JPH11122267A - Atm exchange and method for setting atm virtual path capacity - Google Patents

Abstract

PROBLEM TO BE SOLVED: To set an ATM virtual path capacity for an ATM network actually in operation without the need of a specific measurement form. SOLUTION: In the method for setting a virtual path capacity that satisfies a cell loss rate object rate as to an ATM virtual path terminated to an ATM exchange, a Q length threshold level is set to a cell transmission wait buffer 121 attended with the ATM virtual path, number of arrived cells and number of times of exceeded Q length over the threshold level to the cell transmission wait buffer 121 are counted, a Q length threshold level excess frequency is calculated based on the arrived cell number and the Q length threshold level excess number of times, and whether or not the Q length threshold level excess frequency is higher than the cell loss object rate CLR is decided and the capacity of the ATM virtual path is set depending whether or not number of times of decision to be higher exceeds a prescribed reference number.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to an asynchronous transfer mode (ATM) network.
ATM for setting the capacity of the TM virtual path so as to satisfy a cell loss target value predetermined for the ATM virtual path.
The present invention relates to a switching system and an ATM virtual path capacity setting method, and more particularly to a technique for setting a capacity in accordance with traffic demand with high speed and high accuracy based on lightweight traffic measurement.

[0002]

2. Description of the Related Art As is well known, a circuit switching communication network such as a telephone service employs a known synchronous transfer mode (STM: Synchronous Transfer Mode).
ous Transfer Mode). In this STM network, a line having a fixed band is set in a path set between two exchanges, and a line having various bands cannot be set. When operating a communication network that handles only such fixed bandwidth lines individually,
It has not been possible to flexibly support various communication services and new communication services, and there has been a limit to efficient operation of network resources.

A data communication network such as a packet communication service is provided by known packet communication. In this packet communication network, data from an unspecified number of information sources is converted into packets of various lengths and transferred, but the data from the specified information source affects all other data flowing through the same transmission medium. Therefore, it has not been possible to sufficiently increase the utilization efficiency of the transmission medium while keeping the data loss rate low. Even if the communication network is operated so that the bandwidth of the transmission medium is shared by an unspecified number of people without restriction, it is not possible to flexibly cope with various data characteristics and fluctuations in communication demand, and therefore, this also requires network resources. There was a limit to efficient operation.

[0004] In contrast to the above technologies, in a wideband ISDN which handles a plurality of communication services such as telephone, data communication, and image communication in recent years, an asynchronous transfer mode (ATM:
By using Asynchronous Transfer Mode, all communication information is converted into fixed-length cells (ATM cells) and transferred, so that a unified exchange process independent of the type of communication service can be realized. As a result, it is possible to logically set a band usable between the two exchanges instead of the path as an ATM virtual path.

Further, under the limitation of the capacity of the ATM virtual path,
An unspecified number of information sources can logically set a virtual circuit sharing an ATM virtual path. Information transfer using ATM cells is a technique similar to known time-division multiplexing using time slots. However, time slots for virtual circuits are not allocated in a time-periodical manner and operate in response to a time-varying information transmission request. It is possible to allocate time slots. Various bands can be variably set by changing the number of cells transmitted per unit time. In other words, the ATM is a data transfer mode in which all communication services from voice and data to images can be transmitted in a unified manner. A communication network that performs communication based on this ATM is called an ATM network. The above AT
Regarding the configuration of the transmission path network using the M virtual path, see, for example, “Configuration Method of High-Speed Burst Multiplexing Transmission System” by Sato, Kaneda and Tokizawa (Information and Communication Network Research Group of IEICE, IN 87-84, 1987 )).

Referring to FIG. 1, a description will be given of a functional configuration of a conventional ATM exchange related to setting of the capacity of an ATM virtual path. First, the physical functional configuration will be described. In the cell transmission device 100 of the ATM exchange, a cell transmission interface 101 is provided for each physical medium 200 of a transmission destination, and cell transmission exceeding the capacity of the physical medium 200 is suppressed. It is supposed to. Describing the logical functional configuration, shavers 111, 112,... Are connected to the cell transmission interface 101 for each of the ATM virtual paths 201, 202,. These shavers 111, 112,...
This is to suppress cell transmission exceeding the capacity logically determined for the M virtual paths. Each cell is transmitted between the cell transmission interface 101 and the shavers 111, 112,... At a timing according to a certain scheduling rule.

[0007] Shavers 111, 112, ... and ATM
The virtual paths correspond one-to-one, for example, shaver 1
11 and 112 are ATM virtual paths 201 and 20, respectively.
2, ... are supported. Here, for simplicity, it is assumed that the capacity of each path is the same as the cell transmission speed of the shavers 111, 112,.... Such a shaver 11
1, 112,... Include cell transmission waiting buffers 121, 1.
Are connected to each other, and in each of the cell transmission waiting buffers 121, 122,..., When cells exceeding the virtual path capacity flow in, the buffers 121, 122,.
, And has a timing for transmitting the next cell.

Next, the cell transmission waiting buffers 121, 122,..., Q length (to be described later), shavers 111, 112,. Cells switched according to the physical medium 200 in the ATM exchange are first routed to the cell transmission device 100. Logically, cells are divided for each destination of virtual paths 201, 202,..., And incoming cell flows 301, 302,.

For example, the cell transmission waiting buffer 121 to which the cell stream 301 is added counts the number of cells arriving during a preset period T. The cells arriving at the cell transmission waiting buffer 121 are processed in the order of arrival. If the shaver 111 is in operation, the cells are accumulated in the cell transmission waiting buffer 121 and wait for the turn of processing. The number of cells waiting for transmission is referred to as Q length. Cell transmission waiting buffer 12
The order of the cell processing in 1 is based on the scheduling rule defined by the shaver 111. When the processing order reaches the shaver 111, the cells stored in the cell transmission waiting buffer 121 become the shaver 1.
Then, the packet passes through the cell transmission interface 101 and is transmitted to the physical medium 200.

Next, some conventional ATM virtual path capacity setting methods in the above-described ATM exchange will be described. [Conventional method 1] First, as a conventional method 1, a method of modeling a cell arrival process by a Poisson process will be described. This method is characterized in that the capacity of the ATM virtual path can be easily determined by the only traffic measurement item called the utilization rate of the ATM virtual path. That is, when the buffer size of the cell transmission waiting buffer 121 is b and the cell loss rate target value is CLR, the ATM virtual path 201
The virtual path capacity C that satisfies the cell loss rate target value CLR for the cell flow 301 with the arrival rate λ is calculated by the following equation (1). C = λ · M / ρ (1) where M is a constant for converting the number of cells per unit time into capacity. The usage rate ρ is given by the following equation (2). ρ = CLR ^{1 / (b + 1)} (2)

The above equation (1) is derived based on the following points. That is, the Q length distribution is S, and the Q length is k
If the longer probability is P [S> k], it is known that the following equation (3) is established by the large deviation theory. P [S> k] to Exp (−δ · k) (3) Further, especially when the cell arrives according to the Poisson process, the following equation (4) holds. P [S> k] = Σ _{i = 1,} ... _, ∞ {ρ _i ^ki / ( ^ki )! } · Exp (−ρ _i ) (4) Here, when Molina's formula is applied, the above equation (4) is expressed as an approximate equation (5). P [S> k] to ρ ^{k + 1} (5) The details of this discussion are described in detail in pages 351 to 362 of Communication Traffic Theory (Suzuki, Ganbe).

In the above equation (5), when the Q length k on the left side matches the buffer size b, the cell loss rate target value CL
Equation (6) is obtained when the usage rate ρ at that time is determined so as to match R. CLR = ρ ^{b + 1} (6) When this equation (6) is solved for the usage rate ρ, the above equation (2) is obtained.
Is derived, and the virtual path capacity C that results in the usage rate ρ is derived from the above equation (1).

Next, FIG. 2 is a block diagram showing a functional configuration of an ATM exchange for realizing such a conventional method 1. As shown in FIG. In the following description, the same reference numerals are given to the components already described in FIG. 1 and the description is omitted.

In the ATM exchange, known soft counters 131, 132,...
,... Are attached to the virtual paths 201, 202,... To count the number of cells arriving for each ATM virtual path when the cells are logically distributed to the virtual paths 201, 202,. The arrival rate λ is the arrival rate λ which is the number of cells arriving per unit time for each predetermined cycle T in the setting device 400 that summarizes the counting results of the soft counters 131, 132,.
Is calculated by the following equation (7). λ = N / T (7)

However, in the conventional method 1, there is no guarantee that modeling of the cell flow by the Poisson process is sufficiently valid. That is, the burstiness of traffic is not small enough to be modeled in a Poisson process.
Reports on burstiness and large long-term correlations regarding ATM traffic are reported, for example, in WELeland, MSTaqqu, W. Wi.
llianger, and `` On the self-simil '' by DVWilson
er nature of Ethernettraffic (extender version) '' I
EE / ACM Trans. Networking, vol.2, no.1, pp.1-15, 1994. Therefore, the AT in actual operation by the conventional method 1
Setting the virtual path capacity of the M network is dangerous.

[Conventional method 2] Next, as conventional method 2,
A method of collecting all the arrival times of the cells and restoring the arrival process to set a sufficient virtual path capacity will be described. FIG. 3 shows an AT for realizing the conventional method 2.
It is an example of a functional configuration of an M exchange. In this ATM exchange,
A cell flowing by an external device 500 provided between the cell transmission interface 101 and the physical medium 200 is copied (captured) or a header portion which is a part thereof is copied, and the copied data is communicated. The data is stored in the storage device 510 via the line 501.

For example, when a cell arrives, the storage device 510 stores a time stamp of the arrival time and a cell header. The header of this cell contains an identifier for specifying the ATM virtual path. The storage device 510
Classify these timestamps for each ATM virtual path,
The process of cell arrival for each ATM virtual path is analyzed. This analysis can be done by creating a probability distribution for the arrival process,
There is a method of calculating the average and variance of the arrival intervals, the third moment, and the like.

However, the storage device 510 to which such a time stamp can be added is expensive. The main cause is
This is due to the technical difficulty of time-stamping in sub-microsecond units required to analyze incoming cells. Further, since continuous capture can be performed only while the amount of memory mounted in the storage device 510 permits (capturing for several seconds is performed every few minutes), setting accuracy other than the capturing interval cannot be guaranteed. Also, depending on the specifications of the external device 510, some devices cannot capture only the header information added by the communication provider, and the payload carrying the customer's communication information may be sent to the outside. There must be.

Furthermore, communication disconnection when connecting the external device 500 between the physical interface 101 and the physical medium 200 is inevitable. That is, in this conventional technique, the communication of the customer is hindered due to the setting of the virtual path capacity. The cell flow that can be captured is Shaver 1
Since the cell stream is arranged by 11, 112,..., The burst property will be small, and cells lost in the cell transmission waiting buffers 121, 122,. This has a small effect on low-load operation, but has a large effect on accuracy reduction in setting the virtual path capacity during high-load operation. Therefore, setting the virtual path capacity of the ATM network in actual operation based on the conventional method 2 is prohibited and cannot be executed.

[Conventional method 3] Next, prior art method 3 will be described. Before that, the Q length distribution shown in FIG. 4 will be described. This Q length distribution is created from the traffic data actually measured by the same method as the conventional method 2 described above.
The Q length distribution is a probability distribution of the number of cells (Q length) accumulated in the cell transmission waiting buffers 121, 122,... At the time of arrival of the cell.
The Q length distribution is drawn based on the arrival time data of 10,000 cell traffics. The horizontal axis indicates the number k of cells in the buffer, and the vertical axis indicates the probability P [S> of cells equal to or greater than the number of cells stored in the cell transmission waiting buffers 121, 122,.
k] is expressed as a logarithm. In a range smaller than the reciprocal of the number of digits of the measured number of cells, the Q length distribution has no meaning. For example, if the traffic data captures 130,000 cells, a probability of 10 ^-4 or less has no meaning. In the decay rate of the Q length distribution, the tail indicates that the Q length k is relatively long.

In general, it is known that, in a stochastic process that satisfies stationarity, rarity, and independence, the probability of occurrence of an event follows an exponential distribution (Poisson's law of few numbers). this is,
In the case of the system configuration of FIG. 2, it means that the tail of the Q length distribution is approximated by a straight line. Here, the Q length threshold value and the number of times the Q length threshold value is exceeded will be described. It is technically difficult to grasp the entire distribution of the Q length as described in the conventional method 2. Therefore, in order to partially understand the characteristics of the Q length, a mark (Q length threshold) is put in the middle of the buffers arranged in one line, and the Q length is set to the Q length threshold every time a cell arrives. There is a technique of judging whether or not the number exceeds the threshold value and counting the result (the number of times the Q length threshold value is exceeded). A value obtained by normalizing the number of times the Q-length threshold has been exceeded by the number of arrival cells is the Q-length threshold excess frequency.

Conventional method 3 is characterized in that two points of Q length are used by using the above-mentioned technique, and the Q length distribution is determined by a straight line determined from two Q length threshold excess frequencies obtained therefrom. This is a method of approximating the tail of attenuation. FIG.
The solid line shown in FIG. 7 indicates the Q length threshold value of a certain two points and the corresponding Q length threshold value.
It shows that the Q length distribution is approximated by a straight line passing through the frequency exceeding the long threshold.

The case where two Q length thresholds are set in each of the cell transmission waiting buffers 121, 122,... Of the cell transmission device 100 will be described below with reference to the functional configuration diagram of FIG. Each of the cell transmission waiting buffers 121, 122,... Has a first Q length threshold 151 (1), a second Q length threshold 151 (2), and a first Q length threshold 152, respectively.
(1) and a second Q length threshold value 152 (2),... Are set, and it is determined whether or not the Q length exceeds the Q length threshold value when a cell arrives. Among the determination results, the result that the Q length exceeds the Q length threshold value and the result that the Q length exceeds the first and second Q length threshold values are counted at the same cycle T as the counting of the number of arriving cells. . The counting result is the number of times the Q length threshold has been exceeded.

FIG. 6 shows a table 1100 in which only the data relating to the ATM virtual path 201 among the counting results in the cell transmission device 100 are summarized.
The data counted by the cell transmission device 100 is transmitted from the cell transmission device 100 to the virtual path capacity setting device 400 via the communication line 401 (see FIG. 5).
100 is stored in the setting device 400. In this table 1100, 1110, 1120, 1130,
1140 is a record indicating a column of a counting cycle number, an arrival cell number, a first Q length threshold excess number, a second Q length threshold excess number column, 1111, 1112,... .. Are records of the number of arrival cells corresponding to the counting cycle numbers 1111, 1112,..., 1131, 1132,. The records of the number of times the Q-length threshold has been exceeded, 1141, 1142,..., Are the second records corresponding to the counting cycle numbers 1111, 1112,.
Is a record of the number of times the Q-length threshold has been exceeded.

The setting device 400 operates in the table 1100
Is processed to calculate a set capacity for the ATM virtual path 201. Note that this calculation formula is based on Shioda, Toyoi
zumi, Yokoi, Tsuchiya, Saito, “Self-sizing network: a
new network concept basedon autonomous VP bandwidt
h adjustment, "Proc. of ITC 15, pp. 997-1006, 1997.

First, a record sequence 1120 of the number of arrival cells
Using {d _i }, the virtual path usage rate {ρ _i } in period i is calculated by the following equation (8). _{ρ i = M · d i /} C, i = 1, ......, n (8) However, M is a constant for converting the number of cells per unit time in the capacitance, the capacitance C is set in the ATM virtual path 201 It is. Further, using a record sequence 1130 {q _i ⁽¹⁾ } of the first Q length threshold excess count and a record sequence 1140 {q _i ⁽²⁾ } of the second Q length threshold excess count, According to (9), the first threshold excess frequency {p _i ⁽¹⁾ } and the second threshold excess frequency {p _i ⁽²⁾ } in the cycle i are calculated. p _i ⁽¹⁾ = q _i ⁽¹⁾ / d _i , p _i ⁽²⁾ = q _i ⁽²⁾ / d _i , i = 1,..., n (9)

Further, the Q length distribution attenuation ｛δ in the period i
i｝ is calculated by the following equation (10). Here, s ⁽¹⁾ is the first Q length threshold 151 (1), and s ⁽²⁾ is the second Q length threshold 151 (2). ^{_{δ i = (s (2)}} -s (1)) / (logp i (1) -logp i (2)), i = 1, ......, n (10) In addition, the intercept constant beta _i in period i It is calculated by the following equation (11). β _i = Exp {(s ⁽²⁾ logp ₁ −s ⁽¹⁾ logp ₂ ) / (s ⁽²⁾ −s ⁽¹⁾ )}, i = 1,..., n (11)

Next, the setting capacity C _i in period i is calculated by the equation (12). C _i = (d _i / T) · [1+ {m ₂ / (m ₁ −b)} · log CLR], i = 1,..., N (12) The above m _i ⁽¹⁾ is expressed by the following equation (13). And m _i ⁽²⁾ is as shown in the following equation (14). T is the length of the counting cycle, b is the buffer size, and CLR is the target value of the cell loss rate. m _i ⁽¹⁾ = δ _i logβ _i , i = 1,..., n (13) m _i ⁽²⁾ = {ρ _i / (1-ρ _i )} · (1 / δ _i ), i = 1 , ..., n (14)

The set value C _d of the virtual path capacity to be obtained by using the calculation result to obtain the above result is calculated by the following equation (15).
It is calculated by (16). C _d = Max {C _i | i = 1,..., N} (15) C _d = C _d (n), C _d (i + 1) = (1−α) C _d (i) + αC _i , i = 1,..., N-1 (16) where α is a constant satisfying 0 ≦ α ≦ 1.

However, in the conventional method 3 described above, it is complicated to handle data on the number of times the Q length threshold has been exceeded obtained from the two Q length thresholds. This is because it is necessary to handle twice the data as compared with the case where the Q length threshold value is one point. Also, it is not easy to appropriately select a plurality of Q length thresholds. For example, if a relatively large value is set as the Q length threshold, the Q length threshold excess frequency may not be collected to that point. On the other hand, if a relatively small value is set as the Q-length threshold value, the portion of the Q-length distribution that sharply drops unlike the tail may be measured, and an approximate straight line on the dangerous side may be obtained. In addition, as the distance between the two Q length thresholds increases, the accuracy of approximation of a straight line increases, but it is difficult to find an appropriate interval due to the aforementioned problem.

[0031]

Therefore, the above-mentioned prior art relating to the setting of the ATM virtual path capacity has the following problems. (1) In the conventional method 1, although the ATM exchange can determine the parameters of the virtual path capacity calculation formula with simple traffic measurement items, the ATM cell traffic characteristics are expected to fluctuate only about Poisson process, and the There is a problem that the correlation cannot be considered for a long time. (2) In the conventional method 2, it is possible to accurately calculate a virtual path capacity by capturing cell traffic data by an external device, but an expensive external device is required, and a complicated external device is required by requesting a customer to disconnect the communication. However, there is a problem that the traffic must be captured by the above operation and the data must be analyzed, and the analysis result can be applied only for a short time during which the capture can be performed. (3) In the conventional method 3, it is possible to obtain a straight line that approximates the tail of the attenuation rate of the Q length distribution relatively accurately from the two Q length threshold excess frequencies, and the calculation of the virtual path capacity based on the straight line can be performed. It is possible. However, this method has a problem that it is difficult to select a Q-length threshold, and cannot cope with an inappropriate setting of the Q-length threshold.

The present invention has been made in view of the above circumstances, and does not require a special measurement form such as the external device described above by using data obtained by simple traffic measurement in an ATM exchange. , AT in operation
The purpose is to set the virtual path capacity of the M network.

[0033]

In order to achieve the above object, according to the present invention, as a means relating to an ATM virtual path capacity setting method in an ATM exchange, a cell loss target rate is satisfied for an ATM virtual path terminating at the ATM exchange. In the method for setting the virtual path capacity to be performed, a Q length threshold value is set in a cell transmission waiting buffer associated with the ATM virtual path, and the number of cells arriving at the cell transmission waiting buffer and Number of times that the number of cells (Q length) waiting in the cell transmission waiting buffer exceeds the Q length threshold (Q
The frequency of exceeding the Q length threshold (frequency of exceeding the Q length threshold) is calculated based on the number of arriving cells and the number of times of exceeding the Q length threshold. It is determined whether or not the frequency of exceeding the long threshold is greater than the target cell loss rate CLR, and the number of times that the frequency of exceeding the Q-long threshold is determined to be greater than the target cell loss CLR exceeds a predetermined reference value. Means of setting the virtual path capacity according to whether or not is adopted. Further, in the above means, a means for estimating the attenuation rate of the Q length distribution density function based on the number of arrival cells and the number of times the Q length threshold is exceeded to set the virtual path capacity is also employed. On the other hand, according to the present invention, as a means relating to the ATM exchange, a threshold (Q
(A long threshold value), and a cell transmission waiting buffer attached to the ATM virtual path for which a predetermined period T is set.
Counting means for counting the number of cells arriving at the cell transmission waiting buffer and the number of times the cell transmission waiting buffer Q length has exceeded the Q length threshold (number of times the Q length threshold has been exceeded); The frequency of exceeding the Q length threshold (frequency exceeding the Q length threshold) is calculated, and the ATM is determined according to whether the frequency of exceeding the Q length threshold is greater than the target cell loss rate CLR.
Means for setting the capacity of the virtual path. Further, in this means, the setting device is configured to set the capacity of the ATM virtual path by estimating the attenuation rate of the Q length distribution density function based on the number of arrival cells for the ATM virtual path and the number of times the Q length threshold has been exceeded. Adopt means.

[0034]

According to the present invention, it is possible to set the capacity of the ATM virtual path in the ATM network at high speed and with high accuracy in accordance with the actual traffic demand based on the light traffic measurement. The first characteristic is to grasp the virtual cell loss rate at the Q length threshold in the traffic measurement at the cell level, despite the lightweight measurement of comparing the Q length threshold and the Q length. Is a point. Therefore, the behavior of the cells in the transmission waiting buffer can be grasped relatively easily and accurately, and the capacity of the ATM virtual path according to the traffic demand can be set at high speed and high accuracy. The second feature is that the main term of Q length behavior is estimated in cell-level traffic measurement. Therefore, the Q length distribution in the transmission waiting buffer can be grasped relatively easily, and the capacity of the ATM virtual path according to the traffic demand can be set at high speed and with high accuracy.

Due to these characteristics, in particular, UBR (Unspec
fied Bit Rate) It is possible to set the capacity of the ATM virtual path accommodating the best-effort virtual circuit for starting the bearer class so as to satisfy a certain cell loss rate target value (not a specified value). Further, the resources of the ATM network are appropriately arranged in the ATMATM virtual path,
It is possible to check and record the effective use every day, and it is possible to check that the traffic is flowing evenly throughout the entire network. In addition, new
It is possible to provide information for performing appropriate reconfiguration in a reasonable operation and period in response to a change in network configuration such as relocation.

[0036]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of an ATM switch and an ATM virtual path capacity setting method according to the present invention; In the following description, the same reference numerals are given to the components already described, and the description will be omitted.

[First Embodiment] A first embodiment will be described with reference to FIG. As shown in this figure, in the present embodiment, the Q length threshold 1 is set in each cell transmission waiting buffer 121, 122,... Of the cell transmission device 100 (counting means).
, Are set, and the number of cells arriving at each cell transmission waiting buffer 121, 122,... Is counted by the soft counters 131, 132,. The cell sending device 100 determines whether or not the threshold value has been exceeded. Then, of the determination results, a result in which the Q length exceeds the Q length threshold is counted at every cycle T which is the same as the counting of the number of arriving cells, and the counted result is determined as the number of times the Q length threshold is exceeded. .

FIG. 8 shows a cell transmission device 10 of an ATM exchange.
This is a table in which only the data of the portion related to the ATM virtual path 201 among the count results of 0 is summarized. In the ATM exchange, these data are sent from the cell sending device 100 to the virtual path capacity setting device 400 via the communication line 401, and are stored in the setting device 400 as a table 1000. In this table 1000, 1010, 1
020, 1030 are records indicating columns of the counting cycle number (counting cycle number), the number of arriving cells, and the number of times the Q-length threshold has been exceeded, 1011, 1012,... Are records of the number of arrival cells corresponding to the counting cycle numbers 1011, 1012,..., 1031, 1032,.
11, 1012,... Are records of the number of times the Q threshold has been exceeded.

The setting device 400 processes the data in the table 1000 as described below to calculate the set capacity for the ATM virtual path 201. For the sake of simplicity, it is assumed that there is no missing data in the following description. That is,
If a measurement result in a counting cycle T (for example, 20 seconds) is obtained within a predetermined time (for example, 1 hour),
The total count number n in this case is 180. Further, by using the record column 1020 of the number of arriving cells {d _i} and Q length threshold crossing times 1030 {q _i}, the Q length threshold crossing frequency {p _i} on the basis of the following equation (17) calculate. p _i = q _i / d _i , i = 1,..., n (17)

It is determined whether or not the Q length threshold value excess frequency {pi} is greater than a predetermined target cell loss rate CLR. FIG. 9 shows the Q length threshold value exceeding frequency 104.
It is a table | surface which shows 0 and the determination result 1050. This table 1001 is stored in the setting device 400, and 1041, 1042,.
Are records indicating the Q-length threshold excess frequency corresponding to 1,1012,..., 1051, 1052,.
Are records indicating determination results corresponding to 011, 1012,.... Summarizing the results of the determination result 1050, it is determined whether or not the number of count numbers 1011, 1012,... Determined that the Q value threshold excess frequency is greater than the target cell loss rate CLR exceeds a predetermined reference value. Is determined.

If the number of the count numbers 1011, 1012,... Determined to be “large” exceeds the reference value, a capacity larger than the current virtual path capacity C (for example,
Set the capacity C) which is larger by a predetermined unit capacity to the set value C.
_d . For example, the set value C _d is obtained by the following equation (18). Here, C is the capacity set for the virtual path. C _d = C + △ (18)

Here, it is conceivable to set the same capacity as described above by comparing the number of times the Q-length threshold is exceeded with a predetermined number, but this is not always appropriate. Because the number of times of excess is a value dependent on the number of arriving cells,
This is because normalization by the number of arriving cells provides an appropriate measure. However, if the number of arriving cells is relatively small, there is no large number of samples, so that it is not an appropriate measure. Therefore, when the number of arriving cells is equal to or less than a predetermined value, it is excluded from the above determination. This value which is determined in advance with respect to the number of arriving cells may be obtained by a large number rule based on the number of samples in which the cell loss target rate CLR is included in the required reliability section.

[Second Embodiment] Next, a second embodiment of the present invention will be described. The present embodiment incorporates means for estimating the Q-length distribution attenuation rate using the Q-length threshold excess frequency {p _i }. For example, this means is incorporated in the setting device 400.

[0044] In estimating the Q length distribution attenuation factor, virtual path usage in the periodic i based on the following equation (19) by using the above arrival cells number of records column _{1020 {d i} {ρ i} } are Is calculated. In the equation (19), M is a constant for converting the number of cells per unit time into a capacity, and C is a capacity currently set in the ATM virtual path 201. ρ _i = M · d _i / C, i = 1,..., n (19)

Subsequently, the Q-length distribution decay rate {δ _0i } when the cell arrival in the cycle i follows the Poisson process is calculated based on the equation (20). δ _0i = −log {ρ _i }, i = 1,..., n (20) Further, the record sequence 10 of the Q-length threshold excess frequency {q _i }
Using Q.30, the Q length threshold value excess number {p _i } in period i is calculated based on equation (21). p _i = q _i / d _i , i = 1,..., n (21) Further, the estimated value {δ _1i } of the Q length distribution attenuation rate based on the number of times the Q length threshold has been exceeded is calculated by the following equation (22). Is calculated.
In the equation (22), s is the transmission waiting buffer 1
21 is the Q length threshold value set to 21. δ _1i = −log {p _i } / s, i = 1,..., n (22) Then, the ratio of the two Q length distribution attenuation rates calculated based on the above equations (20) and (22)） w _i ｝ is given by the following equation (23)
It is calculated by w _i = δ _0i / δ _1i , i = 1,..., n (23)

Subsequently, the process proceeds to the step of analyzing the data of the counting cycle of the data number n and calculating the set value of the virtual path capacity. First, the maximum value w of the value calculated by the above equation (23) is calculated by the equation (24). _{w = Max {w i | i} = 1, ......, n} (24) In the same manner, the record sequence number arriving cells {d _i} 1020
Is calculated by equation (25). _{d = Max {d i | i} = 1, ......, n} (25) Then, the set value C _d of the virtual path capacity obtained based on the above calculation result is calculated by the equation (26). In addition,
In equation (26), CLR is a cell loss rate target value, and b is a buffer size. C _d = M · CLR ^{w / d} · d (26)

[Third Embodiment] The third embodiment is different from the first embodiment in that a means for estimating the Q length distribution attenuation rate using the Q length threshold value excess frequency and the average and variance of the number of arriving cells is used. It has been incorporated. For example, this means is incorporated in the setting device 400.

Prior to the description, a general problem related to identifying a moment indicating a characteristic of an object from measured data will be described. When identifying the variance or the third or higher order moment from the measured data of the number of arriving cells in a certain ATM virtual path, a sufficiently large number of samples are required to fit the calculation result in the required reliability section width. It is. However, in the case of measurement in which the same object is repeatedly counted along the time axis, accurate identification cannot be expected even if the number of samples is increased. This is because increasing the number of samples means elapse of time, and accurate identification cannot be performed for an object whose characteristics change with elapse of time.

In order to set the virtual path capacity in accordance with the actual traffic demand with high accuracy due to the general problem of calculating the moment, in the second embodiment, the Q length threshold is exceeded. By using only the average of the frequency and the number of arriving cells that can be identified relatively accurately, the asymptote at the tail of the Q length distribution is a general straight line.

[0050] In contrast, in this embodiment, the following equation using the record 1020 of the number of the arriving cells {d _i} (27)
To calculate the average arrival rate λ during the measurement period. In Soki (27), n is the data size, and T is the counting cycle. λ = Σ _{i = 1, n} d _i / (n · T) (27) Further, the dispersion ratio V of the number of cells arriving during the measurement period is calculated based on the equation (28). V = Σ _{i = 1, n} (d _i −λ) ² / (n · T) (28) Further, a record 1030 of the number of times the Q length threshold value has been exceeded
Based on {q _i }, the Q length threshold value excess frequency p during the measurement period is calculated by equation (29). p = Σ _{i = 1, n} q _i / Σ _{i = 1, n} d _i (29)

Then, using the above calculation results, the Q length distribution attenuation rate δ is estimated based on the following equation (30). Here, C is the capacity of the current ATM virtual path. δ = 2 (C−λ) / V (30) Further, the intercept constant β is estimated by the following equation (31). Note that b is a buffer size. β = p · Exp {2b (C-1 / V)} (31)

Based on these calculation results, the tail of the attenuation rate of the Q length distribution is approximated, and the set value C _{d of the} virtual path capacity to be obtained is obtained.
Is calculated by Expression (32). Note that M is a constant for converting the number of cells into capacity, and CLR is a cell loss rate target value. C _d = M {1 + V · log (b / CLR) / (2b)} (32)

Although the embodiment of the present invention has been described above, the present invention can be applied to a case where the correspondence between the cell transmission waiting buffer and the ATM virtual path is not one-to-one. For example, when the cell transmission waiting buffer and the ATM virtual path correspond to many-to-one, it is necessary to create a setting expression that depends on the number of times the Q-length threshold has been exceeded or a setting expression that estimates the attenuation rate of the Q-length distribution density function. It can be easily inferred from the above embodiment. Also, when the data of the number of arriving cells is classified and collected depending on the priority related to cell discarding indicated in the header of the cell, the setting expression or Q length distribution depending on the number of times the Q length threshold is exceeded is also used. Creating a setting equation for estimating the decay rate of the density function can be easily analogized from the above embodiment.

[0054]

As described in detail above, A according to the present invention
According to the TM exchange and the ATMATM virtual path capacity setting method, as a means related to the ATM virtual path capacity setting method in the ATM exchange, an ATM terminating at the ATM exchange is used.
In the method for setting a virtual path capacity satisfying a target cell loss rate for a virtual path, a Q length threshold value is set in a cell transmission waiting buffer associated with an ATM virtual path, and a cell is set every predetermined period T. The number of cells arriving at the transmission waiting buffer and the number of cells waiting in the cell transmission waiting buffer (Q
Is counted (the number of times the Q-length threshold is exceeded), and exceeds the Q-length threshold based on the number of arrival cells and the number of times the Q-length threshold is exceeded. Frequency (Q length threshold excess frequency) is calculated, and it is determined whether the Q length threshold excess frequency is greater than the target cell loss rate CLR,
Since the virtual path capacity is set according to whether or not the number of times that the Q-length threshold exceeded frequency is determined to be larger than the target cell loss rate CLR exceeds a predetermined reference value, it is called the Q-length threshold exceeded number. By using the traffic measurement items,
It is possible to efficiently estimate the cell loss in the cell transmission waiting buffer associated with the ATM virtual path, and set the ATM virtual path capacity according to the traffic demand with high speed and high accuracy.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a basic functional configuration of a main part of a conventional ATM exchange.

FIG. 2 is a block diagram showing a functional configuration of a main part of an ATM exchange for realizing a conventional method 1 relating to an ATM virtual path capacity setting method.

FIG. 3 is a block diagram showing a functional configuration of a main part of an ATM exchange for realizing Conventional Method 2 relating to an ATM virtual path capacity setting method.

FIG. 4 is a graph for explaining a conventional method 3 relating to an ATM virtual path capacity setting method.

FIG. 5 is a block diagram showing a functional configuration of a main part of an ATM exchange for implementing Conventional Method 3 relating to an ATM virtual path capacity setting method.

FIG. 6 is a table of data stored in a setting device in an ATM exchange for implementing conventional method 3 relating to an ATM virtual path capacity setting method.

FIG. 7 is a block diagram showing a functional configuration of a main part of the ATM exchange according to the first embodiment of the present invention.

FIG. 8 is a first table of data stored in a setting device in the ATM exchange according to the first embodiment of the present invention.

FIG. 9 is a second table of data stored in the setting device in the ATM exchange according to the first embodiment of the present invention.

[Explanation of symbols]

 100 cell transmission device (counting means) 101 cell transmission interface 111, 112 shaver 121, 122 cell transmission waiting buffer 200 physical medium 201, 202 ATM virtual path 301, 302 cell Flow 131, 132 Soft counter 400 Setting device 401, 501 Communication line 500 External device 510 Storage device

Continued on the front page (72) Inventor Shigeo Shioda 3-19-2 Nishishinjuku, Shinjuku-ku, Tokyo Japan Telegraph and Telephone Corporation

Claims (4)

[Claims] 1. A method for setting a virtual path capacity that satisfies a target cell loss rate for an ATM virtual path terminating at an ATM switch, comprising: setting a Q length threshold value in a cell transmission waiting buffer associated with the ATM virtual path; The number of times that the number of cells arriving at the cell transmission waiting buffer and the number of cells waiting in the cell transmission waiting buffer (Q length) exceed the Q length threshold (Q
Is counted based on the number of arriving cells and the number of times the Q length threshold has been exceeded.
Calculate the frequency of exceeding the long threshold (Q long threshold excess frequency), determine whether the Q long threshold excess frequency is greater than the cell loss target rate CLR, Setting the virtual path capacity according to whether or not the number of times that is determined to be greater than the target cell loss rate CLR exceeds a predetermined reference value. 2. The ATM virtual path capacity according to claim 1, wherein the virtual path capacity is set by estimating the attenuation rate of the Q length distribution density function based on the number of arriving cells and the number of times the Q length threshold is exceeded. Setting method. 3. An ATM exchange constituting an ATM network, comprising: a cell transmission waiting buffer attached to an ATM virtual path in which a threshold (Q length threshold) is set for the number of cells to wait (Q length). A count for counting the number of cells arriving at the cell transmission waiting buffer and the number of times the cell transmission waiting buffer Q length exceeds the Q length threshold (the number of times the Q length threshold has been exceeded) at every predetermined fixed period T. Means for calculating the frequency of exceeding the Q length threshold (frequency exceeding the Q length threshold) based on the counting result of the counting means, and determining whether the frequency of exceeding the Q length threshold is greater than the target cell loss rate CLR. A setting device for setting a capacity of an ATM virtual path in accordance with whether or not the ATM virtual path is present. 4. The setting device estimates the attenuation rate of the Q length distribution density function based on the number of cells arriving at the ATM virtual path and the number of times the Q length threshold is exceeded, and sets the capacity of the ATM virtual path. The ATM switch according to claim 3.