KR101980581B1 - Method for managing memory of tandem system - Google Patents

Abstract

Disclosed is a method for managing a memory of a tandem system. According to an embodiment of the present invention, the method for managing a memory of a tandem system comprises: a step of buffering an inputted message; and a step of broadcasting the buffered message as a service process. The step of buffering the message includes: a step of calculating the size of an extended memory area; a step of determining whether a memory use amount is exceeded by the size of the extended memory area; and a step of processing the inputted message to secure an available memory space if the use amount is exceeded. Accordingly, if a processing speed of a large amount of inputted messages exceeds the capacity of an initially declared memory for buffering, an extended memory area is allocated, and the allocated memory area is used to process messages. An available space of the extended memory space is continuously secured to efficiently process inputted messages.

Description

[0001] The present invention relates to a memory management method for a tandem system,

The present invention relates to a memory management method, and more particularly, to a memory management method of a tandem system in which a plurality of processes are connected together to process the same problem at the same time.

Generally, a router process is a single process, which buffers a large number of incoming messages, allowing a large amount of buffered messages to be broadcast to neighboring service processes.

Such a router process causes a problem when the capacity of buffering memory initially declared exceeds the processing speed of a service process that processes a large number of input messages and an interim processing thereof.

In order to solve this problem, a local memory area for increasing the processing speed is used as a buffering buffer. However, when the size of the local memory is limited and the memory capacity is exceeded, a problem arises that the message can not be smoothly processed.

Korean Patent No. 10-0333581 (entitled " I / O data compression module operating in a tandem manner)

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and it is an object of the present invention to provide an apparatus and method for allocating an extended memory area when a capacity of a buffer memory for initialization is exceeded, The present invention provides a memory management method of a tandem system that processes an input message using an extended memory area, and smoothly processes an input message by continuously securing free space in an extended memory area.

According to an aspect of the present invention, there is provided a method of managing a memory of a tandem system, the method comprising: buffering an input message; And broadcasting the buffered message to a service process, wherein buffering the message comprises: estimating an extended memory area size; Determining whether the memory usage exceeds an amount of the extended memory area; And, when the usage amount is exceeded, processing the input message to secure a memory free space.

When the memory used for buffering the input message reaches 95% of the calculated size of the extended memory area, the step of reserving the memory space may include selecting a message among the buffered messages according to a preset condition , And the stored message is removed from the extended memory area so that the free space of the extended memory area reaches 40% of the extended memory area size.

In addition, the step of securing the memory free space may include the steps of: when the memory used for buffering the input message reaches 70%, 80%, 85%, and 90% of the estimated extended memory area size, It is possible to output the overuse warning information of the memory area.

In the step of securing the memory space, when the memory used for buffering the input message becomes 95% of the calculated size of the extended memory area, the input time of the buffered messages is in the oldest order The selected message is stored in the storage unit in the form of a memory image and the stored message is removed from the extended memory area so that the free space of the extended memory area reaches 40% of the size of the extended memory area .

In addition, the memory management method of the tandem system according to the present embodiment may include loading (loading) a message stored in the storage unit into the extended memory area when the memory space is secured; And buffering the invoked message.

And the step of recalling includes the step of storing the message stored in the storage unit when a predetermined first period of time elapses from an exceeded usage time, or when free memory space of 50% or more of the estimated extended memory area size is secured, The extended memory area can be called up.

When the memory used for buffering the input message reaches 95% of the calculated size of the extended memory area, the step of reserving the memory space further comprises: The free space of the extended memory area can reach 40% of the size of the extended memory area by abandoning the selected message according to the oldest time.

The extended memory area may include a transmission queue area, a routing table area, a common queue area, a process handle area, and a memory pool area.

In addition, the transmission queue area allocates 4000.25 Kbytes in order to maintain the transmission buffer until the completion of the write / read command and to process 128 specials, and the routing table area is allocated to one router 110 Kbytes are allocated in order to allocate 22 bytes to the required space for each process, and the common queue area is divided into a single task having a high throughput In order to support load balancing per process so as to be bundled into one queue, 60 Kbytes are allocated, and the process handle area can allocate 120 Kbytes to retrieve process information.

The memory pool area is dynamically managed to process the process information and the process queue. The memory pool area is divided into 2,048 bytes of pages for flexible allocation and release. The memory pool area is divided into a header area Can be allocated at the front.

The memory pool area may include a first layer that configures a memory pool and allocates or releases a region on a page basis; A second layer as an upper layer of the first layer, for allocating or releasing an area of the process information or the process specialty queue using the memory pool; And a third layer that is an upper layer of the second layer and sets or clears actual data in the process information or the process specialty queue, To allocate or release each area.

Accordingly, when the capacity of the buffer for initialization declared exceeds the processing speed of a large amount of input messages, the extended memory area is allocated, the input message is processed using the allocated extended memory area, It is possible to smoothly process the input message by continuously securing the free space of the memory area.

1 is a diagram illustrating a memory allocation area structure of a memory management system using a memory management method of a tandem system according to an embodiment of the present invention.
2 is a diagram illustrating a memory management system using a memory management method of a tandem system according to an embodiment of the present invention.
3 is a diagram illustrating a buffer memory management unit according to an embodiment of the present invention.
4 is a diagram illustrating a memory management method of a tandem system according to an embodiment of the present invention.
5 to 6 are diagrams for explaining a process of processing a message when a memory is over according to an embodiment of the present invention.
7 to 8 are diagrams for explaining a process of retrieving and processing a message stored in the storage unit according to an embodiment of the present invention.
9 is a diagram illustrating a structure of an extended memory area according to an embodiment of the present invention.
FIG. 10 is a diagram illustrating a relationship between a process information area and another area according to an embodiment of the present invention. FIG.
11 to 12 are diagrams illustrating a structure of a memory pool area according to an embodiment of the present invention.
13 is a diagram illustrating a page area according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. The embodiments described below are provided as examples for allowing a person skilled in the art to sufficiently convey the ideas of the present invention. The present invention is not limited to the embodiments described below and may be embodied in other forms.

FIG. 1 is a diagram illustrating a memory allocation area structure of a memory management system using a memory management method of a tandem system according to an embodiment of the present invention. FIG. 2 is a block diagram of a tandem system according to an exemplary embodiment of the present invention. Fig. 2 is a diagram illustrating a memory management system using a memory management method.

A memory management system (hereinafter collectively referred to as a " memory management system ") that uses a memory management method of a tandem system according to an embodiment of the present invention stores the capacity of an initially declared buffering memory, When the processing speed is exceeded, an extended swap memory area is allocated by using the router process 10, a message input by using the allocated extended memory area is processed, And is continuously provided to smoothly process the input message.

To this end, the memory management system may include a buffer memory management unit 100 and a storage unit 200.

When the extended memory area is allocated by the router process 10, the buffer memory management unit 100 processes the input message using the allocated extended memory area, and continuously reserves the free space of the extended memory area, Messages can be processed smoothly.

Specifically, the buffer memory management unit 100 allocates the extended memory area at the request of the operator, calculates the size of the allocated extended memory area, and determines whether the memory usage exceeds the calculated size of the extended memory area When the usage amount is exceeded, the input message is stored in the storage unit 200 and the memory space can be secured.

The storage unit 200 may be implemented as a physical storage unit such as a storage disk so as to receive and store a message from the extended memory area or to transmit the stored message to the extended memory area.

FIG. 3 is a diagram for explaining the buffer memory management unit 100 according to an embodiment of the present invention in more detail.

The buffer memory management unit 100 allocates an extended memory area and processes the input message using the allocated extended memory area. In order to smoothly process the input message by continuously reserving free space in the extended memory area A buffering engine 110, a broadcasting engine 120, a monitoring module 130, an image download / loading module 140, and a message removal module 150.

The buffering engine 110 is provided for buffering messages entered according to the router buffering queue and the broadcasting engine 120 may broadcast the buffered messages to the service process 20. [

The monitoring module 130 may determine whether the memory usage exceeds an amount of the extended memory area calculated by calculating the size of the allocated extended memory area and the image downloading / The selected message may be stored in the storage unit 200 in the form of a memory image or the message stored in the storage unit 200 may be loaded into the extended memory area.

The message removal module 150 can remove the selected message according to predetermined conditions without storing it.

4 is a diagram illustrating a memory management method of a tandem system according to an embodiment of the present invention. In the memory management method of the tandem system according to the embodiment of the present invention, when the buffer memory management unit 100 allocates the extended memory area, the size of the extended memory area is calculated (S410), the input message is buffered (S420) , It is determined whether the memory usage exceeds the calculated size of the extended memory area at step S430. If the used memory area is exceeded at step S430-Y, the input message is stored in the storage unit 200 at step S440. (S450).

More specifically, for example, when the buffer memory management unit 100 has 95% of the size of the extended memory area used for buffering the input message in order to secure memory space, Some of the messages are selected and stored in the storage unit 200 and the stored messages are removed from the extended memory area so that the free space of the extended memory area can reach 40% have.

In addition, when the memory space is secured, the memory management unit loads the message stored in the storage unit 200 into the extended memory area (S460), and buffers the loaded message (S470).

The memory management unit can smoothly process the input message by broadcasting the buffered message to the service process 20 (S480).

5 to 6 are diagrams for explaining a process of processing a message when a memory is over according to an embodiment of the present invention.

The buffer memory management unit 100 according to an exemplary embodiment of the present invention deletes the size of the extended memory area that is used for buffering the input message, Of the buffered messages are stored in the storage unit 200 in the form of a memory image and the stored messages are removed from the extended memory area , The free space of the extended memory area can reach 40% of the size of the extended memory area.

For example, the buffer memory management unit 100 may allocate a memory area occupying 95% of the size of the extended memory area, which is used for buffering the input message, The free space of the extended memory area is reduced to 40 times the size of the extended memory area by abandoning the selected message according to the old input time of the further input message or the existing buffered message, %. ≪ / RTI >

At the time when the memory used for buffering the input message reaches 70%, 80%, 85%, and 90% of the size of the expanded memory area, the buffer memory management unit 100 stores the excess memory area It is possible to cause the danger warning information to be output.

7 to 8 are views for explaining a process of retrieving and processing a message stored in the storage unit 200 according to an embodiment of the present invention.

In the buffer memory management unit 100 according to the embodiment of the present invention, when the memory space is secured, the memory management unit can load the message stored in the storage unit 200 into the extended memory area and buffer the loaded message .

More specifically, for example, the buffer memory management unit 100 determines whether or not a predetermined first period elapses from an exceeded usage time, a memory free space of 50% or more of the size of the extended memory area, The message stored in the storage unit 200 can be loaded into the extended memory area.

FIG. 9 is a view for explaining a structure of an extended memory area according to an embodiment of the present invention. FIG. 10 is a diagram illustrating a structure of an extended memory area according to an embodiment of the present invention. Respectively.

The extended memory area according to an exemplary embodiment of the present invention uses a large amount of extended memory for managing process information and a large number of messages required for broadcasting, and its size is the number of processes registered in 'PCONF' And the allocated extended memory area may be divided into several areas.

Specifically, the extended memory area may be composed of a transmission queue area, a routing table area, a common queue area, a process handle area, a memory pool area, and other areas.

The transmission queue area can allocate 4000.25 Kbytes (128 * 32000/1024) in order to maintain the transmission buffer until the completion of the write / read command and to process 128 specials.

The routing table area allocates 110 Kbytes in order to allocate 22 bytes to the required space for each process by allocating 5,120 processes that can be connected to one router at a maximum. At this time, the routing table stores process information (Allocated in the memory pool area). In order to find the destination, it is desirable to use binary search for faster speed.

Incidentally, binary search is a search algorithm that can search very quickly in an ordered array. This retrieval method has a disadvantage that it must be always kept in an ordered state for retrieval, but it is often used in a case where there is little waste of memory and the number of comparisons is remarkably small.

The reason for using this search method in the routing table is that it does not happen frequently to register / delete the process dynamically.

In addition, when adding for dynamic addition / deletion, it is possible to maintain the alignment by using the insertion alignment internally, and to rearrange the array to maintain the alignment even when the deletion is performed.

The common queue area allocates 60 Kbytes in order to support process-specific load balancing so that a plurality of processes bundle and process a single task having a high throughput into one queue. In this case, In many single tasks, it is a function that can support process-specific load distribution by allowing multiple processes to be grouped into a single queue.

The process handle area can allocate 120 Kbytes to retrieve process information. The process handle area is used as a process handle to retrieve process information. It uses binary search like the routing table area.

The memory pool area is dynamically managed to process the process information and the process queue, and is divided into 2,048-byte pages for flexible allocation and release. The header area for effectively managing the page is the first Can be assigned in advance.

The method of estimating the area size of the extended memory is as follows. (1) The basic area (4425984 Bytes) used by the router is calculated. (2) The area for storing the process information is calculated by the following first calculation expression.

First calculation formula = (number of processes + 1) / 21 * 2048 Bytes

And (3) the area for storing the text can be calculated by the second calculation equation below.

Second calculation formula = (maximum number of proposals to store / 2) * 2048 Bytes

However, the maximum number of specifications is based on 1010 Bytes, which is reduced by a factor of several for longer specifications. That is, one 2000-byte special is stored in two areas.

④ The area for managing the ② and ③ area calculated above can be calculated by the following third calculation formula.

(2048 bytes) Count / 256 + 1) * 2048 = (2048 bytes) (2048 bytes) (2048 bytes) (2048 bytes)

FIGS. 11 to 12 are diagrams for explaining a structure of a memory pool area according to an embodiment of the present invention, and FIG. 13 is a view for explaining a page area according to an embodiment of the present invention.

The memory pool area used as process information and specialty queue is allocated in page unit (2048 Bytes), and the used area is released and used for other purposes. To implement such a function, the implementation of the memory pool has a three-layer structure.

Specifically, the memory pool area is an upper layer of the first hierarchical first hierarchy that constitutes a memory pool and allocates or releases areas on a page basis, and allocates the process information using the memory pool or the area of the process special queue And a third layer that is an upper layer of the second layer and the second layer that sets or clears the actual data in the process information or process professional queue.

At this time, each upper layer can allocate or release each area by requesting the lower layer.

The first layer - the lowest layer - is a layer that configures a memory pool and allocates and releases areas on a page-by-page basis. The second layer is a layer that allocates the process information or the area of the specialized queue that uses the memory pool. And allocates an area in page units to allocate the area of the process information or the specialized queue. After the use of all the areas of one page is finished, the first layer requests the page area Lt; / RTI > Layer 3 - the uppermost layer - is the layer that installs or removes actual data in the process information or specialty queue, and has the ability to connect or disconnect each zone as needed. The third layer can allocate or release each area by requesting the second layer.

The structure of the memory pool is divided into a header area (information for each page) and an allocation area, and each header area and the allocation area are mapped to 1: 1. As shown in FIG. 12, the page 0 header is mapped to page 0, the page 1 header is mapped to page 1, and the 1: 1 mapping thereafter is maintained. Note that page 0 to page N, which make up the header area, are also mapped in the page area.

In addition, when a new page is allocated, a single linked list is managed so that a blank page that is not used is not searched every time. Therefore, when a new page allocation is requested, the first page of the unused page list is allocated. Conversely, if you want to release a used page, you can add the page to be released to the front of the unused list first.

Each header area consists of a pointer to the allocated area, the next pointer to an unused list, and whether it is currently in use.

The second layer manages the page assigned to the first layer by dividing the allocated page into the process information / specialized queue area. The second layer has two kinds of regions as shown in FIG. 13, and the regions are mutually exclusive. The functions are similar to each other, but the units of the region to be managed are different from each other and the stored data are different from each other. It is independent.

The process of allocating the area is similar to the process information area or the specialized queue area. The process can be summarized as follows.

First, it is possible to check whether there is an unused area in the current free list. Specifically, when there is no unused area, the page is allocated to the first layer and divided into the corresponding area, and each area is registered in the unused list. If there is already an unused area, no page is allocated.

The release of the area is also similar to the process information area or the specialized queue area.

Specifically, if all the pages of the page including the release target area are unused areas, the check page has the current number of used areas. If the value is 0, there is no used area. In this case, the page itself must be released. Also, the area included in the unallocated target page in the unused list must be removed from the unused list.

Next, the target area is released and added to the unused list. However, when checking whether there is an unused area in the free list, the entire page is released, so it is not added to the unused list.

Each area can be used only by allocating only the memory area and storing the data. Therefore, the area can be used only if the pointer of the area is managed systematically. The third layer has the function of installing data in the allocated area and managing the pointers systematically. The area management function of the third layer regards the process information area and the professional queue area as completely separate, and the implemented code can also perform processing completely different from each other.

The process information area and the professional queue area will be described separately. In order to add a new process as shown in FIG. 9, other related areas must be changed as well. Data is added to the routing table so that it can be searched by the process symbol name. In case of having a common queue, the common queue area is also changed. However, the process handle table is set after opening the process when the BOOT-UP message is received from the process.

And deletion of process information should also change the other areas associated with it. After deleting the common queue area and releasing the process information area (requesting to the second layer), information on the process can be removed from the routing table.

It goes without saying that the technical idea of the present invention can also be applied to a computer-readable recording medium having a computer program for performing the functions and methods of the apparatus according to the present embodiment. In addition, the technical idea according to various embodiments of the present invention may be realized in the form of a computer-readable programming language code recorded on a computer-readable recording medium. The computer-readable recording medium is any data storage device that can be read by a computer and can store data.

For example, the computer-readable recording medium may be a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical disk, a hard disk drive, a flash memory, a solid state disk (SSD), or the like.

In addition, the computer readable code or program stored in the computer readable recording medium may be transmitted through a network connected between the computers.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is clearly understood that the same is by way of illustration and example only and is not to be construed as limiting the scope of the invention as defined by the appended claims. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.

10: Router process
20: Service Process
100: buffer memory management unit
110: buffering engine
120: broadcasting engine
130: Monitoring module
140: Image Download / Loading Module
150: Message Removal Module
200:

Claims (11)

delete delete Buffer memory management unit buffering an input message; And
And a buffer memory management unit broadcasting the buffered message to a service process,
Wherein buffering the message comprises:
The buffer memory management unit estimating an extended memory area size;
The buffer memory management unit determining whether the memory usage exceeds an amount of the extended memory area; And
Processing the received message to be stored in the storage unit when the usage amount is exceeded, thereby securing a memory free space;
The step of securing the memory space may include:
When a memory used for buffering the input message reaches 95% of the calculated size of the extended memory area, some messages among the buffered messages are selected and stored in a storage unit according to predetermined conditions, Thereby allowing the free space of the extended memory area to reach 40% of the size of the extended memory area,
The step of securing the memory space may include:
And at the time when the memory used for buffering the input message reaches 70%, 80%, 85%, and 90% of the estimated extended memory area size, the overuse risk warning information of the extended memory area is output The memory management method comprising: Buffer memory management unit buffering an input message; And
And a buffer memory management unit broadcasting the buffered message to a service process,
Wherein buffering the message comprises:
The buffer memory management unit estimating an extended memory area size;
The buffer memory management unit determining whether the memory usage exceeds an amount of the extended memory area; And
Processing the received message to be stored in the storage unit when the usage amount is exceeded, thereby securing a memory free space;
The step of securing the memory space may include:
When a memory used for buffering the input message reaches 95% of the calculated size of the extended memory area, some messages among the buffered messages are selected and stored in a storage unit according to predetermined conditions, Thereby allowing the free space of the extended memory area to reach 40% of the size of the extended memory area,
The step of securing the memory space may include:
If the memory used for buffering the input message becomes 95% of the calculated size of the extended memory area, the message sorted according to the oldest input time of the buffered messages is stored in the form of a memory image And allowing the stored message to be removed from the extended memory area so that the free space of the extended memory area reaches 40% of the extended memory area size. 5. The method of claim 4,
Loading the message stored in the storage unit into the extended memory area when the memory space is secured; And
And buffer memory management unit buffering the invoked message. ≪ Desc / Clms Page number 19 > 6. The method of claim 5,
The loading step may include:
When a preset first period elapses from an exceeded usage time or when a memory free space of 50% or more of the estimated extended memory area size is secured, loading the message stored in the storage unit into the extended memory area A method for managing memory in a tandem system. Buffer memory management unit buffering an input message; And
And a buffer memory management unit broadcasting the buffered message to a service process,
Wherein buffering the message comprises:
The buffer memory management unit estimating an extended memory area size;
The buffer memory management unit determining whether the memory usage exceeds an amount of the extended memory area; And
Processing the received message to be stored in the storage unit when the usage amount is exceeded, thereby securing a memory free space;
The step of securing the memory space may include:
When a memory used for buffering the input message reaches 95% of the calculated size of the extended memory area, some messages among the buffered messages are selected and stored in a storage unit according to predetermined conditions, Thereby allowing the free space of the extended memory area to reach 40% of the size of the extended memory area,
The step of securing the memory space may include:
When the memory used for buffering the input message becomes 95% of the calculated size of the extended memory area, the message input or the buffered message, To allow the free space of the extended memory area to reach 40% of the size of the extended memory area by abandoning the extended memory area. delete Buffer memory management unit buffering an input message; And
And a buffer memory management unit broadcasting the buffered message to a service process,
Wherein buffering the message comprises:
The buffer memory management unit estimating an extended memory area size;
The buffer memory management unit determining whether the memory usage exceeds an amount of the extended memory area; And
Processing the received message to be stored in the storage unit when the usage amount is exceeded, thereby securing a memory free space;
Wherein the extended memory area comprises:
A transmission queue area, a routing table area, a common queue area, a process handle area, and a memory pool area.
The transmission queue region includes:
The transfer buffer is maintained until the write / read command is completed, and at the same time, 4000.25 Kbytes is allocated to process 128 specials,
Wherein the routing table area comprises:
Allocates 110 Kbytes in order to allocate 22 bytes to each space required for each process,
Wherein the common queue area comprises:
Allocating 60 Kbytes to support load balancing per process so that a plurality of processes bundle a single task with a high throughput into one queue,
Wherein the process handle area comprises:
And allocating 120 Kbytes to retrieve the process information. Buffer memory management unit buffering an input message; And
And a buffer memory management unit broadcasting the buffered message to a service process,
Wherein buffering the message comprises:
The buffer memory management unit estimating an extended memory area size;
The buffer memory management unit determining whether the memory usage exceeds an amount of the extended memory area; And
Processing the received message to be stored in the storage unit when the usage amount is exceeded, thereby securing a memory free space;
Wherein the extended memory area comprises:
A transmission queue area, a routing table area, a common queue area, a process handle area, and a memory pool area.
Wherein the memory pool area comprises:
In order to process the process information and the process-specific queue, it is managed dynamically and is divided and processed in units of 2,048 bytes of pages for flexible assignment and release, and a header area for effectively managing pages is allocated at the front Of the tandem system. 11. The method of claim 10,
Wherein the memory pool area comprises:
A first layer for configuring a memory pool, allocating or releasing areas on a page basis;
A second layer as an upper layer of the first layer, for allocating or releasing an area of the process information or the process specialty queue using the memory pool; And
A third layer as an upper layer of the second layer, the third layer for setting or clearing actual data in the process information or process professional queue; / RTI >
Each upper layer,
And allocating or releasing each area by requesting a lower layer.

Images