JPH0522426B2 - - Google Patents

Description

[Detailed Description of the Invention] [Summary] Regarding a buffer division method for inter-layer data transfer in an information processing system in which a protocol is layered, an object of the present invention is to reduce the amount of buffer acquired during data division and reduce the frequency of use of the buffer. The purpose of this is to provide a buffer partitioning method that enables buffer partitioning, and the partitioning unit, which is activated when buffer partitioning occurs by adding a header during data transfer from an upper layer to a lower layer, acquires a new buffer area upon activation. a buffer acquisition means, a header copying means for copying a header into the acquired area, a new header data setting means for setting a division position of the divided buffer in the data area start pointer part in the copied header, and a division buffer in the original header. and original header data setting means for setting the head position of the acquired buffer in the header pointer section.

[Field of Industrial Application] The present invention relates to a buffer division method for data transfer between layers of an information processing system in which protocols are layered.

In recent years, research has been widely conducted on information processing systems equipped with communication control functions that comply with the OSI (Open Systems Interconnection) standard standardized by the ISO (International Standards Organization).

In the OSI reference model, the communication functions of open systems are layered into seven functional modules, and each functional layer is called a protocol layer. like this
When communicating using the OSI standard, data is exchanged (transferred) in data units between each layer of multiple layers.
is being carried out. When data is sent from an upper layer to a lower layer, a header is added to the data, and conversely, when data is sent from a lower layer to an upper layer, the data included in the data sent from the lower layer is Removed header. Actual transfer is performed by outputting the head pointer of the transfer data (header section and data section are linked) stored in the memory buffer to the transfer destination layer.

On the other hand, the data unit to be communicated has a fixed capacity (length), and when sending data from an upper layer to a lower layer of OSI, adding a header will overflow the capacity of the data unit. So,
It becomes necessary to divide data to configure multiple data units. Data exchange between layers is performed by buffers as described above, and dividing data means dividing the buffers. However, in the conventional method, there are problems such as a large amount of memory capacity being occupied due to buffer division, and an improvement is desired.

[Prior Art] FIG. 7 is an explanatory diagram of data transferred between layered protocols.

In the case of Figure 7, there are three layers: N+1 layer, N layer, and N-1 layer.
Two layers are shown. In fact, the OSI standard model defines seven layers, from bottom to top: physical layer, data link layer, network layer, transport layer, session layer, presentation layer, and application layer. 7th
Layer N in the figure can be applied to any layer except the top and bottom of these seven layers.

In Figure 7, when data is transferred from the N+1 layer to the N layer, the data at the N+1 layer (or
PDU: Protocol Data Unit)
A header (or PCI: protocol control information) is added to the data and transferred to the N layer.

The N layer receives the entire data including the header sent from N+1 as N layer data, and when further transferring this data to the N-1 layer, a new header created in the N layer is added to the N- layer. Transfer to layer 1.
When data is transferred from an upper layer to a lower layer in this manner, a new header is added to the original data section, resulting in a longer data length. In FIG. 7, when data is transferred from a lower level to an upper level, the headers are sequentially removed, so the data length becomes shorter.

Specifically, the above data transfer is performed using a memory buffer, and when there is insufficient area for transfer, a buffer area is acquired, and when it is no longer needed, the area is discarded.

Conventionally, in data transfer between protocols, one data unit to be transferred is defined to have a certain length (or capacity). However, when data is transferred from an upper layer to a lower layer as described above, the length of the data unit exceeds the specified length due to the addition of a header.

FIG. 8 is an explanatory diagram of conventional buffer division, which is performed when a specified data length is exceeded during transfer.

In the case of FIG. 8, when data (PDU) of the N+1 layer, which has a specified data length, is transferred to the N layer, the specified data length is exceeded by adding a header (PCI). In that case, N+1 data is divided into parts a and b (a+b=PDU of N+1 layer). A header (PCI) is added to each of the divided data a and data b. In this way, two sets of data units are formed, as shown at 1 and 2 in FIG.

FIG. 9 shows an explanatory diagram of a conventional data division operation.

In A of FIG. 9, when the data part 93 in the N+1 layer stored in the memory buffer is transferred to the N layer, a header part 92 is added. Note that the header section is usually provided as an identifier of the data section in which actual data is stored.

However, if the header section 92 is combined with the data section 93, the length of the specified data unit will be exceeded, so the data section 93 will be divided into a and b. In this case, the length of the header and the length of data a become the specified data unit length.

Therefore, as shown in B of FIG. 9, a data section 94 containing only data a is combined with a header section 92 to form one data unit. Then, the header section 92 is copied to the newly acquired buffer area 95 to provide a new header section, and the data b of the data section 94 is also copied to the newly acquired buffer area 96 to provide a new data section. This constitutes another divided data unit.

[Problems to be Solved by the Invention] As described above, when buffer division occurs in data transfer between layers, it is necessary to acquire buffer areas not only for the header area but also for the data area. However, since the buffer uses a finite capacity of memory, if buffer division occurs frequently, the total amount of buffer is constant, so a buffer shortage situation occurs due to an increase in buffer lending. When this buffer shortage occurs, retries are performed until the buffer can be acquired, so it takes time to acquire the buffer, and the processing device is wasted for this purpose, resulting in a problem of lower processing efficiency.

SUMMARY OF THE INVENTION An object of the present invention is to provide a buffer division method that can reduce the amount of buffer acquired during data division and reduce the frequency of buffer use.

[Means to solve the problem] FIG. 1 is a diagram showing the basic configuration of the present invention.

In FIG. 1, reference numeral 10 represents a division processing unit, 11 represents a buffer acquisition means, 12 represents a header copying means, 13 represents a new header data setting means, 14 represents an original header data setting means, and 15 represents a memory.
55 represents a header portion, and 154 represents a data portion.

The present invention provides a storage area for data (division counter and division buffer pointer) representing the number of divided buffers and the position of the divided buffers in the header, and when division occurs, acquires a new buffer in the data section. Instead, only the buffer for the header part of the data after division is obtained, and only the pointer value pointing to the data part stored in the header part is used to indicate the division position.

[Operation] When transferring the contents of the data section 154 of a certain layer of the protocol stored in the buffer to a lower layer, the header section 153 is added. However, since the data length (combined length of the header and data) exceeds the specified unit length, division occurs in the data section 154.

When division occurs, the division processing section 10 is activated, and first the buffer acquisition means 11 of the division processing section 10 is activated.
Acquire a new buffer for the header.
As a result, when the buffer 155 is acquired, the header copying means 12 copies the contents of the header section (referred to as the original header section) 153 to the buffer 155, thereby forming a new header section in the buffer 155.

When copying is completed, new header data setting means 1
3 sets part of the data stored in the new header of the buffer 155. That is, the dividing position of the data section 154 (buffer 154
) is set.

Next, the original header data setting means 14 sets the division buffer pointer (provided according to the present invention), which is the data in the original header part 153, to the new header part 15.
Set to a value representing the starting position of 5.

[Embodiment] FIG. 2 is a process flow diagram of an embodiment, FIG. 3 is a header configuration diagram of an embodiment, and FIG. 4 is a diagram showing an example of an information processing system in which the present invention is implemented.

In the information processing system shown in FIG. 4, a memory 41 is connected to a CPU (central processing unit) 40, and data is transferred to and from each input/output device via interfaces 42, 45 and a line adapter 48 connected to a bus. conduct. The line adapter 48 is connected to a communication line through a modem 49, and communicates with the other party's information processing system in accordance with OSI standards. The communication function of the upper layer of the protocol is executed in the CPU 40, and the buffer area in the memory 41 is used to transfer data between layers.

To explain the header configuration diagram of the embodiment shown in FIG. 3, a pointer 30 to the next buffer in the header section indicates the position of the header of the next succeeding data,
The data area start pointer 31 stores the start pointer of the data section that is the source of this header, the buffer length 32 represents the actual length of the buffer to be used, the data length 33 represents the data length of the data section, and the pool identifier 34 This is the identifier of the application that uses the buffer, and the reception flow control management data 35 and the pointer 38 to the user buffer are used for other purposes. New data provided by the present invention is a division counter 36 and a pointer 37 to the division buffer.

What is stored in the division counter 36 is the number of newly generated headers when division occurs, and a pointer 3 to the division buffer.
7, the header position of the divided data section is set from the header generated by the division.

To explain the processing flow shown in Figure 2,
In this embodiment, the CPU 40 and memory 41 shown in FIG.
It is executed using a buffer of

In the process of transferring data to the lower layer, it is determined whether or not the buffer of the data section has been divided (step 20). When division occurs, a buffer is acquired (step 21), and the contents of the buffer header before division are then copied to the acquired buffer (step 22). After this, the data area head pointer value (31 in FIG. 3) and data length (33 in FIG. 3) are newly set (step 23).

Next, set the pointer value to the next buffer in the header section before division (30 in Figure 3) to 0 (step
twenty four).

Through the above processing flow, the data to be transferred to the lower layer is formed on the buffer, and the data is transferred to the lower layer.

FIG. 5 is a diagram showing the state after the buffer is divided according to the present invention, and FIG. 6 is a diagram showing the relationship between the header section and the data section after division.

In the state shown in FIG. 5, when the header section 52 is added to transfer the data section 53 in the buffer, division occurs in the buffer of the data section 53 and the buffer 54 is
represents the state that has been acquired. Buffer division occurs at the boundary between a and b within the data section 53.

As a result of the division in FIG.
4 and the data section 53 are shown in FIG.

The relationship between each buffer is established using the area of the division counter and the area of the pointer to the division buffer provided in the header according to the present invention, as described with reference to FIG.

That is, the header part (original header part) 5 in FIG.
Data section 53 by the data area start pointer 2
The beginning of the data area within is indicated, and the data length of the line below it represents the length of data area a. In this case, the division counter of the original header section 52 is set to "1", indicating that there is one division buffer. The starting position (address) of the header section (new header section) 54 acquired by the division is set in the area of the pointer to the division buffer that follows this.

Next, the division position of the divided data section 53 is set to the data area head pointer in the new header section 54, and b is set as the data length.

[Effects of the Invention] According to the present invention, even if buffer division occurs, there is no need to acquire the buffer for the data section as in the conventional case, and only the buffer for the header section needs to be acquired, so the amount of buffer usage can be reduced. It is possible to reduce the number of buffers and prevent the occurrence of buffer shortages, thereby improving processing efficiency. Furthermore, conventionally, data was copied when acquiring the buffer of the data section, but since copying of data can be eliminated, processing performance can be improved by eliminating buffer acquisition time and copying time.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the basic configuration of the present invention, FIG. 2 is a processing flow diagram of an embodiment, FIG. 3 is a header configuration diagram of an embodiment, and FIG. 4 is an example of an information processing system in which the present invention is implemented. Figure 5 is a diagram showing the state after buffer division, Figure 6 is a diagram showing the relationship between the header part and data part after division, and Figure 7 is an explanatory diagram of transfer data between layered protocols. FIG. 8 is an explanatory diagram of conventional buffer division, and FIG. 9 is an explanatory diagram of conventional data division operation. In FIG. 1, 10... division processing section, 11... buffer acquisition means, 12... header copying means, 13...
...New header data setting means, 14...Original header data setting means, 15...Memory.

Claims (1)

[Claims] 1. Activated when buffer division occurs by adding a header during data transfer from an upper layer to a lower layer in a buffer division method for data transfer between layers in an information processing system with a layered protocol. The division processing unit 10 includes a buffer acquisition unit 11 that acquires a new buffer area upon startup, a header copy unit 12 that copies a header to the acquired area, and a buffer that is divided into data area head pointer parts in the copied header. In an inter-layer data transfer characterized by comprising a new header data setting means 13 for setting a division position, and an original header data setting means 14 for setting a starting position of an acquired buffer in a division buffer pointer section in the original header. Batsuhua division method.