JP4051644B2 - Relay processing unit control method in network connection device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an inter-network connection device, and more particularly, to a control method for a relay processing unit (Routing Processor; simply referred to as RP) in an inter-network connection device.
[0002]
[Prior art]
Conventionally, in order to connect a plurality of networks to each other, an inter-network connection device composed of a plurality of RPs each connected to one or more communication ports for routing and a processor for controlling the plurality of RPs has been proposed. Has been. However, since the conventional RP control method in the network connection apparatus controls a plurality of RPs installed in the apparatus with a single RP control program, the waiting time of each RP increases as the number of RPs increases. This increases the processing time of the entire apparatus.
[0003]
Next, the reason will be specifically described with reference to the drawings. FIG. 2 shows a configuration of an inter-network connection device that realizes a conventional RP control method. In the figure, the bold line portion represents the hardware configuration, the other portion represents the software configuration, 200 is an inter-network connection device, 201 is a device management unit, 202 is an RP control program, and 210 to 212 are RP (0) to RP (2) are shown. Here, the RP control program 202 is executed as a part of the device management unit 201 and controls all of the RP (0) 210 to RP (2) 212. In this case, the RP control program 202 cannot perform the process for each RP at the same time, and needs to execute the processes for each RP one by one in time series.
[0004]
FIG. 3 is a flowchart for explaining the RP initialization / activation process at the time of device activation according to the conventional RP management method. In step 301, i is set to an initial value “0”, and in step 302, i is determined from the number of RPs installed. It is determined whether or not the number is larger. If the number is less than or equal to the number of mounted RPs, step 303 for performing initialization / activation processing and step 304 for adding 1 to i are performed. Initialization and activation processing are performed for all RPs mounted in steps 302 to 304. While one RP is executing the RP initialization / activation process of step 303, the initialization / activation process of the other RP is not performed and the initialization / activation process performed by the one RP is completed. have to wait.
[0005]
FIG. 4 is a diagram showing the relationship between the flow of the RP initialization / activation process shown in FIG. 3 and time. Reference numerals 410, 411, and 412 denote RP (0), RP (1), and RP (2), respectively. ) Shows the time during which the initialization / start-up process is executed. As can be seen, the execution start of the initialization / startup process 412 in RP (2) is the sum of [RP (0) initialization / startup process 410] + [RP (1) initialization / startup process 411]. We have to wait only for the time indicated by reference numeral 422, which is time.
Then, in order to finally complete the initialization / activation process of all RPs, the processing time indicated by reference numeral 400 is required. The RP initialization / start-up process has been described above. However, this situation can be applied to all processes, not just the initialization / start-up process, and the more time required to execute the process for one RP increases. Other RPs wait longer for processing. As a result, the processing time of the entire apparatus also increases.
[0006]
Further, in the conventional RP control method, when the RP control program 202 is upgraded, it is necessary to rewrite the entire device management unit 201. Therefore, each time the version is upgraded, the device is restarted and the processes of FIGS. 3 and 4 are performed. Must be maintainable. For the prior art, see, for example, JP-A-5-199230.
[0007]
[Problems to be solved by the invention]
Since the conventional RP control method for controlling all RPs with a single RP control program cannot execute processing for each RP in parallel, the processing time for one RP increases as the processing time for one RP increases. There is a problem that the waiting time increases, and as a result, the processing time of the entire apparatus also increases.
Further, since the RP control program is a part of the device management program, there is a problem that the entire device needs to be restarted at the time of rewriting, resulting in poor maintainability.
[0008]
An object of the present invention is to solve the above-mentioned problems and reduce the processing waiting time of each RP by executing the anti-RP processing that has been executed in time series in parallel, resulting in the processing time of the entire apparatus as a result. To provide a relay processing unit control method in an inter-network connection device capable of shortening the network, and when the version of the RP control unit program on the RP is upgraded, it can be handled only by stopping and restarting the RP. It is to provide a relay processing unit control method in an inter-network connection device that can improve performance.
[0009]
[Means for Solving the Problems]
In order to achieve the above object, a relay processing unit control method in an inter-network connection device according to the present invention includes (1) an RP management individual unit that gives a control instruction to each relay processing unit, and operates on each relay processing unit The relay processing unit is controlled by a two-level control program of the RP control unit that controls the relay processing unit, and one RP management individual unit is generated for each relay processing unit, and each RP management individual Each autonomously gives a control instruction to the corresponding RP control part, and each RP control part autonomously controls the relay processing part according to a control instruction from the corresponding RP management individual part, (2) The number of RP management individual units equal to the maximum mountable number of relay processing units specified in the inter-network connection device is generated in advance when the inter-network connection device is started, or the inter-network connection The number of RP management individual units generated dynamically changes according to the number of relay processing units installed in the device (the generation and deletion of RP management units are dynamically controlled according to the number of installed RPs) (3) The RP control unit program is stored in a pre-compressed form in a storage area on the relay processing unit, and is expanded on a temporary memory of the relay processing unit according to an instruction from the RP management individual unit. When the RP control unit program is activated and does not exist in the storage area on the relay processing unit, the RP control unit program is automatically taken in, or the RP control unit is stored in the temporary memory of the relay processing unit when the storage area fails It is characterized by directly writing and executing a program.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 shows a configuration of an inter-network connection apparatus 100 that realizes an RP control method according to the present invention. In the figure, the bold line portion represents the hardware configuration, and the other portions represent the software configuration. The RP management unit includes an RP management common unit 102 and RP management individual units 110 to 112, and the RP management common unit 102 is a part of the device management unit 101. The RP management individual units 110 to 112 manage RP (0) 120 to RP (2) 122, respectively.
[0011]
The RP (0) 120 to RP (2) 122 have RP control unit program storage areas 130 to 132, respectively. The RP control unit programs 150 to 152 are compressed in the RP control unit program storage areas 130 to 132. Are stored in the same format. When the RP (0) 120 to RP (2) 122 are activated, the RP control unit programs 150 to 152 are expanded on the temporary memories 140 to 142 from the RP control unit program storage areas 130 to 132, and are used as the RP control units 160 to 162. Let it run.
[0012]
FIG. 5 is a flowchart illustrating an example of generation processing of the RP management individual unit in the RP management common unit 102. The RP management common unit 102 initializes i = 1 in step 501. In the next step 502, it is determined whether or not i is larger than the number of mountable RPs. If it is not larger than the number of mountable RPs (step 502: N), an RP management individual unit (i) is generated ( Step 503), after adding 1 to i, the process returns to Step 502 again. The generated individual RP management unit can execute the initialization / activation processing of the corresponding RP independently and in parallel with the others.
If i is larger than the number of mountable RPs (step 502: Y), this means that all mountable RP management individual units have already been generated, and the processing is terminated (step 505).
[0013]
FIG. 6 shows the relationship between the RP management individual unit generation processing and the RP initialization / startup processing flow and time of FIG. 5, in which RP management common unit 102 generates each RP management individual unit. 610 to 612 indicate the time during which each RP is initialized / started up. The difference in the processing start times of the initialization / start-up processes 610 to 612 of the individual RP management individual units is small because it is only the difference in the time at which the individual RP management individual unit generation processing by the RP management common unit is completed. The initialization / start-up processes 610 to 612 of the management individual unit are executed in parallel in almost all time zones. In this way, the RP initialization / start-up processing that was executed in time series in the conventional method can be executed simultaneously and in parallel in this embodiment, and as a result, the processing time of the entire apparatus can be shortened. become.
[0014]
FIG. 5 is a flowchart in a case where the RP management common unit 102 generates the RP management individual units as many as the maximum number of RPs that can be mounted. However, FIG. 7 is an example, but FIG. When the unit 102 does not generate the maximum number of RP management individual units on which the RP can be mounted, but dynamically increases or decreases the number of RP management individual units according to the RP mounting state, that is, when a new RP is mounted 6 is a flowchart in which one RP management individual unit is generated and one RP management individual unit is deleted when the RP is removed.
[0015]
Specifically, it is determined in step 701 whether or not there is a change in the RP mounting state. If there is a change in the RP mounting state (step 701: N), then in step 702, the change is newly made. It is determined whether or not there is a change due to the mounting of the RP. If it is newly mounted (step 702: Y), one RP management individual unit is generated (step 704), and the process is terminated. If it is not newly installed (step 702: N), it is a case where the RP is removed, so one RP management individual unit is deleted (step 703), and the process is terminated.
[0016]
FIG. 8 is a flowchart showing an example of the automatic writing function of the RP control unit program on the RP.
This function is performed before execution of the RP initialization / startup process. First, it is checked whether or not there is an RP control unit program in the RP control unit program storage area on the RP (step 801). If there is no RP control unit program (step 801: N), an RP management individual unit Automatically writes the RP control unit program in the RP control unit program storage area (step 802). If there is an RP control unit program (step 801: Y), the RP control unit is activated to execute RP initialization / activation processing (step 803).
[0017]
FIG. 9 is a flowchart illustrating an example of an RP control unit activation function from the RP temporary memory when an RP control unit program storage area failure on the RP occurs.
This function is performed before execution of the RP initialization / startup process. First, a failure check of the RP control unit program storage area is performed to determine whether or not the area is normal (step 901). If the determination result is normal (step 901: Y), the RP control unit program is expanded from the storage area onto the temporary memory (step 903), and then the RP control unit is activated to perform RP initialization / activation processing. Perform (step 904).
On the other hand, if the result of the determination in step 901 is invalid (step 901: N), the RP management individual unit directly writes the RP control unit program on the temporary memory of the RP (step 902), and then it is normal. Then, the RP control unit is activated to perform RP initialization / activation processing (step 904).
[0018]
According to the embodiment described above, by generating the RP management individual unit corresponding to the RP, the processing of the RP can be executed simultaneously in parallel, and the processing waiting time of each RP can be reduced. Ultimately, it is possible to reduce the processing time of the entire apparatus. Further, it is not necessary to restart the apparatus when rewriting the RP control unit program, and it is possible to improve maintainability. Furthermore, the number of RP management individual units is generated by the number of installed RPs, and when the number of mounted RPs changes, the generation / deletion of RP management individual units and the RP Initialization and startup can be performed automatically.
[0019]
【The invention's effect】
According to the present invention, the processing waiting time of each RP can be shortened, and as a result, the processing time of the entire apparatus can be shortened. In addition, when upgrading the version of the RP control unit program on the RP, it is possible to cope with only the stop and restart of the RP, so that maintainability can be greatly improved.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of an inter-network connection device that implements an RP control method of the present invention.
FIG. 2 is a configuration diagram of an inter-network connection device that realizes a conventional RP control method.
FIG. 3 is a flowchart of RP initialization / startup processing by a conventional RP control method.
FIG. 4 is a diagram for explaining the relationship between time and the flow of RP initialization / startup processing according to a conventional RP control method;
FIG. 5 is a flowchart of RP initialization / startup processing according to the RP control method of the present invention.
FIG. 6 is a diagram for explaining the relationship between time and the flow of RP initialization / startup processing according to the RP control method of the present invention.
FIG. 7 is a flowchart of RP initialization / startup processing according to the RP control method of the present invention.
FIG. 8 is a flowchart of RP control unit program automatic writing processing;
FIG. 9 is a flowchart of RP initialization / activation processing when an RP control unit program storage area failure occurs;
[Explanation of symbols]
100, 200: Network connection device 101, 201: Device management unit 102: RP management common unit 110-112: RP management individual unit 120-122, 210-212: Relay processing device (RP (0) -RP (2) )
130-132: RP control unit program storage area 140-142: Temporary memory 150-152: RP control unit program 160-162: RP control unit

Claims (4)

A relay processing unit control method in an inter-network connection device capable of mounting a plurality of relay processing units (hereinafter referred to as RP),
The relay processing unit is controlled by a two-level control program of an RP management individual unit that gives control instructions to each relay processing unit and an RP control unit that operates on each relay processing unit and controls the relay processing unit. , One RP management individual unit is generated for one relay processing unit, each RP management individual unit autonomously gives a control instruction to a corresponding RP control unit, and each RP control unit A relay processing unit control method in an inter-network connection device, wherein the relay processing unit is autonomously controlled by a control instruction from a corresponding RP management individual unit. In the network connecting unit of claim 1, wherein the advance generate RP managing individual parts of the same number as the relay processing unit maximum mountable number that is specified between the network connection device when the connecting device starts between the network Relay processor control method. 2. The relay processing unit control in the inter-network connection device according to claim 1, wherein the number of generated RP management individual units is dynamically changed according to the number of relay processing units mounted on the inter-network connection device. Method.   The RP control unit program is stored in a pre-compressed form in a storage area on the relay processing unit, and is expanded and activated on a temporary memory of the relay processing unit according to an instruction from the RP management individual unit. When the RP control unit program does not exist in the storage area, the RP control unit program is automatically taken in, or when the storage area fails, the RP control unit program is directly written in the temporary memory of the relay processing unit and executed. The relay processing unit control method for an inter-network connection device according to claim 1.

Images