JPS6226745B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to an online information processing system,
Specifically, it relates to a control method for information transfer to a terminal.

[Prior art]

In an online system such as a securities business, reception from a terminal takes priority, so during busy reception times, information output to the terminal may be temporarily delayed within the center device. In such a system, the center device generally has an output queue file, and output information waiting to be transmitted is queued in the file, and outputted in order when it becomes possible to transmit it to the terminal. However, in the past, terminal-specific queue files were often used as output queue files, resulting in the following drawbacks.

(1) Because each terminal has a file, splitting losses inevitably occur.

(2) If the capacity is limited in order to reduce the division loss in (1), the measures to be taken when output queues accumulate beyond the capacity become complicated.

(3) Actual processing in the event of a terminal failure becomes extremely complicated.

On the other hand, in order to overcome the drawbacks of (1) and (2) above, it is possible to use a single sequential output queue file as the output queue file and sequentially queue the output information to the terminal in the order in which it occurs, regardless of the terminal. Although this is possible, there are problems such as the output control becomes extremely complicated and the number of accesses to the output queue file increases dramatically.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a method for easily controlling the output and efficiently accessing the output queue file when using a sequential output queue file in order to avoid division loss of the output queue file to a terminal. It is in.

[Summary of the invention]

The present invention is premised on the use of a sequential terminal output queue file, but separately from this, a queue memory is prepared for each terminal. This queue memory has a completely different function from the conventional queue file for each terminal, and is used as a secondary memory of the sequential output queue file. That is, the number of messages for each terminal that can be queued in the queue memory is limited to a certain extent, and the number of messages that have not been queued is held in the sequential terminal output queue file.

Furthermore, in the present invention, a queue memory management table is prepared in order to simplify the queuing process between the sequential terminal output queue file and the queue memory for each terminal. The registration process from the sequential output queue file to the queue memory is as follows:
This is done in a dedicated task while referring to the above queue memory management table.

[Embodiment of the Invention] FIG. 1 shows a schematic configuration of an online system. In the figure, each terminal 1 is connected to a central processing unit 2 through a common line and/or an individual line.
A main storage device 3 and an external storage device 4 are connected to the central processing unit 2 . In the following description of the present invention, it is assumed that a part of the main storage device 3 is used for a queue memory and a queue memory management table for each terminal, and that the external storage device 4 is used for a sequential terminal output queue file.

FIG. 2 is an explanatory diagram of the process of queuing a response message to a sequential terminal output queue file in response to a service request from a terminal. Second
In the figure, messages input from each terminal 1 are processed by a central processing unit 2 and buffered one after another into a buffer register 21 in the order of processing. In this way, when response messages of a predetermined length are completed, they are written to the sequential terminal output queue file 41 of the external storage device 4. WP is a pointer (latest write pointer) indicating the latest write position of the sequential terminal output queue file 41.
As will be described later, the latest write pointer WP is managed by a queue memory management table, and each time one block of messages is written to the sequential terminal output queue file 41 of the external storage device 4,
Its contents will be updated to indicate the next writing position.

FIG. 3 is an explanatory diagram of the process from registering the message of the sequential terminal output queue file in the queue memory to outputting it to the terminal. In Figure 3,
31 is a queue memory for queuing a fixed number of messages (reply messages) to be output to the terminal for each terminal; 32 is a queue memory management table that manages information necessary for queuing processing of the queue memory 31; A desired area of the main storage device 3 is also allocated. RP is a pointer indicating the latest read position of the sequential terminal output queue file 41 (latest read pointer), and NP is a pointer indicating the lowest read position of the message group remaining unoutputted on the sequential terminal output queue file 41 (unread pointer). output read pointer). These pointers RP and NP are also stored in the queue memory management table 3 in the same way as the latest write pointer WP.
Managed by 2.

Queuing processing from the sequential terminal output queue file 41 to the queue memory 31 is performed using A.
Two dedicated tasks, C and B, are assigned, and a dedicated task C is assigned to output processing from the queue memory 31 to each terminal, and each of these dedicated tasks operates while referring to the queue memory management table 32. .

Task A is a task for queuing a message newly stored in the sequential terminal output queue file 41 into the queue memory 31 for each terminal. If there is a new message in the queue file 41, task A reads the messages after the latest read pointer RP, skips the messages in the terminal area of the queue memory 31 that are full, and reads the messages in the terminal area that is not full. Queuing is performed for each terminal. On the other hand, task B is a task for queuing in the queue memory 31 the message that task A could not register because there was no space in the queue memory 31. In other words, task B refers to the queue memory management table 32, rereads the messages in the sequential terminal output queue file 41 indicated by the unoutput read pointer NP, and queues the messages in the terminal area that has become available in the queue memory 31. , if there is no more space, the read position of the lowest number of the message group in the queue file 41 is set as a new unoutput read pointer NP. Since both tasks A and B perform the queuing process while sequentially referring to the queue memory management table 32, overtaking registration of messages on the queue memory 31 does not occur. Note that detailed operations of tasks A and B will be described later.

Task C is a task of reading out the messages in the queue memory 31, for example, in the order of predetermined terminals or in the order of full terminals, while referring to the queue memory management table 32. The message read from the queue memory 31 is sent to the corresponding terminal via the central processing unit 2. Note that when task C reads the message from the queue memory 31, it updates the queue memory management table 32.

FIG. 4 shows the relationship between the sequential terminal output queue file 41 and the queue memory 31. Now, the writing of messages to the sequential terminal output queue file 41 is n of n blocks.
Assume that -1 record has been completed. At this time, the value of the latest write pointer WP is block number =
n, record number=n. On the other hand, it is assumed that the queuing process by task A has progressed to block n-1 of the sequential queue file 41. Therefore, the values of the latest read pointer RP are block number=n and record number=1. It is also assumed that the unoutput read pointer NP indicates block number=m and record number=m.

When the queuing process has progressed to block n-1 of the sequential terminal output queue file 41, the queue memory 31 stores terminals A, B, and I.
It is still possible to register new messages in this area. In this state, when task A reads the messages after RP in the sequential terminal output queue file 41, message _n1 is newly queued in the terminal A area of the queue memory 31, and message n _o-1 is newly queued in the terminal B area. . Note that the terminal B area is now full, and from now on, when a message from terminal B is input to the sequential queue file 41, it will have to wait in the queue file 41 until there is space in the queue memory 31.

On the other hand, the areas for terminals C and X in the queue memory 31 are already full. Therefore, even in the processing of task B, the message mm in the sequential terminal output queue file 41 indicated by NP is stored in the queue memory 3.
Even if I try to queue to the terminal X area of 1,
The queueing is not possible. In this case, task B moves to the queuing process for the messages remaining in the sequential queue file 41 for terminals other than terminal X, and then attempts to queue the message mm again.

FIG. 5 is a detailed diagram of the queue memory management table 32. That is, the queue memory management table 32
has information on the number of currently registered messages in the queue memory, information on whether messages can be registered after RP, and unregistered pointer information for each terminal, as well as the latest write pointer WP, latest read pointer RP, and unoutput read pointer common to each terminal. Has NP. The contents of the queue memory management table 32 correspond to the states of the sequential terminal output queue file 41 and the queue memory 31 shown in FIG. However, the maximum number of messages that can be registered in the queue memory 31 is 30 for each terminal.
It is said that

In FIG. 5, the unregistered pointer and unoutput read pointer NP have the following relationship. That is,
An unregistered pointer is provided for each terminal, and indicates, for each terminal, the youngest position (pair of block number and record number) in the group of messages remaining in the sequential file 41 in an unoutput state. The unoutput read pointer NP is extracted from the unregistered pointers for each terminal with the lowest number. Note that "-" in the unregistered pointer column means that there is no unoutputted message remaining in the sequential file 41 for the terminal in question.

Next, details of the processing of dedicated task A and task B will be explained.

FIG. 6 is a processing flow diagram of task A. Task A is started by a timer and first sets RP in the working register in the processing unit (step
100), sequential terminal output queue file 4
It is determined whether there is a new message in 1 by comparing RP and WP (step 101). If RP=WP and there is no new message, a timer is set and the process waits for a predetermined period of time (step 107). on the other hand,
If RP<WP, the message indicated by RP is read out from the sequential terminal output queue file 41 (step 102). Next, referring to the queue memory management table 32, it is checked whether the corresponding terminal area in the queue memory 31 can be registered (103), and if it is possible to register, the message is newly queued to the terminal area (104). Next, regarding the queue memory management table (MTBL) 32, count up the number of currently registered messages corresponding to the corresponding terminal, change whether or not future message registration is possible (it will not be possible if it becomes full with new queuing), update the RP, etc. Do (step)
105). After incrementing the contents of the work register (step 106), the process returns to step 101. On the other hand, if it is determined in step 103 that registration is not possible for the corresponding terminal area, the process directly proceeds to step 105. In this case, in step 105, the unregistered pointer is set (when the corresponding unregistered pointer is unset,
RP value is set), RP is updated, etc. Thereafter, the same operation is repeated, and when PR=WP, a timer is set and the process ends.

7a and 7b are processing flow diagrams of task B. FIG. Similar to task A above, task B is also started by a timer, first sets NP in the working register (step 200), and compares NP and RP to see if there is unoutput power in the sequential terminal output queue file 41. Judgment (step
201). If NP=RP and there are no unoutput messages remaining, a timer is set and the process waits for a predetermined period of time (step 213). On the other hand, if NP<RP, the message indicated by the NP is read from the sequential terminal output queue file 41 (step
202), it is checked with reference to the queue memory management table 32 whether or not the message can be registered in the queue memory 31 (step 203). As a result, if registration is still not possible, the next youngest unregistered pointer after NP in the group of unregistered pointers in the queue memory management table 32 is read out and set in the working register (step 214), and steps 201, 202, and 203 are performed. Repeat the process. That is, if the message indicated by NP cannot be registered, the message of the terminal with the next youngest unregistered pointer becomes the processing target.

When it is determined in step 203 that registration is possible, the message is registered in the corresponding terminal area in the queue memory 31 (step 204), and the queue memory management table (MTBL) 32 is updated (step 205). The update processing at this time includes counting up the number of currently registered messages corresponding to the terminal, changing whether or not future messages can be registered, clearing the unregistered pointer, and updating the NP. Note that the queue memory management table 32
The above NP is updated as is if the message indicated by the NP cannot be registered.

Next, the value of the working register is incremented (step 206), and it is determined whether the unoutputted message has been processed up to the RP (step 207). If the RP has been processed, set a timer and wait for a predetermined period of time.
Waits for the next unoutput message to be processed (step 215). On the other hand, if up to RP has not been processed, it is determined whether the next message indicated by the value of the working register has been read (step 208), and if it has been read, the next message is
Returns to 206, but if it has not been read, proceed to the next step.
209 and read out the message from the sequential terminal output queue file 41. Here, whether or not the message has been read can be easily checked by setting the determination bit in the head flag of the message.
Next, it is determined whether the message read from the sequential terminal output queue file 41 can be registered in the queue memory 31 (step 210), and if it is possible to register, the message is registered in the corresponding terminal area (step 210).
211), and updates the queue memory management table 32 (step 212). The update process at this time is step 205.
It is similar to On the other hand, if registration is not possible, the message is returned to the non-output state and the process goes to step 206. Thereafter, the operations from step 206 onward are repeated until reaching RP, and the process ends.

In this way, in the process of task B, once a message that can be registered is found (step 203), from then on,
New queuing progresses by sequentially determining whether the message can be registered as read, one message at a time, up to the RP. This is because in the queue memory management table 32, only the smallest message pointer is registered for each terminal. This is because there is a possibility that a telegram exists.

In addition, in FIGS. 6 and 7, the message reading from the sequential terminal output file is 1
Although it is assumed that reading is performed for each message (one record), it goes without saying that reading may be performed for each block.

〔Effect of the invention〕

As is clear from the above description, the following effects can be obtained according to the present invention.

(1) Making the terminal output queue file a sequential file allows for effective use of external storage devices.

(2) By providing a queue memory and a queue memory management table, output processing has become easier, and by devising the processing of the output read pointer, the terminal output queue file has been made into a sequential file. You can fully overcome your weaknesses.

(3) By appropriately adjusting the queue memory size, it is possible to balance the capacity of each terminal and the load on the central processing unit.

[Brief explanation of the drawing]

Figure 1 is an overall configuration diagram of an online information processing system, Figures 2 and 3 are diagrams showing an embodiment of the present invention, Figure 4 is a detailed diagram of a sequential terminal output queue file and queue memory, and Figure 5 is a diagram of a queue memory. Detailed views of the management table, FIGS. 6 and 7, are processing flow diagrams of the embodiment of the present invention. 1... terminal, 2... central processing unit, 3... main storage device, 4... external storage device, 31... queue memory, 32... queue memory management table, 41
...Sequential terminal output queue file,
A, B, C...dedicated task, NP, RP...read pointer, WP...write pointer.

Claims (1)

[Scope of Claims] 1. An online information processing system comprising a central device and a plurality of terminal devices, in which the central device executes desired processing in response to a service request from the terminal device, and sends response information to the corresponding terminal device. , on the central device side, a first memory means for sequentially storing response information to the terminal device one after another;
A second memory means for queuing and storing the answer information in the memory means for each terminal device, and a management file means for managing the storage status of the answer information in the second memory means for each terminal device, Reading the answer information from the first memory means while referring to the management file means and queuing it to the corresponding terminal device area of the second memory means, and reading the answer information from the second memory means and sending it to the terminal device. An online information processing system characterized by: 2. The central device has first and second dedicated tasks for reading the answer information from the first memory means and queuing it to the corresponding terminal device area of the second memory means, and the first dedicated task The response information of each terminal device whose memory means has free space is read out one after another from the first memory means, and newly queued to the corresponding terminal device area of the second memory means, and the second dedicated task is executed by the first dedicated task. Online information processing according to claim 1, characterized in that the unread response information is read from the first memory means, and is queued to the corresponding terminal device area after waiting for the second memory means to become empty. system. 3. The management file means has, for each terminal device, the current number of registered response information, whether or not it can be registered thereafter, and unread addresses of the first memory means, and further includes the first memory means in which the first and second dedicated tasks are to be executed one after another. 3. The online information processing system according to claim 2, wherein the online information processing system has a read address of .