JPH0243662A - Communication path control system - Google Patents

Abstract

PURPOSE:To improve the throughput of the whole of a network by repeating the informing and updating of information to be informed between all the processors until the title system is made into a state in which no information exists and managing the state of a virtual path with a self processor as either a start edge or a relay point in the respective processors. CONSTITUTION:For example, when the establishment of the paths between a self processor #0 and processors #1-#3 adjacent to the self processor #0 is detected, the self processor #0 checks whether virtual path information 2-11 managed in the self processor #0 is to be informed to the adjacent processors #1-#3 or not. Further, when the virtual path information 2-11 is to be informed, the self processor #0 informs an applicable processor out of the processors #1-#3 of the virtual path information 2-11, and when virtual path information 2-13 is informed from the other adjacent processors #1-#3, or when the self processor #0 detects the disconnection of the paths between the self processor #0 and the adjacent processors #1-#3, and self processor #0 changes the virtual path information 2-11 managed in the self processor #0, checks whether the virtual path information 2-11 managed in the self processor #0 is to be informed to the other adjacent processors #1-#3 or not, and informs the applicable processor out of the processors #1-#3 the virtual path information 2-11 when the virtual path information 2-11 is to be informed. Thus, the throughput of the whole of the network can be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a communication route control method, and particularly to a communication route control method for a connectionless network system having a virtual route via a relay processor.

[Conventional technology]

Conventionally, when building a network system between multiple processors, PVC was used to secure a fixed route in the system.
<Permanent virtual circuit: fixed connection)
A VC (virtual call) method is used in which one route available at that time is established by call control when communication becomes necessary.

[Problem to be solved by the invention]

When using the PVC method, the route is fixed in one communication path, so detour control cannot be performed in the event of a failure, etc. When multiple communication paths are set up with separate routes between the same start and end points. However, in a large-scale network system, the adoption of such a method has the problem that there may be limitations due to memory resources and the like.

In addition, when the VC method is adopted, if at least one connectable route exists at the time of connection establishment, it is possible to automatically select that route, but once the connection is established, that route becomes an obstacle. In this case, automatic detour cannot be performed and the connection must be released. Furthermore, overhead for data transmission and reception for connection establishment always occurs, which affects the throughput of the entire network. Also, it is difficult to provide certain prerequisites for setting the route when establishing a connection, and if there is a route that can be connected in any case, the connection will be established and that route will remain fixed until the connection is released. For example, if a route is taken that uses many processors as relay points, there is a problem in that the response time is greatly affected.

[Means to solve the problem]

According to the present invention, a connectionless network is created in which adjacent processors are connected using a permanent virtual circuit method in a network made up of multiple processors, and virtual routes with all directions that can fit between each processor are set on the network. When the system detects the establishment of a route between its own processor and an adjacent processor.If the system checks whether to notify other adjacent processors of the virtual route information managed within its own processor, and notifies it.Applicable. Notifies the processor of virtual route information,
When virtual route information is notified from another adjacent processor,
Alternatively, if a disconnection of the route between the own processor and an adjacent processor is detected, the virtual route information managed within the own processor is changed, and the virtual route information managed within the own processor is transmitted to other adjacent processors. A communication path control method is obtained which is characterized in that it checks whether notification is to be made and, if notification is to be made, the corresponding processor is notified of virtual path information.

〔Example〕

Next, the present invention will be described in more detail with reference to the drawings showing one embodiment of the present invention.

Referring to FIG. 1, one embodiment of the present invention includes host computers H1, H2 and network processors NPI, NP2, NF2, . NF2 and
Distributed processor DPI, DP connected to the network
2, and terminals Tl, T2.2 connected to the network. T3
．． It is composed of T4. host computer H1
and network processors NPI, NP2 are installed in the A center, and a host computer H2 and network processors NP3. NP4 is installed in the B center.

FIG. 2 shows a specific implementation example of the communication route control method when the network according to the present invention is constructed using the network processor in FIG. 1. In FIG. 2, 2-40.
2-20.2-30.2-40 each represent a processor that constructs a network, and 2-11 represents a processor #
2-12 is the virtual route information managed by 0, 2-12 is the virtual route information with processor #0 as the terminal, and 2-13 is the virtual route with processor #0 as the starting end or relay point notified from other processors. 2-14 represents a mask table for extracting virtual route information to be notified from processor #0 to other adjacent processors based on certain preconditions, and virtual route information currently being notified to other adjacent processors. ing. Here, as a prerequisite,
Between processors 0 and processor #1. Processor #2 and processor #3 each belong to the same district, and do not have a virtual route that uses a processor in another district as a relay point as a virtual route between processors in the same district. The two conditions are that there is no virtual route that can be used. Further, 2-11 has a value that is the logical sum of the information of 2-12 and 2-13. Each processor is
2-11 in a state where all routes are disconnected
It has only mask tables 2-12 and 2-14 as information.

FIG. 3 shows identification numbers of virtual paths that have bidirectionality between each processor.

The operation of this embodiment will be explained with reference to FIG. If processor #3 goes down while all inter-processor paths are connected, processor #0. Processor #1. Each processor #2 detects the disconnection of the route with processor #3, changes all the virtual route information notified by processor #3 to a disconnected state, and updates the virtual route information managed by its own processor. Update. Next, virtual route information for notification to each adjacent processor is extracted from the updated virtual route information using a mask table. Only when the extracted virtual route information differs from the previously notified information, the processor is notified of the extracted virtual route information.0 In the processor that received the notification,
The notification information area (2-13) from the adjacent processor is changed based on the notified virtual route information, and the information managed by the own processor is updated. In this way, each processor manages the state of the virtual route with its own processor as the start end or relay point by repeating notification→update until there is no more information to be notified among all the processors.

When one processor (A) wants to send a message addressed to another processor (B), check the state of the virtual route with A as the starting end and B as the ending end using 2-11. -13, the transmission path is controlled by transmitting a message to the processor with the highest priority among those whose virtual path is in the connected state.

〔Effect of the invention〕

As described above, according to the present invention, route detour control is possible after grasping the state of the route at the terminal end or relay point at the starting end, without being limited by memory resources or the like. In addition, by creating a mask table, it is possible to add certain preconditions to possible routes within the network, so it is possible to select routes that use a large number of processors as relay points, which can affect response time. It is also possible to control the situation so that it does not occur. Furthermore, since it is a connectionless method, it is extremely unlikely that communication will be interrupted due to a path failure during communication, and since there is no overhead for establishing a connection, the throughput of the entire network can be improved.

[Brief explanation of the drawing]

FIG. 1 is a system configuration diagram of an embodiment of the present invention, FIG. 2 is a diagram showing a concrete implementation example of the communication path control method using the configuration shown in FIG. 1, and FIG. 3 is a diagram showing the direction between each processor. FIG. 4 is a flowchart showing the process outline of the present invention. Hl, H2...Host system, NPI~NP4...Network processor, DPI
, DP2...Distributed processor, T1-T4...Terminal, 2-10.2-20.2-30.2-40...Network processor configuring the network, 2-11.
...Virtual route information table managed by processor #0,
2-12... Virtual route information table with processor #0 as the terminal, 2-13... Processor #1. Processor #2. Processor #0 notified by processor #3
A virtual route information table with 2- as the starting point or relay point.
14... Processor #1. Processor #2. A mask table for extracting virtual route information to be notified to processor #3, and a virtual route information table to be notified to each processor at the current time. Agent Patent Attorney Susumu Uchihara Figure 1 Figure 2 Puzzle #2 Figure 4 ZU Jin, 7f is 3

Claims (1)

[Claims] A connectionless network system in which adjacent processors are connected using a permanent virtual circuit method in a network of multiple processors, and virtual routes with all possible directions are set between each processor. When the establishment of a route between the own processor and an adjacent processor is detected, it is checked whether to notify other adjacent processors of the virtual route information managed within the own processor, and if notification is to be made, the corresponding processor When the virtual route information is notified from other neighboring processors,
Alternatively, if a disconnection of the route between the own processor and an adjacent processor is detected, the virtual route information managed within the own processor is changed, and the virtual route information managed within the own processor is changed to other adjacent processors. A communication route control method characterized by checking whether notification is to be made and, if notification is to be made, notifying the corresponding processor of virtual route information.