JPH0257399B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a file access control method in an electronic exchange, and in particular, to a method for controlling file access in an electronic exchange, and particularly to an electronic exchange using a multi-processor system having a plurality of processors each having a file memory. The present invention relates to a file access control method in cases where:

Generally, large electronic exchanges are equipped with multiple networks, and so-called load distribution is performed. In other words, multiple networks are individually controlled by individual call processing processors (CPRs), and these call processors are managed by the main processor, which collectively controls the multiprocessor system. Configure. In this case, each call processor stores subscriber data or station data such as programs for all subscribers belonging to the network it controls in its own file memory (external storage device). This allows call processing to be performed without the main processor being supplied with station data one by one, resulting in the advantage of load distribution. Subscriber data here refers to data necessary to check and control each subscriber, for example whether the subscriber is connected to a central office line or a dedicated line. In addition, a program is procedural information required to perform a task such as a certain exchange connection.

In an electronic switching system configured with such a multiprocessor system and using load distribution control, the file memory belonging to each call processor stores only its own station data.
In order to process a call between the call processor and subscriber B under another call processor, the station data regarding subscriber B is read out from the file memory belonging to the call processor and temporarily stored in its own main memory.
After that, the desired call processing must be performed.
The present invention refers to reading from the partner file memory in this case. That is, file access.

[Conventional technology]

FIG. 9 is a schematic diagram showing the overall configuration of a general electronic exchange to which the access file control method of the present invention is applied. In this figure, NW is a network, and for example, there are multiple time division networks NW. However, in this figure, two networks NW(X) and NW(Y) are illustrated. These networks NW(X) and NW(Y) are each controlled by a processor. In this figure, these processors are call processor CPR.
(X) and CPR (Y), these call processors are managed by the main processor MPR via the inter-channel connection device CCA and the common bus CB. Note that these call processors and main processors are composed of similar components as shown in the figure, and CC is a central control unit,
CHC is channel control device, MN is main memory,
FM is file memory. Since the present invention particularly refers to the file memory among the above-mentioned components, only the meanings of the symbols for the other components will be made clear. SBU is a subscriber terminal, SLC is a subscriber circuit, LC is a subscriber line concentrator, DT is a digital terminal, AT is an analog trunk, REC is a signal receiver, TNG is a signal tone generator, CSE is a common line signaling device, RLC is a remote line concentrator, SWBD is a switchboard, PSC is a switchboard controller, LTCNS is a subscriber line test stand, STCNS is a monitoring test stand, TYP is a typewriter, VDU is a character display unit, and MTU is a magnetic tape unit.

As shown in Figure 9, there are multiple call (call processing) processors in the electronic exchange (two in the figure).
CPR(X) and CPR(Y) are introduced to perform so-called load balancing. call processor here
CPR (X) is your own file memory FM (hereinafter referred to as
FM (also written as X)) is accessed,
The required file (station data mentioned above) is read and call processing is performed. Similarly Coleop Processor CPR
(Y) is your own file memory FM (hereinafter referred to as FM
(also written as Y)), reads the required file, and performs call processing. In this case, when subscriber A who belongs to network NW (X) under call processor CPR (X) makes a call to subscriber B who belongs to network NW (Y) under call processor CPR (Y). , since there is no file related to subscriber B in the file memory FM(X), the file is read by the file memory FM(Y) in the call processor CPR(Y), and the main file in the call processor CPR(X) is read. After it is temporarily stored in the memory MM (hereinafter also referred to as MM(X)), the required call processing is performed.

The series of operations described above, such as accessing the file memory FM (Y) from the call processor CPR (X) and transferring the file from the file memory FM (Y) to the main memory MM (X), is called inter-processor communication. It is done by folding. The communication means in this case are the inter-channel connection device shown in FIG. 9 and the common bus CB in the main processor MPR.

[Problem that the invention seeks to solve]

When one call processor accesses the file memory in the other call processor, communication between the processors using the communication means must be performed each time. Therefore, the first problem is that the communication overhead becomes large and the time required for this communication makes it impossible to speed up call processing.

In addition, for each processor, a processing request section (user)
The second problem is that the method is complicated because a special access routine must be prepared in advance for each file.

[Means for solving problems]

The present invention proposes a file access control method that solves the above problems, and includes a plurality of networks each connected to a large number of subscribers, a plurality of processors that individually control the plurality of networks, and a plurality of processors that individually control the plurality of networks. and a communication means responsible for communication between a plurality of processors, each processor including a file memory for storing station data or programs regarding the subscriber as a file, and reading out the file. In the electronic exchange for setting up lines between subscribers, a plurality of buffer file areas formed in the file memory are provided in each of the processors, and each of the files is further distinguished by a file number and assigned individual attributes. A directory table is provided in each of the processors, and under the management of the directory table, (a) the file for which an access request was made in one of the processors is stored in the file memory or the buffer in the processor; (b) the received access request file is read from the file memory in which the access request file is stored in the other processor and received via the communication means; Store a copy of the file in the buffer file area of No. 1, and perform the line settings using the access request file; (c) In the processor No. 1, store the same access request file as in (a) above. When accessing a file, the copy is read from the first buffer file area and the line settings are performed.

[Effect]

If the file requested by one processor is not stored in its internal file memory,
The request file is read from the file memory in which the request file is stored in the other processor through the communication means, and predetermined call processing is performed, and a copy of the read request file is stored in the internal file memory. The buffer file area is stored in one of a plurality of buffer file areas, and when the first processor accesses the same file as the requested file again, the copy is read from the buffer file area in its own file memory, and Do not perform inter-processor communication using communication means. In this case, information necessary for accessing the requested file is stored in the directory table.

〔Example〕

FIG. 1 is a diagram showing an overview of processing in a file access method based on the present invention. In this figure, the left side of the dotted line in the center is the processor CPR(X) system, and the right side is the processor CPR(Y) system, and both construct the same system. That is, the processing request unit (user) 11(X), 11(Y)
Processing request reception (user reception) section 12 (X), 12
(Y), file access section 13 (X), 13
(Y), input/output control device IOCS14 (X), 14
(Y) and file memory FM (X), FM (Y)
It is. The interprocessor communication section 15 connects both systems. Here, the processing request unit 11
(X) and 11(Y) are programs that operate the electronic exchange, for example, commands are call processing,
Here, it means the requestor of "call processing". Processing request reception units 12 (X) and 12 (Y) are processing request unit 1
It accepts requests from 1(X) and 11(Y) and performs predetermined processing. For example, the processing includes setting the processing request interface, or editing the interface to be used when passing the processing request interface IOCS 14X, 14(Y). Y) can be released from further processing. This also eliminates the complexity of having the processing requesting section for each processor prepare a special access routine for each file in advance and perform a series of processing (the second problem mentioned above). File access section 13 (X), 1
3 (Y) is actually the file memory FM (X), FM
This is the part responsible for accessing (Y), for example,
Request file is internal processor CPR(X), CPR
(Y) or other processor CPR
(Y), CPR (X), and if it is internal, access the file memories FM (X), FM (Y) through IOCS14 (X), 14 (Y) and read the desired file. Or write.
If it is determined that the file is in the other processors CPR(Y) and CPR(X), the other party's file access unit 13(Y) and 1
3(X) and further IOCS14(Y),1
4 (X) and file memory FM (Y),
Access FM(X) and read the desired file. The read file is again IOCS14
(Y), 14 (X) → File access section 13
(Y), 13(X)→interprocessor communication section 15→file access section 13(X), 13(Y) and is returned to the processing request reception section 12(X), 12(Y). The above IOCS is Input/Output
It is an abbreviation for Control System, and is equivalent to a handler or device controller.

In this FIG. 1, to clarify the comparison with the conventional one, conventionally, the processing requesting parts 11(X), 11(Y)
The means provided for accessing file memories FM(X) and FM(Y) are IOCS14(X),
14(Y) and the interprocessor communication unit 15, therefore, the processing requesting units 11(X), 11
The burden on (Y) was quite heavy. On the other hand, in the present invention, the processing request reception units 12(X), 12(Y) and the file access units 13(X), 13(Y)
With the intervention of , the burden is considerably reduced.

FIG. 2 is a diagram showing an outline of processing in an electronic exchange that employs the method according to the present invention. However, the case where processor CPR(X) executes file access is shown. In this figure, similar indications are attached to those already explained, but task 1
1'(X) corresponds to the processing request section described above, and IOCS is
FM side 14'(X), 14'(Y) and CC side (side for inter-processor communication) 15'(X), 15'(Y)
FM1 and FM2 indicate buffer file areas, respectively. However, for the sake of simplicity, a case where two areas are formed will be shown. Note that the main memory MM is divided into multiple work areas, etc., but here they are grouped into one area.
Denoted as MM.

Although various processing requests are issued in task 11'(X), the subject matter of the present invention is call processing requests, particularly file access requests. When a certain file access request occurs, the request is made via the processing request unit interfaces 12' (X) and 13' (Y).
Leading to FMIOCS. The processing request unit interface will be described later, but its outline is to set the data necessary to transfer the task (file access request) to FMIOCS(X).
FMIOCS (X) is MM, FM, fm as shown in the figure.
1. Collaborates with fm2 and directory table. The main memory MM includes an area for temporarily storing data transferred from CCAIOCS(X) and a transfer area. The transfer area may be a transaction (TR) or an overlay buffer that buffers an overlay program. If the data is a file, a transaction is entered. Local station data is stored in the file memory FM as a file. The buffer file areas fm1 and fm2 are areas for temporarily storing files read from other processors, and play an important role in the present invention.
The directory table will be described later, but it is a part that stores, for example, storage locations for request files and the like and FM addresses (file memory addresses). Note that CCAIOCS is a part that controls the inter-channel connection device CCA shown in FIG. 9, that is, a part that controls inter-processor communication.

FIG. 3 is a diagram for diagrammatically explaining the characteristic operation of the present invention.
(X) shows the operations performed in fm1 and fm2. When the processor CPR (X) receives a file access instruction from the processing request unit (task), the requested file is stored in its own file memory.
If it is in FM(X), it is sucked up to overlay area MM ₁ in main memory MM and the process proceeds (). On the other hand, file memory FM (X)
If not, look at the buffer file areas fm1 and fm2, and if any of them contains the requested file, download it to _MM1 and proceed with the process (,). If FM
(X), fm1, and fm2, the requested file is read from another file memory, such as FM(Y), by interprocessor communication, and is temporarily stored in the overlay buffer area _MM2 of the main memory. (), on the other hand, transfers it to overlay area MM ₁ () and proceeds with the process. On the other hand, a copy of the read file is stored. This is stored in the buffer file area fm1 or fm2 (,).
File areas fm1 and fm for these buffers
2 has already stored a copy of another file, clear the file area for one of the buffers before storing. The clearing is performed on the area that stores the most recently accessed (oldest) copy. Next, if the processing request unit requests access to a file in file memory FM(Y), and a copy of this requested file is stored in the buffer file area, the process request will be processed immediately without starting interprocessor communication. The copy can then be transferred to overlay area MM ₁ , realizing high-speed file access (resolving the first problem mentioned above).

The buffer file area itself described above is similar to the concept of a "page" in virtual memory in general information processing. However, a page is an event used when performing hierarchical control in a memory system that has a main storage device and an auxiliary storage device.
It is not of a nature that is used, for example, for call processing in an electronic switching system having a multiprocessor configuration.

FIG. 4 is a diagram showing a simplified configuration of an electronic exchange to which the method of the present invention is applied. In this figure,
The dotted line in the center is processor CPR (X) and CPR
The system of (Y) is divided, and the processor CPR (X)
Subscriber A's terminal under central control unit CC(X)
Central control unit CC for SBU and processor CPR (Y)
(Y) The terminal SBU of subscriber B under its control is connected via speech paths SP(X) and SP(Y).
In this case, the file memory FM(X) contains a file related to subscriber B, that is, station data B(Y).
is not stored, so first open the buffer file area fm1 (fm2 is also fine) in FM(X).
Incorporate into. This import is performed by the file access units 13(X) and 13(Y) via interprocessor communication means (CCA). after that,
When subscriber A reconnects to subscriber B, the line can be set up immediately using the station data B(Y) stored in fm1 in FM(X).

FIG. 5 is a diagram for explaining an example of the processing request section interface. The processing request section interface has already been briefly explained as the processing request receiving section 12(X), 12(Y) in FIG. 1 and the "processing request section interface" in FIG. The processing request unit interface includes RQ, WC, MHA, FLTYP, FMA,
Handles information such as RLDAR and FLNO. The other blank fields contain other information and are not related to file access, so they will not be described in detail. RQ indicates a request for file access, and MMA indicates the address of the request source in main memory. WC is the word count, that is, the number of transferred words. FLTYP means file type, and is information of analects "0" or "1". When it is "0", it indicates that the requested file is stored in the internal file memory (FM(X) on the processor CPR(X) side). On the other hand, when it is "1", it indicates that the requested file is not stored in the internal file memory. When FLTYP="0", the requested file exists in the internal file memory, so FMA is enabled at this time. FMA indicates the file memory address at this time. on the other hand,
When FLTYP="1", the file memory in the other processor (if the other processor is CPR (Y))
Since the request file exists in FM(Y),
In this case, FMA is disabled and RLDAR and
FLNO is enabled and follows the directory table. FLNO is a file number, which is a logical number assigned to each file, and the file memory is accessed using this logical number. When reading a file using this access
Use RLDAR. RLDAR is a relative address when reading. in the form of logical numbers like this
The logical number is input to FMIOCS (see FIG. 2), and FMIOCS starts accessing other processors according to the logical number. At this stage, the processing requesting section is released, so the second problem mentioned above is solved.

FIG. 6 is a diagram showing an example of a directory table. The outline of this directory table has already been explained with reference to Figure 2, but in detail, as shown in the figure, the directory head pointer
It consists of a DP, a priority directory D1, and a general directory D2, and is placed in the main memory MM. The directory head pointer DP has areas TS1 and TS2 representing the table size,
It consists of areas HA1 and HA2 representing head addresses of directories D1 and D2, respectively. Here, the set of TS1 and HA1 is for priority, TS
2. HA2 set is for general use. TS1 and HA1 specify directory D1, and TS2 and HA2 specify directory D2. Furthermore, it flies to a predetermined position according to the file number FLO mentioned above. Here, the general use corresponds to a file that does not need to be copied, and the priority use corresponds to a file whose copy may be stored in the buffer file area. In priority directory 1, PROC, FMA,
Contains BUFA, TIME, etc. PROC, FMA, etc. are stored in the general directory D2. PROC indicates another processor that has the requested file, and FMA indicates the address of the file memory in the other processor. On the other hand, BUFA is the address of the buffer file area within its own processor, and when there is no copy, BUFA="0". When BUFA is other than "0", the usage time TIME records the time from when the copy was created to the present. This TIME is used as a guideline when discarding old copies and replacing them with new copies in order to effectively use the limited number of buffer file areas.

7A and 7B are flowcharts showing the flow of processing when implementing the method based on the present invention on the side that has sent a file access request, and mainly refers to the file access section 13(X) (on the sending side) in FIG. This is the process performed when CPR (X) is used. File access starts at START in FIG. 7A, and first, in step a, it is determined whether the access-requested file is in the own processor or another processor using FLTYP in FIG.
I saw it and cut it in half. If Yes, it is in the own processor, so FMA is enabled, and in step b,
The file memory address (FMA) is obtained from the processing request unit interface, and the file memory FM in the own processor is accessed in step c. After that, the process reaches the state shown in FIG. 7B.

If the result of step a is No, that is, if the file is in another processor, the priority of the access request file is checked in step d.
If it is "general", there can be no copies within the company, so jump to Figure 7A. On the other hand, if it is "priority", there is a possibility that a copy exists within the company, so
In step e, a copy of the access request file is stored in the buffer file area (fm1 or fm
2) Check whether it is stored in . Yes
If so, step f is reached and the copy is used to immediately complete the file access. That is, the line setting data is loaded into a predetermined area (MM ₁ in Figure 3) in the main memory of the processing requester, and the speech path shown in Figure 4 (the connection line at the top in the figure) is established. do. At this point, the file access is completed, and the process returns to RETURN to wait for the next file access.

If the answer in step e is No, that is, the copy of the access request file is not stored in the buffer file area (fm1 or fm2), the communication transaction (TR) is hunted in step g, and then The inter-processor communication indicated by CCA in the center of FIG. 4 is performed by the inter-processor communication control section 15 shown in FIG. That is, as shown in step h, communication information is requested from the directory table (Fig. 6), and which processor among the other processors (PROC in Fig. 6) is requested.
When which address of the file memory (FMA in the figure) is set, a communication request is issued at step i, and the above-mentioned inter-processor communication is started (step j).

The access request file is transferred by another processor through the interprocessor communication in step j, so that the access requesting processor is restarted. This is the restart (step J) in FIG. 7B, which continues to step k. Here, the priority of the transferred file is checked in the same manner as in step d above. "priority"
If so, it is necessary to make a copy, so in step 1, the most recently accessed buffer file area (fm1 or fm2) is found. This is done by referring to TIME (FIG. 6) in the directory table. Once this area is determined, a copy of the file (line setting data) is stored there (step m). On the other hand, if it is determined to be "general" in step k above, the process jumps directly to step n, and, like step f, the process is moved to a predetermined area in the main memory (third Load the line setting data into Figure MM ₁ ). This is the same even after step m has passed.

FIG. 8 is a flowchart showing the processing flow when the method based on the present invention is implemented on the side requested to access a file. The positioning of this flowchart is between step j in FIG. 7A and step J in FIG. 7B. In step a of FIG. 8, in response to step i of FIG. 7A,
Based on the specified communication information (FMA in the directory table in Figure 6), the file memory (FM)
access the specified address. In step b, the file obtained by the access (access result) is sent back to the processor that requested the access via inter-processor communication. In response to this, the system is restarted at step J in FIG. 7B, thus completing a series of file accesses, and the same operation is repeated for the next file access.

〔Effect of the invention〕

As explained above, according to the present invention, by providing a buffer file area, even files in other processors can be obtained without performing inter-processor communication each time, resulting in high-speed processing. File access can be performed and call processing time can be shortened.

In addition, a new access routine is required because it is necessary to judge whether or not to perform the above-mentioned inter-processor communication, and since an element called a file number is introduced to realize this access routine, processing requests The system is freed from preparing and controlling special file routines for all files as in the past, and the system can be simplified.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an overview of processing in the file access method based on the present invention, FIG. 2 is a diagram showing an outline of processing in an electronic exchange that employs the method according to the present invention, and FIG. A diagram for schematically explaining the operation, FIG. 4 is a diagram showing a simplified configuration of an electronic exchange to which the method of the present invention is applied, and FIG. 5 is a diagram for explaining an example of the processing request unit interface. , FIG. 6 is a diagram showing an example of a directory table, and FIGS. 7A and 7B show the processing flow when implementing the method based on the present invention.
FIG. 8 is a flowchart showing the process flow when implementing the method based on the present invention for the side that has requested file access, and FIG. 9 is a flowchart showing the process flow when implementing the method based on the present invention. 1 is a schematic diagram showing the overall configuration of a general electronic exchange to which an access file control method is applied. 13(X), 13(Y)...File access section, 15...Interprocessor communication section, NW(X),
NW (Y)...Network, CPR (X), CPR
(Y)... Processor, FM (X), FM (Y)...
File memory, fm1, fm2...File area for buffers.

Claims (1)

[Claims] 1. A plurality of networks each connected to a large number of subscribers, a plurality of processors that individually control the plurality of networks, and a communication means responsible for communication between the plurality of processors. , each of the processors includes a file memory for storing station data or programs regarding the subscribers as files, and an electronic exchange that reads the files and sets up lines between the subscribers. , comprising a plurality of buffer file areas formed in each of the file memories in each of the processors, and a directory table formed in each main memory for managing files, and under the management of the directory table, 1. A file access control method in an electronic exchange, characterized by controlling access to files in a file memory.