ′ ₂ calculation unit 6 by the holding time h ₁ and h ₂ calculation unit 7. N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ , the first application band BW ₁ and the second application band BW ₂ selected by the first and second application selector 5, and the number L _C and used bandwidth value L _BW of call connection call number and use bandwidth calculation unit 4 is _determined, and, by the number 4 in the above equation using the observation period T, the first application occupying the band The holding time h ₁ that is the time to perform and the holding time h ₂ that is the time that the second application occupies the band are calculated (step S406).

Also, the call volume Er ₁ , Er ₂ calculation unit 8 calculates the number of completed calls N ₁ , N ₂ , the number of incomplete calls N ′ ₁ , the number of incomplete calls N ₁ calculated by the N ′ ₂ calculation unit 6. N _'1, completed call number _{N 2,} uncompleted calls the number N' and _2, holding time _h 1, _{h 2} calculator 7 holding time _h 1 was calculated, _{h 2,} and described above using the observation period T the expression of equation 5 is calculated call of the first application Er ₁ and the traffic intensity Er ₂ of the second application (step S407).

Then, if necessary bandwidth calculation unit 9, the multi indicated by traffic density _Er 1, Er ₂ calculator traffic intensity 8 is calculated _Er 1, _{Er 2} Erlang-B formula or several of above 7 indicate the number 6 above using The number of lines K ₁ required to set the call loss rates B ₁ (K), B ₂ (K) for the first and second applications calculated by any one of the dimension Erlang-B formulas to a predetermined value or less. , K _2, and using the calculated number of lines K ₁ , K ₂ and the bandwidths BW ₁ , BW _{2 of} the first and second applications, the call loss rate in the network 10 is equal to or less than a predetermined value. The bandwidth BW _des required for the calculation is calculated (step S408).

As a procedure for calculating the bandwidth BW _des , the first and second application bandwidths K ₁ and K ₂ calculated by the Erlang-B equation and the bandwidths BW ₁ and BW _{2 of} the first and second applications are used. For each second application, the bandwidths BW _des1 and BW _des2 required to make the call loss rate equal to or lower than the predetermined value are calculated, and the total value of the calculated bandwidths BW _des1 and BW _des2 is defined as the bandwidth BW _des. Maximum pledge between the calculation procedure or the larger number of lines K ₁ and K ₂ calculated by the multidimensional Erlang-B equation and the bandwidths BW ₁ and BW ₂ of the _first and _second applications One of the procedures for calculating a value obtained by multiplying the number as the band BW _des is used.

  Further, in the second embodiment, as shown in FIG. 5, the network bandwidth calculation system 1 is first updated by the call count / bandwidth calculation unit 3 within the observation period T in which the call control unit 2 is predetermined. The total number C of completed calls that have been able to capture the bandwidth at the time of call reception and the total number C ′ of incomplete calls that have not been able to acquire the bandwidth at the time of reception of a new call are counted in the storage devices 3a and 3b. Store (step S501).

Then, the connection call number and use bandwidth calculation unit 4, any number of calls connected to the network 10 at time T ₁ L _C and used bandwidth value L _BW and storage device 4a seeking in the observation period T (Step S502).

  Further, the call number / bandwidth calculation unit 3 calculates the total value B of the required bandwidths of the completed calls for the total number C of which the bandwidth could be acquired within the observation period T and stores them in the storage devices 3a and 3b ( Step S503).

  Further, the first and second application selection unit 5 selects the first and second applications from the applications mixed in the network 10 (step S504).

Then, the number of completed calls N ₁ , N ₂ and the number of incomplete calls N ′ ₁ , N ′ _{2 are} calculated by the first application band BW ₁ and the second bandwidth selected by the first and second application selection unit 5. and band BW _2, the application of the total number of calls the call number, band amount calculating section 3 obtains C, by the numerical formula 8 described above using the integrated value B and C ', the number of completed calls first application N ₁ and 'a completed call number _{N 2} of the _first and second application not completed call number N' uncompleted calls the number N for calculating the ₂ (step S505).

Furthermore, the number of completed calls N ₁ and N ₂ , the number of incomplete calls N ′ ₁ and the number of incomplete calls N ₁ calculated by the N ′ ₂ calculation unit 6 by the holding time h ₁ and h ₂ calculation unit 7. N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ , the first application band BW ₁ and the second application band BW ₂ selected by the first and second application selector 5, and the number L _C and used bandwidth value L _BW of call connection call number and use bandwidth calculation unit 4 is _determined, and, by the number 9 equations described above using the observation period T, the first application occupying the band The holding time h ₁ that is the time to perform and the holding time h ₂ that is the time that the second application occupies the band are calculated (step S506).

Also, the call volume Er ₁ , Er ₂ calculation unit 8 calculates the number of completed calls N ₁ , N ₂ , the number of incomplete calls N ′ ₁ , the number of incomplete calls N ₁ calculated by the N ′ ₂ calculation unit 6. N _'1, completed call number _{N 2,} uncompleted calls the number N' and _2, holding time _h 1, _{h 2} calculator 7 holding time _h 1 was calculated, _{h 2,} and described above using the observation period T the number 10 equation to calculate traffic density of the first application Er ₁ and the traffic intensity Er ₂ of the second application (step S507).

Then, if necessary bandwidth calculation unit 9, the multi indicated by traffic density _Er 1, Er ₂ calculator traffic intensity 8 is calculated _Er 1, _{Er 2} Erlang-B formula or several of above 7 indicate the number 6 above using The number of lines K ₁ required to set the call loss rates B ₁ (K), B ₂ (K) for the first and second applications calculated by any one of the dimension Erlang-B formulas to a predetermined value or less. , K _2, and using the calculated number of lines K ₁ , K ₂ and the bandwidths BW ₁ , BW _{2 of} the first and second applications, the call loss rate in the network 10 is equal to or less than a predetermined value. The bandwidth BW _des required for the calculation is calculated (step S508).

As a procedure for calculating the bandwidth BW _des , the first and second application bandwidths K ₁ and K ₂ calculated by the Erlang-B equation and the bandwidths BW ₁ and BW _{2 of} the first and second applications are used. For each second application, the bandwidths BW _des1 and BW _des2 required to make the call loss rate equal to or lower than the predetermined value are calculated, and the total value of the calculated bandwidths BW _des1 and BW _des2 is defined as the bandwidth BW _des. Maximum pledge between the calculation procedure or the larger number of lines K ₁ and K ₂ calculated by the multidimensional Erlang-B equation and the bandwidths BW ₁ and BW ₂ of the _first and _second applications One of the procedures for calculating a value obtained by multiplying the number as the band BW _des is used.

  Further, in the third embodiment, as shown in FIG. 6, the network bandwidth calculation system 1 is first updated by the call count / bandwidth calculation unit 3 by the call control unit 2 within a predetermined observation period T. The total number C of completed calls that have been able to capture the bandwidth at the time of call reception and the total number C ′ of incomplete calls that have not been able to acquire the bandwidth at the time of reception of a new call are counted in the storage devices 3a and 3b. Store (step S601).

Then, the connection call number and use bandwidth calculation unit 4, any number of calls connected to the network 10 at time T ₁ L _C and used bandwidth value L _BW and storage device 4a seeking in the observation period T (Step S602).

  In the third embodiment, the call count / bandwidth calculation unit 3 shown in FIGS. 4 and 5 does not calculate the integrated value of the requested bandwidth, and the first and second application selection unit 5 establishes the network 10. First and second two applications are selected from the mixed applications (step S603).

Then, the number of completed calls N ₁ , N ₂ and the number of incomplete calls N ′ ₁ , N ′ _{2 are} calculated by the first application band BW ₁ and the second bandwidth selected by the first and second application selection unit 5. band BW 2 of the _application, and by the number 11 equations described above using a number of connection call connection call number and use bandwidth calculation unit 4 is determined L _C and the used bandwidth value L _BW, completion of the first application calculating the number of calls _{N 1} non complete call number N _'1 and complete the call number _{N 2} and uncompleted call number N of the second application' ₂ (step S604).

Furthermore, the number of completed calls N ₁ and N ₂ , the number of incomplete calls N ′ ₁ and the number of incomplete calls N ₁ calculated by the N ′ ₂ calculation unit 6 by the holding time h ₁ and h ₂ calculation unit 7. N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ , the first application band BW ₁ and the second application band BW ₂ selected by the first and second application selector 5, and the number L _C and used bandwidth value L _BW of call connection call number and use bandwidth calculation unit 4 is _determined, and, by the number 12 equations described above using the observation period T, the first application occupying the band The holding time h ₁ that is the time to perform and the holding time h ₂ that is the time that the second application occupies the band are calculated (step S605).

Also, the call volume Er ₁ , Er ₂ calculation unit 8 calculates the number of completed calls N ₁ , N ₂ , the number of incomplete calls N ′ ₁ , the number of incomplete calls N ₁ calculated by the N ′ ₂ calculation unit 6. N _'1, completed call number _{N 2,} uncompleted calls the number N' and _2, holding time _h 1, _{h 2} calculator 7 holding time _h 1 was calculated, _{h 2,} and described above using the observation period T the number 13 equation to calculate traffic density of the first application Er ₁ and the traffic intensity Er ₂ of the second application (step S606).

Then, if necessary bandwidth calculation unit 9, the multi indicated by traffic density _Er 1, Er ₂ calculator traffic intensity 8 is calculated _Er 1, _{Er 2} Erlang-B formula or several of above 7 indicate the number 6 above using The number of lines K ₁ required to set the call loss rates B ₁ (K), B ₂ (K) for the first and second applications calculated by any one of the dimension Erlang-B formulas to a predetermined value or less. , K _2, and using the calculated number of lines K ₁ , K ₂ and the bandwidths BW ₁ , BW _{2 of} the first and second applications, the call loss rate in the network 10 is equal to or less than a predetermined value. The bandwidth BW _des required for the calculation is calculated (step S607).

As a procedure for calculating the bandwidth BW _des , the first and second application bandwidths K ₁ and K ₂ calculated by the Erlang-B equation and the bandwidths BW ₁ and BW _{2 of} the first and second applications are used. For each second application, the bandwidths BW _des1 and BW _des2 required to make the call loss rate equal to or lower than the predetermined value are calculated, and the total value of the calculated bandwidths BW _des1 and BW _des2 is defined as the bandwidth BW _des. Maximum pledge between the calculation procedure or the larger number of lines K ₁ and K ₂ calculated by the multidimensional Erlang-B equation and the bandwidths BW ₁ and BW ₂ of the _first and _second applications One of the procedures for calculating a value obtained by multiplying the number as the band BW _des is used.

  In the above estimation of the call volume using the two virtual applications, the estimation is performed from the statistic indicating the traffic situation. Therefore, it may be difficult to estimate the appropriate call volume in an area where the traffic is heavy. . A technique for dealing with such a case will be described below with reference to FIGS.

In FIG. 7, the network bandwidth calculation system 71, call control unit 72, number of calls / bandwidth calculation unit 73, number of connected calls / bandwidth calculation unit 74, first and second application selection unit 75, number of completed calls N ₁ , N _{2. The} number of incomplete calls N ′ ₁ , N ′ ₂ calculation unit 76, hold time h ₁ , h ₂ calculation unit 77, call volume Er ₁ , Er ₂ calculation unit 78, and necessary bandwidth calculation unit 79 are shown in FIG. Network bandwidth calculation system 1, call control unit 2, number of calls / bandwidth calculation unit 3, number of connected calls / bandwidth calculation unit 4, first and second application selection unit 5, number of completed calls N ₁ , N _2. It has the same functions as the number of incomplete calls N ′ ₁ , N ′ ₂ calculation unit 6, hold time h ₁ , h ₂ calculation unit 7, call volume Er ₁ , Er ₂ calculation unit 8, and necessary bandwidth calculation unit 9. The network bandwidth calculation system 71 has a computer processing program. As means for executing the processing, a call loss number C ₀ calculation unit 78a and a call loss number C ₀ evaluation unit 78b are newly provided.

In the network bandwidth calculation system 71 provided with the call loss number C ₀ calculation unit 78a and the call loss number C ₀ evaluation unit 78b, steps S401 to S407 in FIG. 4, steps S501 to S507 in FIG. 5, and steps in FIG. The process shown in FIG. 8 is performed using the call volumes Er ₁ and Er ₂ of the first and second applications calculated in the processes of S601 to S606.

  The first and second application selection unit 75 selects an application using the maximum bandwidth as the first application, and selects an application using the minimum bandwidth as the second application. .

Call blocking numbers _{C 0} calculator 78a uses the traffic intensity _Er 1, Er ₂ calculator 78 obtains call amount _Er 1, Er _2, either Erlang-B formula or multidimensional Erlang-B formula above the first, to calculate the blocking probability B _1, B ₂ for the second application, using the completed call number N ₂ of the first application of the completed call number N ₁ and the second application, the expression "N _The call loss number C ₀ is calculated from “ ₁ × B ₁ + N ₂ × B ₂ ” (step S801).

The call loss number C ₀ evaluation unit 78b has a threshold value in which the difference between the call loss number C ₀ calculated by the call loss number C ₀ calculation unit 78a and the total number C ′ of incomplete calls generated within the observation period T is determined in advance. A comparison / determination is made as to whether or not the result is below (step S802). If the comparison determination result is equal to or greater than the threshold (step S803), the first and second application selectors 75 are activated, and the first and second In the application selection unit 75, the bandwidth of the second application is increased by a predetermined value (step S804).

Then, the number of completed calls N ₁ , N ₂ , the number of uncompleted calls N ′ ₁ , N ′ ₂ by the calculation unit 76, the number of completed calls N ₁ , N ₂ and the number of uncompleted calls N ′ ₁ , N ′ ₂ , holding time h ₁ , h ₂ calculating unit 77 calculating holding time h ₁ , h ₂ , call volume Er ₁ , Er ₂ calculating unit 78 calculating call volume Er ₁ , Er ₂ , and Then, the call loss rates B ₁ and B ₂ are calculated by the call loss number C ₀ calculation unit 78a (step S805), and then the process returns to steps S801 to S803 by the call loss number C ₀ evaluation unit 78b.

The processes in steps S801 to S805 described above are repeated until the result of the comparison determination process in step S803 is less than or equal to the threshold value. If the result of the comparison determination process in step S803 is less than or equal to the threshold value, Call loss rates B ₁ (K), B for the first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula using the call volumes Er ₁ , Er _{2 2} Calculate the number of lines K ₁ and K ₂ necessary to set (K) to be equal to or less than a predetermined value, and calculate the calculated number of lines K ₁ and K ₂ and the bandwidths of the first and second applications. Is used to calculate the bandwidth required to set the call loss rate in the network to a predetermined value or less (step S806).

  In this way, this technology introduces a feedback mechanism for aligning the number of call losses with respect to the call volume estimated from the statistics indicating the traffic situation, and adds a constraint condition at the time of estimation, thereby adding a virtual condition. As a result, the estimation result can be obtained more than in the case where two applications are determined.

In the example described with reference to FIG. 8, if the result of the comparison by the call loss number C ₀ evaluation unit 78b in step S802 is a comparison / determination result with a threshold value or more, the first and second application selection unit 75 in step S803. In this process, the bandwidth of the second application is increased by a predetermined value. However, as the processing by the first and second application selection unit 75 in step S804, the bandwidth of the first application is predetermined. It is also possible to decrease the value by the specified value.

As described above, when the maximum and minimum values are used as the first and second types of applications, the calculated call loss number C ₀ is always less than or equal to the actual call loss number. Therefore, the process of increasing the bandwidth of the second application with the smallest initial value by a predetermined value or decreasing the bandwidth of the first application with the largest initial value by a predetermined value is deterministic. However, as the first and second types of applications, it is possible not to limit the use of the maximum value and the minimum value as the initial values.

  In other words, the first and second application selection unit 75 sandwiches both the average bandwidth of completed calls (B ÷ C) and the average bandwidth of incomplete calls (B ′ ÷ C ′) as the first and second applications. The larger one of any two bands may be selected as the first application, and the smaller one may be selected as the second application.

In this case, the call loss number C ₀ calculation unit 78a uses the call volumes Er ₁ and Er ₂ obtained by the call volume Er ₁ and Er ₂ calculation unit 78 to use the above Erlang-B formula or multidimensional Erlang-B formula. by any of the first, to calculate the blocking probability _B 1, _{B 2} for the second application, in step S801, the completed call number _{N 2} of the first application of the completed call number _{N 1} and the second application It was used to calculate the call blocking numbers _{C 0} by the equation _{"N 1 × B 1 + N 2} × B 2 ", in step S802, the in call loss number _{C 0} evaluating unit 78b, the call loss number _{C 0} calculator 78a The calculated call loss number C ₀ is compared with whether or not the difference between the calculated call loss number C ₀ and the total number C ′ of incomplete calls generated within the observation period T is equal to or less than a predetermined threshold. Incomplete call Number C 'greater than or performs small or the comparison determination.

In step S803, the comparison determination result between the above threshold and until the calculated call loss number C ₀ was within when the total number C 'larger than the non-completed call becomes below the threshold comparison determination result, the second application The bandwidth of the first application is decreased by a predetermined value or the bandwidth of the first application is increased by a predetermined value, and the number of completed calls N ₁ and N ₂ of the first and second applications in step S805 and incomplete calls Calculation of numbers N ′ ₁ , N ′ ₂ , holding times h ₁ , h ₂ , call volume Er ₁ , Er ₂ , calculation of call loss rates B ₁ , B _2, and calculation of call loss number C ₀ in step S 801, In addition, the comparison between the difference between the call loss number C ₀ and the total number C ′ of incomplete calls C ′ and the threshold in steps S802 and S803 is repeated.

Further, in step S803, the comparison determination result between the above threshold and if the calculated call loss number C ₀ that the total number C 'is smaller than the uncompleted call until threshold value or less in comparison determination result, the second The bandwidth of the first application is increased by a predetermined value or the bandwidth of the first application is decreased by a predetermined value, and the number of completed calls N ₁ and N ₂ of the first and second applications in step S805 is not equal. Calculation of the number of completed calls N ′ ₁ , N ′ ₂ , holding times h ₁ , h ₂ , call volume Er ₁ , Er ₂ , calculation of call loss rates B ₁ , B ₂ , and call loss number C ₀ in step S 801 The calculation and the comparison / determination of the difference between the call loss count C ₀ and the total number C ′ of incomplete calls and the threshold in steps S802 and S803 are repeated.

If the result of the comparison / determination process in step S803 is less than or equal to the threshold value, in step S806, the required bandwidth calculation unit 79 uses the call volumes Er ₁ and Er ₂ when the required bandwidth is less than or equal to the threshold value to determine the Erlang-B formula. Alternatively, the number of lines necessary to set the call loss rates B ₁ (K) and B ₂ (K) for the first and second applications calculated by either of the multidimensional Erlang-B formulas to be equal to or less than a predetermined value. Necessary for calculating the call loss rate in the network to be equal to or less than a predetermined value by calculating K ₁ and K ₂ and using the calculated number of lines K ₁ and K ₂ and the bandwidth of the first and second applications. Calculate the bandwidth.

  FIG. 9 is a comparison of the accuracy of required bandwidth calculation between the first technique that determines two virtual applications and the second technique described in FIGS. 7 and 8. In the technology, the fluorescence for calculating the necessary bandwidth is excessively small in the overloaded region, but the second technology shows that the necessary bandwidth can be estimated with a small error even in the overloaded region. Therefore, the second technique makes it possible to calculate the necessary bandwidth with high accuracy regardless of the traffic conditions even in an area where the traffic is heavily loaded.

  In FIG. 9, the hold times of the three applications of 100 kbps, 2 Mbps, and 10 Mbps are given as “0.1”, “0.1”, and “0.9”, respectively, and the added bands are similar. In this case, the accuracy of calculation of the required bandwidth is compared with the technology for deciding the two virtual applications, and the link bandwidth is 1 Gbps.

As described above with reference to FIGS. 1 to 9, in this example, when there is an application requesting a plurality of bandwidths in the network 10, it is necessary to set the call loss rate to a predetermined value or less. When calculating the bandwidth, without using the traffic information for each bandwidth, the total number C of completed calls that have been able to secure the bandwidth, the total number C 'of incomplete calls that have not been able to secure the bandwidth, and the bandwidth can be secured. 6 of the total number C of the requested bandwidth of each completed call of the total number C, the total value B ′ of the requested bandwidth of each incomplete call of the total number C ′ whose bandwidth could not be secured, the number of connected calls L _C , and the used bandwidth value L _BW One information or five pieces of information obtained by removing the integrated value B ′ from the six pieces of information, or four pieces of information obtained by removing the integrated values B and B ′ from the six pieces of information are used. There are only types Approximate, using only average information content for a call, to calculate a band necessary for below a specified blocking probability.

  As described above, according to this example, only the maximum six pieces of information are used for calculating the necessary bandwidth, and the cost associated with the bandwidth calculation can be kept constant even when the types of required bandwidth increase. Network operation costs can be greatly reduced.

  That is, in the technology for calculating the necessary bandwidth using the conventional multidimensional Erlang-B method, when applications that request a plurality of bandwidths exist in the same network, in order to perform bandwidth calculation, traffic information for each requested bandwidth Therefore, there is a problem that the amount of information to be stored becomes enormous as the bandwidth type increases, but according to this example, it is possible to calculate the bandwidth with a small amount of information Thus, even if the types of requested bandwidths increase, the amount of information used for bandwidth calculation is not increased, so that the network operating cost can be greatly reduced.

In addition, as the two types of applications, the first application using the maximum bandwidth and the second application using the minimum bandwidth are selected, and the Erlang-B formula or the multidimensional Erlang-B formula is selected. the first is calculated by either of the blocking probability B _1, B ₂ for the second application, using the completed call number N ₂ of the first application of the completed call number N ₁ and the second application, wherein The call loss number C ₀ is calculated by “N ₁ × B ₁ + N ₂ × B ₂ ”, and a difference between the calculated call loss number C ₀ and the total number C ′ of incomplete calls generated within the observation period T is determined in advance. If the comparison determination result is equal to or greater than the threshold, the second application bandwidth is determined in advance until the comparison determination result is equal to or less than the threshold. Is increased by a value, or reduces by a predetermined value the band of the first application, first, the number of completed calls second application N _1, N ₂ and uncompleted call number N _'1, N ′ ₂ , holding times h ₁ , h ₂ , call volume Er ₁ , Er ₂ , call loss rate B ₁ , B ₂ and call loss number C ₀ , and call loss number C ₀ and incomplete The comparison between the difference between the total number of calls C ′ and the threshold value is repeatedly performed, and the Erlang-B formula or the multidimensional Erlang-B is used by using the call volumes Er ₁ and Er ₂ when the comparison determination result falls below the threshold value. The number of lines K ₁ , K ₂ required to make the call loss rates B ₁ (K), B ₂ (K) for the first and second applications calculated by any of the equations below a predetermined value calculated conducted, and the calculated number of lines _K 1, _{K 2} and 1, by using the band of the second application, to calculate a band necessary for a predetermined value or less that the call loss probability in the network.

Alternatively, as two types of applications, the larger of any two bands sandwiching both the average band of completed calls (B ÷ C) and the average band of incomplete calls (B ′ ÷ C ′) is the first application, The smaller one is selected as the second application, and the call loss rates B ₁ and B ₂ for the first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula, and the first application using the completed call number _{N 1} and the second application and a completed call number _{N 2,} call loss number is calculated call loss number _{C 0,} which is calculated by the equation _{"N 1 × B 1 + N 2} × B 2 ' C ₀ is compared with the total number C ′ of incomplete calls that occur within the observation period T, and the difference is equal to or greater than the threshold value, and the calculated call loss number C ₀ is greater than the total number C ′ of incomplete calls. The difference is below the threshold Until the second application bandwidth is decreased by a predetermined value or the first application bandwidth is increased by a predetermined value to complete the number of completed calls N ₁ , N ₂ and the number of incomplete calls N ′ ₁ , N ′ ₂ , hold times h ₁ and h ₂ , call volume Er ₁ and Er ₂ , and call loss rates B ₁ and B ₂ and call loss number C ₀ , In addition, the comparison between the call loss number C ₀ and the total number C ′ of incomplete calls is repeated, the difference between the call loss number C ₀ and the total number C ′ of incomplete calls is equal to or greater than a threshold value, and the number of call losses If C ₀ is smaller than the total number of incomplete calls C ′, the bandwidth of the second application is increased by a predetermined value or the bandwidth of the first application is predetermined until the difference becomes equal to or less than the threshold value. The first and second applications are reduced by the specified value. Down the number of completed calls _N 1, _{N 2} and uncompleted call number N _'1, N' _2, holding time _h 1, _{h 2,} the calculation of the call weight _Er 1, Er _2, and a blocking probability _B 1, _{B 2} calculation of call blocking numbers C _0, and, 'repeats the comparison of a call loss number C ₀ total C not completed call' lost calls number C ₀ total C not completed calls the difference between the predetermined The call loss rate B ₁ (K) for the first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula using Er ₁ and Er ₂ when the threshold value is below the threshold value. The number of lines K ₁ and K ₂ necessary to make B ₂ (K) equal to or less than a predetermined value is calculated.

As described above, the call loss rates B ₁ and B ₂ for the first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula, and the number of completed calls N ₁ and the _first application for the first application are calculated. by using the complete call number _{N 2} of the second application, to calculate the call blocking numbers _{C 0} by the equation _{"N 1 × B 1 + N 2} × B 2 ", the call blocking numbers _{C 0} calculated for the observation period T In order to approach the total number C ′ of incomplete calls that occur, the bandwidth of the first and second applications is increased or decreased, and then the call volume estimation is repeated, so that the call loss number C ₀ is the total number C ′ of incomplete calls. Specified by the Erlang-B formula or the multidimensional Erlang-B formula using the call volume that finally gives the total number of incomplete calls C ′ and the number of lost calls. Call loss rate Calculating the bandwidth required to.

  As a result, even in an area where the traffic is high, it is possible to calculate the necessary bandwidth with high accuracy regardless of the traffic conditions.

In addition, this invention is not limited to the example demonstrated using FIGS. 1-9, In the range which does not deviate from the summary, various changes are possible. For example, in the present embodiment, connection call number and use bandwidth calculation unit 4 is, when obtaining an optional in time T ₁ to the number L _C of calls connected to the network 10 using band value L _BW in the observation period T The number of connected calls obtained at an arbitrary plurality of times Tn within the observation period T and the average value of the used band values are obtained. For example, the number of connected calls obtained at an arbitrary plurality of times Tn within the observation period T And the number of connected calls obtained at any of a plurality of times Tn within the observation period T, including the maximum value and the maximum value of the used band value, or the minimum value of the number of hourly connected calls and the minimum value of the used band value. The value of the statistical processing result for the band value may be used.

  Further, regarding the computer configuration example of this example, a computer configuration without a keyboard or an optical disk drive device may be used. In this example, an optical disc is used as a recording medium, but an FD (Flexible Disk) or the like is used as the recording medium. It may be used as. As for the program installation, the program may be downloaded and installed via a network via a communication device.

It is a block diagram which shows the structural example of the network band calculation system which concerns on this invention. It is explanatory drawing which shows the concept of the bandwidth amount and overflow bandwidth amount which are used with the network bandwidth calculation system in FIG. It is explanatory drawing which shows the comparative example of the information amount required in the network band calculation system in FIG. 1, and the information amount required by a prior art. 3 is a flowchart showing a first processing operation example according to the present invention by the network bandwidth calculation system in FIG. 1. 6 is a flowchart showing a second processing operation example according to the present invention by the network bandwidth calculation system in FIG. 1. 7 is a flowchart showing a third processing operation example according to the present invention by the network bandwidth calculation system in FIG. 1. It is a block diagram which shows the other structural example of the network band calculation system which concerns on this invention. It is a flowchart which shows the processing operation example which concerns on this invention by the network band calculation system in FIG. It is explanatory drawing which shows the example of an effect by the network band calculation system in FIG.

Explanation of symbols

1: network bandwidth calculation system (“bandwidth calculation system”), 2: call control unit, 2a: call with bandwidth supplement (completed call), 2b: call with bandwidth supplement (uncompleted call), 3: number of calls - bandwidth amount calculating unit, 3a, 3b: storage, 4: connection call number and use bandwidth calculation unit, 4a: storage, 5: first, second application selection section, 6: completed call number _N 1, _{N 2} Number of incomplete calls N ′ ₁ , N ′ ₂ calculating unit, 7: holding time h ₁ , h ₂ calculating unit, 8: call volume Er ₁ , Er ₂ calculating unit, 9: necessary bandwidth calculating unit, 10: network, 70: network, 71: network bandwidth calculation system (“bandwidth calculation system”), 72: call control unit, 72a: call that has been able to supplement the bandwidth (completed call), 72b: call that has not been able to supplement the bandwidth (uncompleted call), 73: Number of calls / bandwidth calculation unit, 73a, 73b: storage device 74: connection call number and use bandwidth calculation unit, 74a: storage device, 75: first, second application selection unit, 76: completed call number _N 1, _{N 2,} uncompleted calls the number N _'1, N' ₂ calculated , 77: holding time h ₁ , h ₂ calculation unit, 78: call volume Er ₁ , Er ₂ calculation unit, 78a: call loss number C ₀ calculation unit, 78b: call loss number C ₀ evaluation unit, 79: necessary bandwidth Calculation unit.

Claims (12)

In a network where applications with different required bandwidths are mixed and admission control is performed for each call of the application, the bandwidth required to set the call loss rate indicating the ratio of calls to be dropped to a predetermined value or less is programmed. A network bandwidth calculation method for calculating by computer processing,
As a programmed computer processing procedure
The total number C of calls (completion calls) that were able to capture a band when a new call was accepted and calls that were not able to capture a band when a new call was accepted, occurring within a predetermined observation period T A first procedure for counting the total number C ′ of (incomplete calls) and storing it in the storage device;
A second step of storing in a storage device in search of the number L _C and used bandwidth value L _BW calls connected to the network at any time T ₁ of the within the observation period T,
An integrated value B of the required bandwidth of each completed call of the total number C that was able to capture the bandwidth within the observation period T;
A third procedure for calculating and storing in the storage device the integrated value B ′ of the requested bands of each incomplete call of the total number C ′ that could not be acquired within the observation period T;
A fourth procedure for selecting the first and second applications from the applications mixed in the network;
First band BW ₁ of the application and band BW 2 of the second application selected in the fourth _step, and, the first, the total number of calls determined in the third step C, C 'and the integrated value B , B ′ and the following equation 1
A fifth step of calculating a completed call number N ₁ and uncompleted call number N _'1 and the number of completed calls the second application N ₂ and uncompleted call number N' ₂ of the first application,

The number of completed calls N ₁ , the number of incomplete calls N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ calculated in the fifth procedure, the band BW ₁ and the band BW ₂ , and the number the number L _C and used bandwidth value L _BW of calls determined in step _2, and by expression 2 below using the observation period T,
A sixth procedure for calculating a holding time h ₁ that is a time during which the _first application occupies the band and a holding time h ₂ during which the second application occupies the band;

The number of completed calls N ₁ , the number of incomplete calls N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ calculated in the fifth procedure, and the hold time h ₁ calculated in the sixth procedure, Using h ₂ and the above observation period T,
A seventh step of calculating the call weight Er ₂ of the first application of traffic intensity Er ₁ and the second application,

Using the call volumes Er ₁ and Er ₂ calculated in the seventh procedure, the above calculation is performed by either the Erlang-B equation shown by the following equation 4 or the multidimensional Erlang-B equation shown by the following equation 5. 1, an eighth procedure for calculating the number of lines K ₁ , K ₂ required to set the call loss rates B ₁ (K), B ₂ (K) for the second application to a predetermined value or less;

Using the number of lines K ₁ , K ₂ calculated in the eighth procedure and the bandwidths BW ₁ , BW _{2 of} the first and second applications, the call loss rate in the network is equal to or less than the predetermined value. And a ninth procedure for calculating a bandwidth BW _des necessary for the network bandwidth calculation. In a network where applications with different required bandwidths are mixed and admission control is performed for each call of the application, the bandwidth required to set the call loss rate indicating the ratio of calls to be dropped to a predetermined value or less is programmed. A network bandwidth calculation method for calculating by computer processing,
As a programmed computer processing procedure
The total number C of calls (completion calls) that were able to capture a band when a new call was accepted and calls that were not able to capture a band when a new call was accepted, occurring within a predetermined observation period T A first procedure for counting the total number C ′ of (incomplete calls) and storing it in the storage device;
A second step of storing in a storage device in search of the number L _C and used bandwidth value L _BW calls connected to the network at any time T ₁ of the within the observation period T,
A third procedure for calculating the total value B of the required bandwidths of each completed call for the total number C of which the bandwidth could be acquired within the observation period T and storing it in the storage device;
A fourth procedure for selecting the first and second applications from the applications mixed in the network;
First band BW ₁ of the application and band BW 2 of the second application selected in the fourth _step, and, the first, the total number of calls determined in the third step C, C 'and the integrated value B Using the following equation (6),
A fifth step of calculating a completed call number N ₁ and uncompleted call number N _'1 and the number of completed calls the second application N ₂ and uncompleted call number N' ₂ of the first application,

The number of completed calls N ₁ , the number of incomplete calls N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ calculated in the fifth procedure, the band BW ₁ and the band BW ₂ , and the number The number of calls L _C obtained in the procedure of 2 and the use band value L _BW , and the following equation 7 using the observation period T,
A sixth procedure for calculating a holding time h ₁ that is a time during which the _first application occupies the band and a holding time h ₂ during which the second application occupies the band;

The number of completed calls N ₁ , the number of incomplete calls N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ calculated in the fifth procedure, and the hold time h ₁ calculated in the sixth procedure, Using h ₂ and the following equation (8):
A seventh step of calculating the call weight Er ₂ of the first application of traffic intensity Er ₁ and the second application,

Using the call volumes Er ₁ and Er ₂ calculated in the seventh procedure, the above-mentioned calculation is performed by either the Erlang-B formula shown by the following formula 9 or the multidimensional Erlang-B formula shown by the following formula 10. 1, an eighth procedure for calculating the number of lines K ₁ , K ₂ required to set the call loss rates B ₁ (K), B ₂ (K) for the second application to a predetermined value or less;

Using the number of lines K ₁ , K ₂ calculated in the eighth procedure and the bandwidths BW ₁ , BW _{2 of} the first and second applications, the call loss rate in the network is equal to or less than the predetermined value. And a ninth procedure for calculating a bandwidth BW _des necessary for the network bandwidth calculation. In a network where applications with different required bandwidths are mixed and admission control is performed for each call of the application, the bandwidth required to set the call loss rate indicating the ratio of calls to be dropped to a predetermined value or less is programmed. A network bandwidth calculation method for calculating by computer processing,
As a programmed computer processing procedure
The total number C of calls (completion calls) that were able to capture a band when a new call was accepted and calls that were not able to capture a band when a new call was accepted, occurring within a predetermined observation period T A first procedure for counting the total number C ′ of (incomplete calls) and storing it in the storage device;
A second step of storing in a storage device in search of the number L _C and used bandwidth value L _BW calls connected to the network at any time T ₁ of the within the observation period T,
A third procedure for selecting the first and second applications from the applications mixed in the network;
First band BW ₁ of the application and and band BW _2, the second application, the following numbers using the total number of calls determined in the first procedure C, the C '11 selected in the third step From the equation
A fourth step of calculating the complete call number N ₁ and uncompleted call number N _'1 and the number of completed calls the second application N ₂ and uncompleted call number N' ₂ of the first application,

The number of completed calls N ₁ , the number of incomplete calls N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ calculated in the fourth procedure, the band BW ₁ and the band BW ₂ , and the number The number of calls L _C and use band value L _BW obtained in the procedure of 2, and the following equation 12 using the observation period T,
A fifth procedure for calculating a holding time h ₁ that is the time that the first application occupies the band and a holding time h ₂ that is the time that the second application occupies the band;

The number of completed calls N ₁ , the number of incomplete calls N ′ ₁ , the number of completed calls N ₂ , the number of incomplete calls N ′ ₂ calculated in the fourth procedure, and the hold time h ₁ calculated in the fifth procedure, Using h ₂ , the following equation (13):
A sixth step of calculating the call weight Er ₂ of the first application of traffic intensity Er ₁ and the second application,

Using the call volumes Er ₁ and Er ₂ calculated in the sixth procedure, the above calculation is performed by either the Erlang-B formula shown by the following formula 14 or the multidimensional Erlang-B formula shown by the following formula 15. 1, a seventh procedure for calculating the number of lines K ₁ , K ₂ necessary for making the call loss rates B ₁ (K), B ₂ (K) for the second application equal to or less than a predetermined value;

Using the number of lines K ₁ , K ₂ calculated in the seventh procedure and the bandwidths BW ₁ , BW _{2 of} the first and second applications, the call loss rate in the network is equal to or less than the predetermined value. And an eighth procedure for calculating a bandwidth BW _des necessary for the network bandwidth calculation. A network bandwidth calculation method according to any one of claims 1 to 3, wherein
The procedure for calculating the band BW _des is as follows:
Using the number of lines K ₁ and K ₂ calculated by the above Erlang-B equation and the bandwidths BW ₁ and BW ₂ of the first and second applications, the call loss rate for each of the first and second applications. Calculating bandwidths BW _des1 and BW _des2 necessary to set the bandwidth BW _des1 and BW _des2 to be equal to or lower than the predetermined value, and calculating the total value of the calculated bandwidths BW _des1 and BW _des2 as the bandwidth BW _des ,
Or
Multiplying the larger number of lines K ₁ and K ₂ calculated by the multidimensional Erlang-B equation and the greatest common divisor of the first and second application bands BW ₁ and BW ₂ One of the procedures for calculating the calculated value as the band BW _des is included. A network bandwidth calculation method according to any one of claims 1 to 4, wherein
In the procedure for selecting the first and second applications,
One application that uses a band larger than the average band value obtained by dividing the total value C of the requested bands of the completed calls for the total number C of the total calls C that can be acquired within the observation period T is the first band. Select as the application
A network bandwidth calculation method, wherein one of applications using a bandwidth smaller than the average bandwidth value is selected as the second application. A network bandwidth calculation method according to any one of claims 1 to 5,
Select the application that uses the maximum bandwidth as the first application,
A network bandwidth calculation method, wherein an application using a minimum bandwidth is selected as the second application. A network bandwidth calculation method according to any one of claims 1 to 5,
Select the application using the maximum bandwidth as the first application, select the application using the minimum bandwidth as the second application,
Call loss rates B ₁ and B ₂ for the first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula, the number of completed calls N ₁ of the first application, and the above Using the number of completed calls N ₂ of the second application and calculating the call loss number C ₀ by the formula “N ₁ × B ₁ + N ₂ × B ₂ ”;
A procedure for comparing and determining whether or not the difference between the calculated call loss number C ₀ and the total number C ′ of incomplete calls generated within the observation period T is equal to or less than a predetermined threshold;
If the comparison determination result is greater than or equal to the threshold value, the second application bandwidth is increased by a predetermined value until the comparison determination result becomes equal to or less than the threshold value, and the first and second application completion calls are completed. The number N ₁ , N ₂ and the number of incomplete calls N ′ ₁ , N ′ ₂ , the holding time h ₁ , h ₂ , the call volume Er ₁ , Er ₂ , and the call loss rates B ₁ , B ₂ and the call loss Calculating the number C ₀ , and repeatedly comparing and determining the difference between the call loss number C ₀ and the total number C ′ of incomplete calls and the threshold value,
The first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula using the Er ₁ and Er ₂ when the comparison discrimination result is below the threshold value. A network bandwidth calculation method, comprising: calculating the number of lines K ₁ and K ₂ necessary for setting the call loss rates B ₁ (K) and B ₂ (K) to be equal to or less than the predetermined value. A network bandwidth calculation method according to any one of claims 1 to 5,
Select the application using the maximum bandwidth as the first application, select the application using the minimum bandwidth as the second application,
Call loss rates B ₁ and B ₂ for the first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula, the number of completed calls N ₁ of the first application, and the above Using the number of completed calls N ₂ of the second application and calculating the call loss number C ₀ by the formula “N ₁ × B ₁ + N ₂ × B ₂ ”;
A procedure for comparing and determining whether or not the difference between the calculated call loss number C ₀ and the total number C ′ of incomplete calls generated within the observation period T is equal to or less than a predetermined threshold;
If the comparison determination result is greater than or equal to the threshold, the first application bandwidth is decreased by a predetermined value until the comparison determination result is less than or equal to the threshold, and the first and second application completion calls are completed. The number N ₁ , N ₂ and the number of incomplete calls N ′ ₁ , N ′ ₂ , the holding time h ₁ , h ₂ , the call volume Er ₁ , Er ₂ , and the call loss rates B ₁ , B ₂ and the call loss Calculating the number C ₀ , and repeatedly comparing and determining the difference between the call loss number C ₀ and the total number C ′ of incomplete calls and the threshold value,
The first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula using the Er ₁ and Er ₂ when the comparison discrimination result is below the threshold value. A network bandwidth calculation method, comprising: calculating the number of lines K ₁ and K ₂ necessary for setting the call loss rates B ₁ (K) and B ₂ (K) to be equal to or less than the predetermined value. A network bandwidth calculation method according to any one of claims 1 to 5,
As the first and second applications, the larger of any two bands sandwiching both the average band of the completed call (B ÷ C) and the average band of the incomplete call (B ′ ÷ C ′) Select the first application, the smaller one as the second application,
Call loss rates B ₁ and B ₂ for the first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula, the number of completed calls N ₁ of the first application, and the above Using the number of completed calls N ₂ of the second application and calculating the call loss number C ₀ by the formula “N ₁ × B ₁ + N ₂ × B ₂ ”;
A procedure for comparing and determining whether the calculated call loss number C ₀ and the total number C ′ of incomplete calls generated within the observation period T are larger or smaller and whether or not the difference is equal to or less than a predetermined threshold; ,
If the calculated call loss number C ₀ is larger than the total number C ′ of the incomplete calls, the second application until the comparison determined result is equal to or less than the threshold value. Is decreased by a predetermined value, or the first application band is increased by a predetermined value, and the number of completed calls N ₁ and N _{2 of} the first and second applications is not increased. Calculation of the number of completed calls N ′ ₁ , N ′ ₂ , holding times h ₁ , h ₂ , call volume Er ₁ , Er ₂ , calculation of the call loss rates B ₁ , B ₂ and call loss C ₀ , and A procedure for repeatedly comparing and determining the difference between the call loss number C ₀ and the total number C ′ of incomplete calls and the threshold value;
If the calculated call loss count C ₀ is smaller than the total number C ′ of the incomplete calls, the second application until the comparison determined result is equal to or less than the threshold. Is increased by a predetermined value, or the first application band is decreased by a predetermined value, and the number of completed calls N ₁ and N _{2 of} the first and second applications is not increased. Calculation of the number of completed calls N ′ ₁ , N ′ ₂ , holding times h ₁ , h ₂ , call volume Er ₁ , Er ₂ , calculation of the call loss rates B ₁ , B ₂ and call loss C ₀ , and A procedure for repeatedly comparing and determining the difference between the call loss number C ₀ and the total number C ′ of incomplete calls and the threshold value,
By using the above Er ₁ and Er ₂ when the difference between the calculated call loss number C ₀ and the total number C ′ of incomplete calls generated within the observation period T is equal to or less than a predetermined threshold value, Call loss rates B ₁ (K) and B ₂ (K) for the first and second applications calculated by either the Erlang-B formula or the multidimensional Erlang-B formula are set to be equal to or lower than the predetermined value. A network bandwidth calculation method characterized in that the number of lines K ₁ and K ₂ necessary for the calculation is calculated. A network bandwidth calculation method according to any one of claims 1 to 9,
As the number _{L C} and the used bandwidth value _{L BW} of the call,
The average value of the number of connected calls and the bandwidth used obtained at any of a plurality of times T _n within the observation period T;
Or
The maximum value of the number of connected calls and the maximum value of the used bandwidth value obtained at any of a plurality of times T _n within the observation period T;
Or
A connection call obtained at any of a plurality of times T _n within the observation period T, including the minimum value of the number of connection calls obtained at any time T _n within the observation period T and the minimum value of the bandwidth used. A network bandwidth calculation method characterized by using a value of a statistical processing result with respect to a number and a used bandwidth value. In a network in which applications with different requested bandwidths are mixed and admission control is performed every time a call from the application occurs, the bandwidth necessary for setting the call loss rate indicating the ratio of calls to be dropped below a predetermined value, A network bandwidth calculation system for calculating by programmed computer processing,
As a means of performing programmed computer processing,
11. A network bandwidth calculation system comprising means for executing each procedure in the network bandwidth calculation method according to claim 1.   The program for making a computer perform each procedure in the network bandwidth calculation method in any one of Claims 1-10.

Images