JPH09114765A - Distributed data access system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To quickly access only desired data without any damage or loss. SOLUTION: A B-system 200 for holding data is provided with the attribute 215 of the event occurrence time in the record 212 of a table 210. An A-system 100 for accessing the data holds a time stamp 130 composed of the start time 131 and the object time 132 after that or the like for the access object of the data. In the A-system 100, the time stamp 130 is regularly accessed, the present time of a clock 140 is compared with the access object start time, and when the activation time comes, a polling request is originated including a table name, the start time of the access object and the object time. In the B-system 200, when the polling request is received, the attribute 215 of the records of the pertinent table is checked and only the records in a pertinent time band are selected and returned to the A-system.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed data access system for collecting data in one or some systems and accessing and utilizing data held by the systems in other systems. It relates to a method for efficiently accessing only the relevant data in the data access between such systems.

[0002]

2. Description of the Related Art Conventionally, in a distributed data access system in which a system holds a table for collecting and managing data, and another system accesses this table, the accessing side system, that is, the data utilization system I don't know if it has changed,
A general method is to access all the records in a table at regular intervals, check the attributes of the records on the side of the data utilization system, and select or discard the corresponding records (data).

[0003]

The above-mentioned conventional technique has a problem that it is not efficient because it also accesses records that are not required to be collected. In addition, there is a problem that the collected data may be lost when the accessing system is temporarily stopped.

An object of the present invention is to solve the above problems of the prior art and to provide a distributed data access system capable of quickly and efficiently accessing only the relevant record from the relevant table.

[0005]

In order to achieve the above object, a system for collecting and managing data adds an event occurrence time, which is a time when the record is created or updated, to a data record in the table, and accesses the data. When a data access request for a table is received from the request system, a function is provided to sequentially check the event occurrence times of the records, select a record having an event occurrence time that matches the data access request, and return the record to the corresponding system. When the data access request system holds the data access start time and the access target time from the start time for each table, compares the start time with the current time, and determines that it is the data access start time. The table name of the access target and the start time corresponding to the table Characterized in that a function of generating an access request including the target time to the system.

According to the present invention, an event occurrence time is set in advance in a record to be accessed, and in a system that makes an access request, a table name, a start time, and a time stamp consisting of a target time which is a time zone from the start time are used for indication. A data access request is made to another system that holds the data, including the access target of the obtained data. The accessed system selects only the corresponding record from the corresponding table using this time stamp and returns it. Thereby, when data is accessed between the systems, only the target data can be efficiently accessed.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION A specific example of the application of the present invention is data access in a communication network operation system. FIG. 1 shows a conceptual diagram of this type of system. In FIG. 1, an exchange system 200 forming a communication network in an operation system of the communication network holds a plurality of tables 210, and when the event occurrence detecting unit 230 detects an event such as a failure, writes a corresponding record in the corresponding table. On the other hand, operation system 1
At 00, a polling request is issued to the exchange system 200 to collect operation information such as a failure and a log, and the corresponding record is received and written in the collection table 110. In this case, in the operation system 100,
Switch system 2 for events that occur momentarily in the communication network
It is necessary to collect from 00. Events include equipment failures, communication network utilization event logs, and exchange expansion events. Generally, these events have different demands for immediacy of collection depending on their types. The present invention provides rational data access between such systems according to the type of data.

FIG. 2 shows an overall configuration diagram of an embodiment of the present invention. The A system 100 is a system that uses the accessed data as a data access to the B system 200, and the B system 200 is a system that returns the corresponding data when a data access request is received after collecting and holding the data. For example, the A system 100 corresponds to the operation system of FIG. 1, and the B system 200 corresponds to the operation system of FIG.
Corresponding to the exchange system. Although data access can be performed regularly or irregularly, this embodiment will be described by applying a polling method of regularly accessing, which is a general method when the system is normal. Reference numerals 120 and 220 are polling processing programs.

In the B system 200, there are a file, a table, an object, etc. as a form for holding data, and the present invention can be applied to any form. In this embodiment, the table 210 and the record 212 will be explained. . The table 210 means a data group for each data type, and the record 212 means data of an access unit in the table 210. Here, in the B system 200 that holds the accessed table, the record 212 has the event occurrence time attribute 215. In this attribute 215, the update or generation time of the record 212 or the occurrence time of the event represented by the record 212 is set.

On the other hand, in the accessing A system 100, the data access target is represented by a time stamp 130 composed of a start time 131 and a target time 132 which is a time zone from the start time, and this is stored in each storage table 110 for each storage device. Etc. Furthermore, the clocks 140 managed by the system generally have a difference between the systems, and in order to guarantee the polling operation, the polling waiting time, that is, the start waiting time 133 is set to the time stamp 13.
Add to 0. The activation waiting time 133 defines that the target time has elapsed from the start time, and then the polling is performed after a fixed time, and in the ideal case where this can be set to 0, the time stamp 130 indicates that the start time is 1.
Only 31 and target time 132 are required.

In the A system 100, the time stamp 13 in the system is set by the polling processing program 120.
0 is regularly accessed, the current time of the clock 140 and the access target start time 131 are searched, and when it is determined that the polling start time has come, the table name, the access target start time 131, and the access target time 132 are determined. , And issues a polling request to the B system 200. When the system is normal, the current time is the access start time 13
The access target time 132 is added to 1, and the polling is started when the start waiting time 133 is added. In the B system 200, when the polling request is received, the polling processing program 220 looks at the event occurrence time attribute 215 of the record 212 of the corresponding table 210, and records this from the access target start time to the access target time zone. Only the selected item is returned to the A system 100. In this case, there may be other conditions as record selection conditions.

As described above, the data access target record can be generally expressed in a time zone corresponding to the event occurrence time. By using this, only the relevant record can be returned, so that efficient data access can be realized. The access target time indicating the target time of data access is equal to the polling cycle in the periodic access by polling.

Next, the update of the access target start time will be described with reference to FIG. First, both systems 10
A case where 0, 200 is operating normally will be described. Data access is performed periodically at a polling cycle.
The target record at this time is the record updated and generated in the previous cycle. In FIG. 3, a record in which an event occurs in the access target time T ₁ from the access target start time t is a target. When the current time becomes the sum of the access target time T ₁ and the activation waiting time T _{2 at} the access target start time t, the polling request, that is, the data access is started. When one data access is completed, the value of the access target start time t is added by the access target time T ₁ to obtain the next access target start time, and the data access of the next polling cycle is prepared.

Next, the case where the system is out of order will be described. A system 100 polling processing program 1
20 is temporarily stopped due to a bug etc.,
It may restart after that. Even in this case,
By storing the time stamp 130 in a database or a file, its contents can be retained correctly. After restarting the polling processing program 120, comparing the access target start time with the current time,
It can be seen that the data generated in the past has not been accessed. In this case, data can be collected without loss by restarting the data access from the access start time.

In the operation system, data such as traffic, which has high immediacy, is added to information such as exchange expansion information.
It is necessary to access various tables such as non-immediate data. In this case, in the A system 100, the time stamp 130 is stored in the database table for each accessed table. By storing the time stamps in the table, these parameters can be easily updated and can be flexibly dealt with when readjustment of the access target time becomes necessary. Therefore, when the communication time between the systems is changed, or when the time difference between the systems of the clocks, which is the basis of the operation of the respective systems 100 and 200, occurs, or the record generation in the B system 200 is performed. Even when the processing time required for updating differs depending on the system, or when the demand for immediacy of data collection changes, there is an advantage that it can be flexibly dealt with by changing the parameter value of the time stamp. In FIG. 2, it is obvious that the table 210 holding the data accessed by the B system 200 may be a database table or an object.

FIG. 4 shows a flow example of the polling processing program 120 of the A system 100. This processing flow is regularly activated. First, the current time of the clock 140 and the access target start time 131 of the time stamp 130 are compared, and when the current time has passed the access target time (T ₁ ) from the access target start time (t), it is determined that polling is activated (step 401), and waits for the start waiting time (T ₂ ) (step 402), and issues a polling request to the B system 200 (step 403). In contrast,
From B system 200, access start time (t)
To the target time (T ₂ ) from the corresponding table is returned (step 404). The returned record is written in the collection table 110 (step 405), and the access target start time 131 of the time stamp 130 is updated to t + T ₁ (step 4).
06). Then, it is determined whether or not to end the record collection (step 407), and in the case of continuing, step 401
Return to

When the system is temporarily stopped due to an abnormality, the access target start time 131 of the time stamp 130 is not updated. In this case, after restarting the system,
When the polling processing program 120 is started, the current time is compared with the access target start time in step 401, and it can be determined that the polling start time is reached.
If the polling activation is determined again after executing the polling once, it can be determined that the polling activation time is still reached, so the polling is performed several times in succession. In the A system 100, it is conceivable to collect the collected log data for each time zone and process it. For this purpose, it is convenient if the data can be accessed at each time zone. Therefore, even when the data is collectively accessed after the system is stopped and restarted, the data is continuously accessed at the same access target time as the processing time zone. in this way,
Although the number of times of access is increased by accessing the access target time unit, there is an advantage that data can be collected in processing units.

[0018]

As described above, according to the present invention, the event occurrence time is set in advance in the record to be accessed, and in the system that issues an access request, the table name, the start time, and the time zone from the start time are used. By making a data access request to another system that holds the data including the access target of the data designated by the time stamp consisting of a certain target time, only the relevant data can be efficiently accessed. Furthermore, even if the system is suspended, there is an effect that data can be collected without loss. Further, there is also an advantage that the time stamps including the polling start time and the polling target time are expressed in a table in a unified manner so that the readjustment can be easily performed.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a system showing a specific example of an application target of the present invention.

FIG. 2 is an overall configuration diagram of a system according to an embodiment of the present invention.

FIG. 3 is a diagram illustrating updating of an access target start time.

FIG. 4 is a diagram showing an example of a polling processing flow according to the present invention.

[Explanation of symbols]

 100 A system 110 Collection table 120 Polling processing program 130 Time stamp 131 Access target start time 132 Access target time 133 Startup waiting time 140 Clock 200 B system 210 Table 212 Record 215 Event occurrence time 220 Polling processing program

Claims (1)

[Claims] 1. A distributed data access system comprising a first system having one or more tables and a second system for accessing the tables of the first system, wherein the first system is a table. To the record, the event occurrence time that is the time when the record was created or updated is added, and when the data access request for the table is received from the second system, the event occurrence time of the record is sequentially inspected, and It has a function of selecting a record having an event occurrence time that matches a data access request and returning it to the second system, and the second system provides a data access start time and an access target time from the start time. , Hold for each table, compare the start time with the current time, start data access A distribution having a function of issuing an access request including a table name of an access target, the start time corresponding to the table, and the target time to the first system when it is determined Type data access system.