JP3487688B2 - GCRA calculation self-monitoring method for cell rate monitoring device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a cell rate monitoring device in an ATM (asynchronous transfer mode) transmission network, and more particularly to a UNI (User-Network Intemet) which is an interface between a user terminal of the ATM transmission network and the network.
rface: NNI (Network-Nod) which is an interface between the user and the network and the network
e Interface: network / node interface)
(GCRA (Generic Cell Ra) of a cell rate monitoring device for monitoring ATM cell traffic in
teAlgorithm) calculation self-monitoring method for ensuring the reliability of the calculation.

[0002]

2. Description of the Related Art In an ATM transmission network, if each user is allowed to freely send cells, excessive cell sending may cause congestion in the network, resulting in deterioration of network quality. To prevent this, each user monitors whether or not cells are being sent within the range declared in advance when setting up a connection, and if a user tries to send excessive traffic, the quality of other users Policing control is performed to regulate the input of the cell in order to guarantee that. There is a cell rate monitoring device as a device for performing this policing control.

A cell rate monitoring device used in UNI between a user terminal and a network is a UPC (Usage Parameter).
A cell rate monitoring device used for NNI between networks is an NPC (NePC) that monitors excessive traffic between networks.
network Parameter Control (network parameter control) device. The parameters usually used as the traffic parameters monitored by the cell rate monitoring device are the peak cell rate Np and the average cell rate Na, and the units of Np and Na are both cells / second. In the cell rate monitoring device, Np and N of the arriving cell are
a is independently monitored, and if both Np and Na are within the range declared by the user, the arriving cell is judged to be conforming and passed, and if either cell rate exceeds the declared range, A cell that arrives is judged to be a violation, and that cell is unconditionally discarded, or it is passed with a violation tag, and if the load in the network is high, the cell with the violation tag is discarded. Take control.

An algorithm for monitoring Np and Na is VSA (Virtual Scheduling A).
lgorithm) and CSLBA (Continuou)
s-State Leaky Bucket Alg
orithm) etc., and these algorithms are G
CRA (Generic Cell Rate Alg
orithm).

FIG. 9 is a flow chart for explaining an algorithm of VSA which is one method of GCRA. The operation of determining whether or not the inflow cell satisfies the user report will be described with reference to FIG. This algorithm makes a judgment by comparing the theoretical arrival time of the cell obtained from the contents declared by the user with the actual arrival time of the cell.

In the initial setting (S25 of FIG. 9), after the theoretical arrival time TAT, the threshold value τ, and the addition parameter N are set, the arrival time ta of the first arriving cell is determined (S26 in the same figure), The size is compared with the theoretical arrival time TAT set in the initial setting (at step S27).

If "TAT <ta", it is determined that the target cell satisfies the declaration, and the theoretical arrival time TAT = ta is rewritten (S28) (the state at this time is "state: NEW.
_TAT, equivalent to conformance ").

On the other hand, if "TAT <ta" is not satisfied in S27, the threshold value is set to τ and the size of TAT and ta + τ are compared (S29).

At this time, if "TAT> ta + τ" is not satisfied, it is determined that the target cell satisfies the declaration (the state at this time is defined as "state: OK, corresponding to conformity").

In the case of these conforming states satisfying the declaration, the theoretical arrival time TA is calculated using the addition parameter N.
T is incremented to "TAT = TAT + 1 / N" (at step S30).

On the other hand, if "TAT> ta + τ" in S29, it is judged that this cell is in violation of the declaration (S3 of the same).
1) (The state at this time is defined as "state: NG, equivalent to violation").

In this way, the VSA compares the values of TAT and ta, and TAT and ta + τ for cells that sequentially flow in (OK / NG / NEW_TA).
The three states of T) are determined. Further, in the case of other GCRA such as CSLBA, similarly, the three states of (OK / NG / NEW_TAT) are determined according to the size of TAT and ta + τ.

In the above algorithm, CDV (Cell Delay Variation:
The cell delay fluctuation) allowable value τp is the VSA for peak cell rate monitoring using the peak cell rate Np as the addition parameter N, and the theoretical arrival time TAT of the cell used at this time is TATp.

Further, the burst allowable value τa is set as the threshold τ.
Is an average cell rate monitoring VSA using the average cell rate Na as the addition parameter N, and the TAT used at this time is TATa.

On the other hand, if the peak cell interval which is the minimum cell interval is Tp and the average cell interval is Ta, then Tp = 1.
Since there is a relationship of / Np and Ta = 1 / Na, S3 in FIG.
0 is TATp = TATp + T during peak cell rate monitoring
p, when monitoring the average cell rate, TATa = TATa + Ta
Can be

In the cell rate monitor, Np and Na are monitored independently by the GCRA which is composed of a hardware logic circuit.

FIG. 10 shows an example of the circuit configuration of the conformity / violation determination unit of the cell rate monitor, which determines whether the arrival cell is conformity or violation. In the figure, 3 is for Np
CRA operation circuit, 4 is a GCRA operation circuit for Na, and 6 is a G that determines whether or not NG is included in both GCRA operation results.
Reference numeral 5 denotes a CRA determination circuit, and reference numeral 5 denotes a GCRA calculation memory that stores each variable during the GCRA calculation.

When the cell arrival time ta is input to both the Np and Na GCRA operation circuits 3 and 4, both GCRA operation circuits independently monitor the cell rate, and both GCRA operation results (peak, average) = (OK / NG / NEW_TATp,
Whether or not NG is included in any of (OK / NG / NEW_TATa) is determined by the GCRA determination circuit 6, and the determination result Ro
= (Match / violate) is output.

By the way, if a circuit error occurs in both the Np and Na GCRA operation circuits 3 and 4 and the GCRA determination circuit 6 and an error occurs in the operation result, accurate cell rate monitoring becomes impossible, and therefore GCRA is always performed. It is necessary to monitor the calculation result. In order to perform the cell rate monitoring operation in-service without interruption, self-monitoring is advantageous rather than external monitoring. As a configuration of such GCRA operation self-monitoring, there is a method of making the GCRA operation circuit and the GCRA comparison circuit redundant.

FIG. 11 shows an example of one configuration of the dual system in which two NCR GCRA operation circuits and two Na Na GCRA operation circuits are provided to form each GC.
The RA calculation result (OK / NG / NEW_TAT) is compared and collated by a comparison circuit. In the figure, 3 and 13
Is an Np GCRA operation circuit, and 4 and 14 are Na GCRA
Arithmetic circuit, 16 is the arithmetic result of 3 and 13 (OK / NG / N
EW_TATp, OK / NG / NEW_TATp) comparison circuit, 17 is a calculation result of 4 and 14 (OK / N
G / NEW_TATa, OK / NG / NEW_TAT
Comparing circuit for comparing a), 5 and 15 are GCRA operation memories, 6 is a GCRA judging circuit, and 10 is a circuit failure alarm issuing circuit. If any of the comparison and comparison results of the comparison circuits 16 or 17 does not match, an alarm is notified to the device management function.

FIG. 12 shows another configuration example of the GCRA operation self-monitoring, which is a dual system in which two GCRA operation circuits for Np and Na and a GCRA comparison circuit are provided. In the figure, 3 and 13 are Np GCRA operation circuits, 4 and 14 are Na GCRA operation circuits, 6 is an operation result of 3 and 4 (OK / NG / NEW_TATp, OK / NG).
/ NEW_TATa) GCR for conformance / violation
A determination circuit, 18 is the calculation result of 13 and 14 (OK / NG
/ NEW_TATp, OK / NG / NEW_TATa)
GCRA judgment circuit for judging conformity / violation from 12 and 6
Comparing circuits 5 and 15 for comparing and collating whether the judgment results of 18 and 18 are coincident with each other are GCRA operation memories. And
Outputs of GCRA comparison circuits 6 and 18 Ro = (match / violate)
Are compared and collated by the comparison circuit 12, and if they do not match, an alarm is notified to the device management function.

[0022]

The configuration for performing the GCRA operation self-monitoring by making the GCRA operation circuit and the GCRA comparison circuit shown in FIGS. 11 and 12 according to the above-described conventional technology redundant, has three GCRA operation circuits. 4, 13, 1
In addition to providing a total of four circuits of four, there is a drawback that the circuit scale increases because it is necessary to also provide two GCRA operation memories 5 and 15.

Further, in the configuration shown in FIGS. 11 and 12 according to such a conventional technique, the combination of the calculation result of Np and the monitoring result of Na (peak, average) = (OK, OK), (O
K, NG), (OK, NEW_TATa), (NG, O
K), (NG, NG), (NG, NEW_TATa),
(NEW_TATp, OK), (NEW_TATp, N
G) and (NEW_TATp, NEW_TATa), even if an abnormality occurs in any of the 9 patterns of circuits, there is a risk that the circuit abnormality will not be detected until the pattern occurs.

[0024]

In view of the problems in the prior art as described above, the method for self-monitoring the cell rate monitoring apparatus according to the present invention is capable of self-monitoring the cell rate monitoring operation without increasing the circuit scale. Another object of the present invention is to provide a cell rate monitoring device capable of self-monitoring while performing a normal cell rate monitoring operation in an in-service state.

GCR of cell rate monitoring apparatus according to the present invention
The A operation self-monitoring method has an algorithm operation circuit for independently monitoring the peak cell rate and the average cell rate based on the arrival time of the cell, and a cell rate monitoring device that monitors the amount of cells flowing into the ATM network A periodic cell detection circuit that detects cells that are arriving in a random manner, a self-monitoring parameter creation unit that creates a parameter of theoretical arrival time based on the periodic cell, and a comparison circuit. In the self-monitoring parameter creation unit, based on the theoretical arrival time of the periodic cell, create a plurality of combinations of the parameters of the calculation result pattern of the peak cell rate monitoring algorithm calculation circuit and the average cell rate monitoring algorithm calculation circuit, the parameter concerned. The peak cell rate calculated using each combination of The circuit abnormality is detected by comparing the calculation results of the arithmetic circuit of the monitoring algorithm and the arithmetic circuit of the average cell rate monitoring algorithm with the prediction result by the comparison circuit.

Further, in the GCRA operation self-monitoring method of the cell rate monitoring apparatus according to the present invention, the operation result patterns of the peak cell rate monitoring algorithm operation circuit and the average cell rate monitoring algorithm operation circuit are such that the cell arrival time has a specified value. The first self-monitoring parameter creating unit has a first conforming state that satisfies the condition, a second conforming state that does not satisfy a specified value but satisfies a value to which a threshold value is added, and a violation state that does not satisfy all of the conditions. In creating a plurality of combinations of parameters of the calculation result pattern of the calculation circuit and the average cell rate monitoring algorithm calculation circuit, next to the calculation result pattern including the violation state, the peak cell rate monitoring algorithm calculation circuit and the average cell rate monitoring Any operation result of the algorithm operation circuit Patterns and wherein the creating includes one or more combinations of parameters to be the first filling state.

The GCRA operation self-monitoring method of the cell rate monitoring apparatus according to the present invention further comprises the peak cell rate monitoring algorithm operation circuit and the average cell rate monitoring algorithm operation circuit included after the operation result pattern including the violation state. When any of the calculation result patterns is calculated using the combination of the parameters that are in the first suitable state, the calculation is performed using the actual arrival time of the cell.

[0028]

According to the GCRA operation self-monitoring method of the cell rate monitor according to the present invention, the set self-monitoring pattern Ps is set.
Since the self-monitoring method of comparing and collating (k) with the GCRA operation result Po is used, it is not necessary to make the GCRA operation circuit, the GCRA comparison circuit, and the operation memory redundant configurations.
Only the circuits need to be provided. Therefore, it is possible to suppress the increase in the circuit scale that occurs in the conventional method.

Ps (k) (k = 0 to 8) is calculated as (peak,
Average) = (NEW_TATp, NEW_TATa) (O
K, OK) (OK, NG) (OK, NEW_TATa)
(NG, OK) (NG, NG) (NG, NEW_TAT
a) (NEW_TATp, OK) (NEW_TATp,
In the case of a continuous pattern of (NG), since the states that the GCRAs of the peak cell rate and the average cell rate can take are periodically monitored, the circuit abnormality of the portion of the hardware logic circuit of the GCRA where the probability of execution of operations is low is immediately detected. Can be detected. Moreover, since cells in the cell stream are used, self-monitoring can be performed while performing normal policing operation in service.

When "NG" is included in either the peak cell rate part or the average cell rate part of Ps (k),
The GCRA operation self-monitoring in the next periodic cell can be smoothly performed. In the update process of TATsp and TATsa, there is no problem in the method of using the values obtained by predicting the arrival time ta of the periodic cell independently of the GCRA calculation as TATp and TATa, but the theoretical arrival time TAT obtained as a result of the GCRA calculation is used as it is as TATp. In the method used as the update process for TAs and TATa, the next periodic cell must be NEW_TAT without fail if a violation is forcibly made by including NG in either the peak cell rate part or the average cell rate part of Ps (i).
Therefore, Ps (i + 1) = (NEW_TATp, N
EW_TATa), and simply (peak, average) = (OK, OK) (OK, NG) (OK, NEW_
TATa) (NG, OK) (NG, NG) (NG, NE
W_TATa) (NEW_TATp, OK) (NEW_
TATp, NG) (NEW_TATp, NEW_TAT
By repeating 9 patterns in which (a) is arbitrarily arranged, the GCRA operation self-monitoring result is determined to be abnormal even if no circuit abnormality has occurred. Therefore, when NG is included in either the peak cell rate portion or the average cell rate portion, the next pattern is protected by NEW_TAT (NEW
_TATp, NEW_TATa), the TAT can be protected so that normal GCRA operation self-monitoring can be continued.

Furthermore, it is possible to automatically recover from the abnormal state when the arrival of the periodic cell does not arrive at the scheduled arrival time.

This system utilizes the fact that the arrival sequence ta of the periodic cells can be predicted with certainty, so that the three states of OK, NG, and NEW_TAT can be reliably created by artificially changing the TAT. However, when the periodic cells do not have a predetermined function interval due to a delay in the ATM network, a loss of synchronization in the device, and the like, and ta is not a predetermined value, ta and TATsp and ta and TAT are
The relation of sa is a state (peak, average) artificially created based on Ps (k) = (OK / NG / NEW_TATp,
(OK / NG / NEW_TATa) may be different. Therefore, even if a circuit failure does not occur in the GCRA operation circuit or the GCRA determination circuit, a circuit failure alarm is generated due to an error in the arrival time of the periodic cell. After that, when the arrival time of the periodic cell returns to normal, it is necessary to automatically restore the monitoring result of this self-monitoring method.
The relationship between p and TATsa deviates from Ps (k). Therefore, a protective NEW_TAT that satisfies (peak, average) = (NEW_TATp, NEW_TATa) between Ps (k) (k = 0 to 8) is arbitrarily sandwiched, and T at that time is sandwiched.
Set p = Ta = 0 and once TATp = TATa = t
If reset to a, ta, TATsp, TATsa again
The relationship of Ps (k) is now in agreement with normal GC
RA operation self-monitoring can be continued.

[0033]

BEST MODE FOR CARRYING OUT THE INVENTION A GCRA operation self-monitoring method of a cell rate monitor according to the present invention will be described with reference to the drawings.

FIG. 1 shows G of the cell rate monitoring apparatus of the present invention.
It is a flow chart for explaining an operation principle of a CRA operation self-monitoring method.

Here, a cell whose arrival time ta can be predicted, such as a cell included in a cell flow, which periodically flows, or a cell based on a function of a certain interval between the flowing cells, is temporarily called a periodic cell. In the present invention, this periodic cell is used for GCRA operation self-monitoring.

In FIG. 1, after initial setting (S1 in FIG. 1), it is first determined whether or not the arriving cell (S2 in FIG. 1) is a periodic cell (S3). In order to monitor the traffic parameters, both Np and Na algorithm operation circuits perform normal GCRA operation (see the same S
4).

On the other hand, in the case of a periodic cell, the GCRA operation self-monitoring process is performed.

Since the arrival sequence ta of the periodic cell can be predicted with certainty, it is possible to artificially create three states of OK, NG, and NEW_TAT by operating the theoretical arrival time TAT.

First, the self-monitoring pattern Ps (k) (k =
0 to j-1) a combination of states that the GCRA can take (peak, average) = (OK / NG / NEW_TATp,
Any j of (OK / NG / NEW_TATa), that is, (peak, average) = (OK, OK), (OK, N
G), (OK, NEW_TATa), (NG, OK),
(NG, NG), (NG, NEW_TATa), (NE
W_TATp, OK), (NEW_TATp, NG), and (NEW_TATp, NEW_TATa) are defined by a pattern in which they are arbitrarily arranged in series.

After detecting the periodic cell, k = i-th Ps
(I) and the self-monitoring TAT (T) preset so that the theoretical value of the GCRA operation result is the same as Ps (i).
(ATsp and TATsa) are read from the GCRA calculation memory (at step S5).

Using TATsp and TATsa instead of the theoretical arrival times TATp and TATa of the periodic cell, Np
And Na are calculated by GCRA (at step S6).

The calculation result Po = (OK / NG / NEW)
_TATp, OK / NG / NEW_TATa) and Ps
(I) is compared and collated (S8 of the same), and if they match, the GCRA operation unit determines that they are normal (S9 of the same), but if they do not match, an error is notified to the device management function (S1 of the same).
1).

Then, i is incremented by +1 (mod
uro j-1) (S11), TATsp, TAT
sa is updated (at step S12).

The update calculation of TATsp and TATsa includes
The arrival time of the periodic cell predicted independently of the GCRA calculation is used. First, the theoretical arrival times TATp and T
Instead of ATa, the arrival time of the periodic cell predicted independently of the GCRA operation is replaced with TATp and TATa, and Ps (k) (k = 0 to j-1) is stored in the memory in advance and k = When i, read Ps (i) and
TA so that the theoretical value of CRA calculation is equal to Ps (i)
Appropriately changing Tp and TATa to change TATsp and TATs
Create a).

Further, as a second invention, in FIG.
CRA possible state combinations (peak, average) =
(OK, OK), (OK, NG), (OK, NEW_T
ATa), (NG, OK), (NG, NG), (NG,
NEW_TATa), (NEW_TATp, OK),
(NEW_TATp, NG), (NEW_TATp, N
A sequence of arbitrary j pieces of EW_TATa) arranged in succession is defined as a sequence Q, and this is defined as Ps (k) (k = 0 to j−1,1 ≦
j ≦ 9), and at the position next to the pattern including “NG” in either the peak cell rate portion or the average cell rate portion of the order Q (peak, average) = (NEW_TATp,
Insert one or more (NEW_TATa) to make a total of M (M
Is a non-negative integer) The combination of the inserted patterns is newly set as Ps (k) (k = 0 to M + j−1), and this is Ps (k).
Used as (k).

At the same time, +1 increment (i.e., S11) of i in FIG. 1 is performed by modulo M + j-1. The inserted (NEW_TATp, NEW_TATa) is called a protection NEW_TAT. In addition, TATsp, TATs
For the update calculation of a, the arrival time of the periodic cell predicted independently of the GCRA calculation is used. First, instead of the ideal arrival times TATp and TATa of the periodic cell, the predicted arrival time of the periodic cell calculated independently of the GCRA calculation is used. Arrival time is TATp and TAT
a, and Ps (k) (k = 0 to M + j-
1) is stored in the memory in advance, and when k = i, Ps
(I) is read and the theoretical value of GCRA operation is Ps (i)
TATsp and TATsa are created by appropriately changing TATp and TATa so as to be equal to.

In this TATsp and TATsa updating process, the TA obtained by the GCRA is not required to replace TATp and TATa with the predicted periodic cell arrival time.
It is also possible to use Tp and TATa as they are.

FIG. 2 shows an example in which the present invention is applied to a UPC device. As a periodic cell, for example, a cell that arrives at a 4k period is targeted, and this is tentatively called a 4k period cell. Figure 2
1 is a 4k cycle cell detection circuit, 3 is a peak cell rate N
GCRA operation circuit for p, 4 GC for average cell rate Na
RA operation circuit, 5 is a memory for GCRA operation, 6 is a GCR
An A determination circuit, 11 is a violating cell processing unit, 2 is a GCRA calculation self-monitoring parameter creating unit, 9 is a self-monitoring calculation memory, 7 and 8 are comparison circuits, and 10 is a circuit failure alarm issuing circuit.

When the cell flow is input, the 4k cycle cell detection circuit 1 first determines whether or not the arriving cell is a 4k cycle cell.

When the cell is not a 4k cycle cell, both GCs for Np and Na are used for normal traffic parameter monitoring.
RA calculation circuits 3 and 4 perform normal GCRA calculation. Then, the GCRA determination circuit 6 determines conformity / violation of the monitoring target cell, and the violating cell processing unit 11 discards the cell for which the violation is determined or performs tagging processing, and performs passing processing of the cell for which conformity determination is performed.

When the 4k-cycle cell detection circuit 1 detects a 4k-cycle cell, the 4k-cycle cell is allowed to pass and the GCRA operation self-monitoring is performed. In the GCRA operation self-monitoring parameter creation unit 2, Ps (i), TATsp, TATs
a is generated, Ps (i) is input to the comparison circuits 7 and 8, and T
ATsp is the Np GCRA operation circuit 3, TATsp is N
It is input to the GCRA arithmetic circuit 4 for a. Both G for Np and Na
The CRA operation circuits 3 and 4 perform GCRA operation using TATsp and TATsa instead of TATp and TATa. Then, in the comparison circuits 7 and 8, both Np and Na GCR
A Calculation result combination Po = (OK / NG / NEW_
TATp, OK / NG / NEW_TATa) and Ps
(I) is compared and collated, and when Ps (i) and Po do not match, the circuit failure alarm issuing circuit 10 notifies the device management function of an alarm.

FIG. 3 shows an example in which the present invention is applied to another UPC device. Similar to the example shown in FIG. 2, a case where a 4k periodic cell is targeted as a periodic cell is shown.

In FIG. 3, 1 is a 4k period cell detection circuit,
3 is a GCRA operation circuit for peak cell rate Np, 4 is a GCRA operation circuit for average cell rate Na, 5 is a memory for GCRA operation, 6 and 13 are GCRA determination circuits, 11 is a violating cell processing unit, 2 is a GCRA operation self-monitoring parameter. A creation unit, 9 is a self-monitoring calculation memory, and 12 is a comparison circuit.

Whether or not the arriving cell is a 4k-cycle cell is determined by the 4k-cycle cell detection circuit 1, and if it is not a 4k-cycle cell, both Np and Na GCRA operation circuits 3 and 4 perform normal GCRA operation. And the GCRA determination circuit 1
In 3, the conformance / violation of the monitoring target cell is determined, and the violating cell processing unit 8 discards the cell for which the violation is determined or performs tagging processing, and performs passing processing of the cell for which conformity determination is performed.

When the 4k-cycle cell detection circuit 1 detects a 4k-cycle cell, the 4k-cycle cell is allowed to pass and the GCRA operation self-monitoring is performed. In the GCRA operation self-monitoring parameter creation unit 2, Ps (i), TATsp, TATs
a to generate TATsp as the Np GCRA operation circuit 3,
The TATsp is input to the GCRA operation circuit 4 for Na. When "NG" is further included in either the peak cell rate portion or the average cell rate portion of Ps (i), "violation",
In other cases, the self-monitoring pattern Rs that is "fit"
(I) is defined and input to the comparison circuit 6. N
In both the GCRA operation circuits 3 and 4 for p and Na, the GCRA operation is performed using TATsp and TATsa instead of TATp and TATa, and the result is calculated by the GCRA determination circuit 13
Determine with. This is set to Ro = (conformance / violation), Ro is input to the comparison circuit 12 and compared and collated with Rs (i). If Rs (i) and Ro do not match, the device management function is notified. Send an alarm notification.

4A and 4B show TATsp and TATs when VSA is used as the GCRA algorithm.
It is a flowchart which shows the update process of a.

FIG. 4A shows one method for updating TATsp. Ps (k) (k = 0 to 8) is stored in the memory in advance. Further, the auxiliary parameter δp for monitoring calculation is defined, and the values of these parameters are set in advance and stored in the memory. The arrival time ta of the periodic cell that will arrive next time can be predicted with certainty.
G, NEW_TATp 3 states are artificially created. Therefore, the arrival time ta of the periodic cell is predicted, and this is set as the ideal arrival time TATp. First start (S in Fig. 4 (a)
13) Then, when k = i, the peak cell rate part of Ps (i) is read (S14), and if it is “OK” (S15), TATsp = TATp + τp−.
δp (at step S16). On the other hand, if the read peak cell rate portion of Ps (i) is "NG" (S15 in the same), TA
Tsp = TATp + τp + δp (at step S17), “N
If EW_TATp ”(at step S15), TATsp = TA
Tp + δp (step S18). This completes the update of TATsp (at step S19).

The method of updating TATsa in FIG. 4B is also the same, and the auxiliary parameter δa for monitoring calculation is defined and stored in the memory. Further, the predicted ta is defined as TATa.
First, after the start (S13 of FIG. 4B), the average cell rate part of Ps (i) is read (S20 of the same), and it is "O".
If it is K ”(at step S21), TATsa = TATa +
τa-δa (at step S22), and if "NG" (at step S2)
1) Set TATsa = TATa + τa + δa (S2
3), if “NEW_TATa” (S21), TAT
Let sa = TATa + δa (S24). This is TAT
The update of sa ends (S19 in the same figure).

In FIGS. 4A and 4B, τp, τa,
δp and δa do not have to be fixed values, and can be changed to various numerical values.

FIG. 5 shows the GCRA operation monitoring method of the cell rate monitoring apparatus according to the present invention, which is used for UPC using VSA.
It is the figure which showed the timing at the time of applying to a device. Tp
= 90, Ta = 90, periodic cells arriving periodically are included in the cell flow, and τp = 20, τa
= 20, δp = 5, δa = 5, and TATsp and TATsa are updated by the method described in FIGS. 4A and 4B, and TATp and TATa at that time are predicted independently of the VSA operation. You are using a value. Ps (k) is a combination of patterns (peak, average) = (NEW_TAT
p, NEW_TATa), (OK, OK), (OK, N
G), (OK, NEW_TATa), (NG, OK),
(NG, NG), (NG, NEW_TATa), (NE
W_TATp, OK) and (NEW_TATp, NG) are repeated.

The uppermost row in FIG. 5 indicates the number of arrival period cells (N
o. ) And arrival time ta, and TATsp and TATsa corresponding to the corresponding cell number in the lower stage, and it is possible to determine which state of OK / NG / NEW_TAT from comparison with ta. The bottom table shows Ps of the corresponding cell number.
(K), TATsp and TATsa or ta, Ps (k) and Po comparison, VSA operation self-monitoring result, expected arrival time of the cell of the next number, and expected operation Ps (k +
1), TATsp and TATsa are shown. First, immediately after the power is turned on, since NEW_TAT, Ps (0) = (N
EW_TATp, NEW_TATa) and TATs
Let p = TATsa = 0. The first periodic cell (No.
It is assumed that the arrival time of 1) is ta = 100. At this time, assuming that the VSA operation is normally performed, Po = (NEW_TATp, NEW_TATa), which matches Ps (0), so that the self-monitoring result of the VSA operation is determined to be normal. It When the arrival time of the next second periodic cell (No. 2) is predicted independently of the VSA calculation, (arrival time ta of No. 1) +
(Periodic cell interval) = 100 + 90 = 190
If the update calculation of FIGS. 4A and 4B is performed using ATp and TATa, No. 2 o'clock expected monitoring operation Ps (1) = (O
K, OK), so TATsp = 205, TAT
sa = 205. No. The actual arrival time of the periodic cell of 2 is ta = 190, and if this is subjected to VSA calculation, P = P
Since o = (OK, OK) matches Ps (1), it is determined to be normal. Predicting the arrival time of the next third periodic cell (No. 3) is (arrival time ta of No. 2) + (periodic cell interval) = 190 + 90 = 280, which is TATp.
4 (a) and (b) as the TATa and TATa, the No. Expected monitoring operation at 3:00 Ps (2) = (OK, N
Since G), TATsp = 295, TATsa =
It becomes 305. No. Actual arrival time t of the periodic cell of 3
a = 280, and if this is subjected to VSA calculation, Po =
Since (OK, NG) matches Ps (3), it is determined to be normal. Predicting the arrival time of the next fourth periodic cell (No. 4) is (arrival time ta of No. 3) + (periodic cell interval) = 280 + 90 = 370.
If the update operation of FIGS. 4A and 4B is performed as ATa, N
o. Expected monitoring operation at 4:00 Ps (3) = (OK, NEW_
Since it is TATa), TATsp = 385, TAT
sa = 365. No. The actual arrival time of the periodic cell of 4 is ta = 370, and if this is subjected to VSA calculation, P = P
Since o = (OK, NG) matches Ps (4), it is determined to be normal. The same applies hereinafter. In this way, VS
A arithmetic self-monitoring can be performed.

FIG. 6 is an example of a self-monitoring pattern according to the second invention of the GCRA operation self-monitoring method of the cell rate monitoring device according to the present invention. This is (peak, average) =
(NEW_TATp, NEW_TATa), (OK, O
K), (OK, NG), (OK, NEW_TATa),
(NG, OK), (NG, NG), (NG, NEW_T
ATa), (NEW_TATp, OK), (NEW_T
ATp, NG) for a total of 9 patterns during protection NEW_
TAT (NEW_TATp, NEW_TATa) is inserted once to form a pattern of 18 pattern periods. FIG. 7 is a diagram showing timing when the present invention is applied to a UPC device using VSA. Tp = 9
It is assumed that a periodic cell of 0 and Ta = 90 that periodically arrives is included in the cell flow. TATsp and TATs
The updating of a is performed by the method described in FIGS. 4A and 4B, and TATp and TATa at that time are TATs obtained by VSA calculation.
The values predicted independently of p and TATa are used. P
s (k) is the repetition of a total of 18 patterns in FIG. τ
p = 20, τa = 20, δp = 5, δa = 5,
If Ps (k) is NEW_TAT for protection, δp =-
90, and δa = −90.

The number of arrival period cells (N
o. ) And arrival time ta, and TATsp and TATsa corresponding to the corresponding cell number in the lower stage, and it is possible to determine which state of OK / NG / NEW_TAT from comparison with ta.

The bottom table shows Ps of the corresponding cell number.
(K), TATsp and TATsa or ta, Ps (k) and Po comparison, VSA operation self-monitoring result, expected operation Ps (k + 1) at the next cell number, theoretical arrival times TATp and TATa, TATsp Each parameter and update value for updating TATsa and TATsa are shown.

First, Ps (0) = (protection NEW_TA
Tp, NEW_TATa for protection, and TATsp =
TATsa = 0. First periodic cell (No. 1)
Arriving time was ta = 100. At this time, V
Assuming that the SA operation is performed normally, Po
= (NEW_TATp, NEW_TATa), which matches Ps (0), the self-monitoring result of the VSA operation is judged to be normal.

The theoretical arrival time of the next second periodic cell (No. 2) is TATp = 190, T from the VSA calculation process.
ATa = 190. Then, if the update calculation of FIGS. Expected monitoring operation at 2:00 Ps (1) =
Since (NEW_TATp, NEW_TATa), TATsp = 185 and TATsa = 185. N
o. The actual arrival time of the periodic cell of 2 is ta = 190, and if this is subjected to VSA calculation, Po = (NEW_
Since TATp, NEW_TATa) matches Ps (1), it is determined to be normal.

The theoretical arrival time of the next third periodic cell (No. 3) is TATp = 28 from the VSA calculation process.
No. 0 and TATa = 280, the update operation of FIGS. Since the monitoring expected operation Ps (2) at 3 o'clock = (NEW_TATp for protection, NEW_TATa for protection), TATsp = 190 and TATsa = 190. No. Actual arrival time ta = 28 of the periodic cell of 3
0, and if this is subjected to VSA operation, Po = (NEW_
Since TATp, NEW_TATa) matches Ps (3), it is determined to be normal.

The theoretical arrival time of the next fourth periodic cell (No. 4) is TATp = 370, TAT from the VSA calculation process.
When the update calculation of FIGS. 4A and 4B is performed with ATa = 370, No. Expected monitoring operation at 4:00 Ps (3) = (OK, O
K), TATsp = 385, TATsa =
It becomes 385. No. Actual arrival time t of the periodic cell of 4
a = 370, and if this is subjected to VSA calculation, Po =
Since (OK, OK) matches Ps (4), it is determined to be normal. The same applies hereinafter. In this way, VSA operation self-monitoring can be performed.

On the other hand, FIG. 8 is a diagram showing the timing when the present invention is applied to another UPC device using a VSA. The conditions are exactly the same as those in FIG. There is a case that arrives earlier or later than scheduled without arriving periodically.

No. 1-No. 3 and No. 4 T
Up to ATsp and TATsa are exactly the same as in FIG. However, the periodic cells arriving at equal intervals are No. In 4 I arrived by mistake early. Assuming that no circuit abnormality has occurred, VSA calculation result Po = (NG, NG) and Ps
Does not match (3). In this way, even if no circuit abnormality occurs, the VSA operation self-result is judged to be abnormal due to the abnormality of the periodic cell.

The theoretical arrival time of the next fifth periodic cell (No. 5) is not updated because the VSA operation result is NG, TATp = 385, TATa = 385. Then, if the update calculation of FIGS. The expected monitoring operation Ps (4) at 5 o'clock = (protection NEW_TATp, protection NEW_TATa), so TATsp = 3.
85 and TATsa = 385. No. Assuming that the normal time is returned at the arrival time ta = 460 of the periodic cell of No. 5, if this is subjected to VSA calculation, Po = (NEW_TATp, NE
W_TATa) matches Ps (4), and is automatically restored. The same applies hereinafter.

By inserting the protection NEW_TAT in this way, the VSA operation self-monitoring result can also be automatically recovered after the abnormal state of the periodic cell is recovered.

The above-described embodiment can also obtain the same effect when the cells based on a certain function such as a pseudo-random number are used as the periodic cells.

[0074]

As described above, according to the GCRA operation self-monitoring method of the cell rate monitoring apparatus according to the present invention, the self-monitoring pattern Ps (k) and the GCRA set using the periodic cells included in the arrival cell stream are used. Since the GCRA calculation self-monitoring is performed by comparing and collating the calculation result Po, the GCR
The A arithmetic circuit, the GCRA comparison circuit, and the arithmetic memory do not need to be redundantly configured, and further, the self-monitoring can be performed while the normal cell rate monitoring operation is performed in the in-service state. .

[Brief description of drawings]

FIG. 1 is a flowchart for explaining the operating principle of a GCRA operation self-monitoring method of a cell rate monitoring device according to the present invention.

FIG. 2 is an example in which the present invention is applied to a UPC device.

FIG. 3 is an example in which the present invention is applied to another UPC device.

FIG. 4A is VS as an algorithm of GCRA.
TATsp update process when A is used, (b) is T
It is a flowchart which shows the update process of ATsa.

FIG. 5 is a GCRA of a cell rate monitoring device according to the present invention.
It is a figure showing the timing when the operation monitoring method is applied to the UPC device using VSA.

FIG. 6 is an example of a self-monitoring pattern according to the second invention of the GCRA operation self-monitoring method of the cell rate monitoring device according to the present invention.

FIG. 7 is a diagram showing timing when the present invention is applied to a UPC device using VSA.

FIG. 8 is a diagram showing timing when the present invention is applied to another UPC device using VSA.

FIG. 9: VS, which is one method of GCRA in the related art
It is a flow chart explaining the algorithm of A.

FIG. 10 is an example of a circuit configuration of a conformity / violation determination unit of a cell rate monitoring device in the related art.

FIG. 11 is an example of a redundant configuration of GCRA operation self-monitoring in the prior art.

FIG. 12 is another example of the redundant configuration of the GCRA operation self-monitoring in the prior art.

[Description of Symbols] S3 Periodic cell determination S5 Ps (i), TATsp, TATsa set S4, S6 GCRA operation of Np and Na S8 Po and Ps (i) comparison S11 Increment processing S12 TATsp and TATsa update processing

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Haruhiko Matsunaga 1-1-6 Uchiyuki-cho, Chiyoda-ku, Tokyo Inside Nippon Telegraph and Telephone Corporation (56) Reference JP-A-6-112969 (JP, A) JP-A 5-110585 (JP, A) JP-A 4-192832 (JP, A) JP-A 5-191433 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04L 12/24 H04L 12/26 H04L 12/28

Claims (3)

(57) [Claims] 1. A cell rate monitoring apparatus for monitoring the amount of cells flowing into an ATM network, which comprises an arithmetic circuit of an algorithm for independently monitoring a peak cell rate and an average cell rate based on a cell arrival time. The device has a periodic cell detection circuit that detects cells that arrive periodically, a self-monitoring parameter creation unit that creates parameters for theoretical arrival time based on the periodic cells, and a comparison circuit. In this case, the self-monitoring parameter creation unit creates a plurality of combinations of parameters of the calculation result patterns of the peak cell rate monitoring algorithm calculation circuit and the average cell rate monitoring algorithm calculation circuit based on the theoretical arrival time of the periodic cell. Then
A circuit abnormality is detected by comparing the calculation results of the calculation circuit of the peak cell rate monitoring algorithm and the calculation circuit of the average cell rate monitoring algorithm calculated by using each combination of the parameters with the prediction result and the comparison circuit. A method for performing a self-monitoring GCRA operation of a cell rate monitoring device, which is characterized in that 2. A calculation result pattern of the peak cell rate monitoring algorithm calculation circuit and the average cell rate monitoring algorithm calculation circuit is a first conforming state in which a cell arrival time satisfies a specified value, and does not satisfy the specified value but a threshold value. The self-monitoring parameter creating unit has a second conforming state that satisfies the added value and a violating state that does not satisfy all of the values, and In creating a plurality of combinations of parameters, next to the operation result pattern including the violation state, any operation result pattern of the peak cell rate monitoring algorithm operation circuit and the average cell rate monitoring algorithm operation circuit is the first conformance. 1 for the combination of parameters
The GCRA operation self-monitoring method of the cell rate monitoring device according to claim 1, wherein the method is created by including more than one. 3. A parameter in which any operation result pattern of the peak cell rate monitoring algorithm operation circuit and the average cell rate monitoring algorithm operation circuit included after the operation result pattern including the violation state is the first conforming state. 3. The GCRA calculation self-monitoring method of the cell rate monitoring device according to claim 2, wherein the calculation is performed using the actual arrival time of the cell when the calculation is performed using the combination.