JP5085479B2 - Network system - Google Patents

Description

  The present invention relates to a network system compliant with ARCNET (Attached Resource Computer Network).

  In recent years, many of the old analog broadcasting facilities have been digitized, and this trend is certain to continue to develop. Against this background, facilities for controlling various broadcasting devices have also been digitized. This type of automatic program transmission control equipment is formed by connecting controlled devices such as a recorder that stores video materials and a switcher that switches a plurality of videos via a LAN (Local Area Network). For such applications, ARCNET, which has strength in real-time control, is generally used.

  The feature of ARCNET is that it is a network that employs a token passing system, and that configuration data reconstruction processing at the physical layer level is started as soon as the network topology changes. By providing such a function, it can be said that ARCNET has a relatively high degree of freedom in topology.

  On the other hand, if a node that continues to issue a reconstruction request signal (burst signal) occurs due to some failure, the node on the LAN cannot pass the token (usage right) to the next node (token pass). If such a situation occurs, the entire LAN goes down, and there is a risk that no communication will be possible.

Various documents disclosing countermeasures against network failures are known (see, for example, Patent Documents 1 and 2), but none of them indicate a response to such a system down that occurs in a form specific to ARCNET. .
JP 2002-319941 A JP-A-10-290243

As described above, in ARCNET, when a node that continues to output a burst signal appears, tokens cannot be transferred between the nodes because of network specifications. If such a situation occurs, the entire system may go down, and some countermeasure is required.
The present invention has been made under the circumstances described above, and an object of the present invention is to provide a network system with improved resistance to system down.

  In order to achieve the above object, according to an aspect of the present invention, in a network system that operates a network formed by a plurality of nodes using a token passing protocol (for example, ARCNET), a token circulating in the network is captured. Reconstructing request means for sending a burst signal to the network to request restructuring of the circulation route of the token when the state where the token has not been reached continues for a certain period of time. Reconstructing means for resetting the recirculation route after detecting the burst signal, and the monitoring node monitors the destination node identifier recorded in the captured token; and From the result, the burst signal is transmitted. Network system comprising: a specifying means for specifying a is provided.

  By taking such means, the monitoring node can grasp the destination node described in each token. In ARCNET, node IDs (identifiers) from # 00 to # 255 can be set as destinations. For example, these lists may be managed by setting a flag. Since the destination node ID of the token is incremented as the reconstruction progresses, monitoring the change over the specified period makes it possible to identify the possibility of the reconstruction and further identify the node that sent the burst signal. Become. Therefore, fault orientation can be performed promptly, and it becomes possible to increase resistance to system down.

  According to the present invention, it is possible to provide a network system with improved resistance to system down.

FIG. 1 is a diagram showing an embodiment of a network system according to the present invention. In this embodiment, a network conforming to ARCNET is assumed. The system of FIG. 1 is configured by connecting a plurality of nodes. The physical form of the connection may be any form such as a ring shape or mesh shape in addition to a bus shape as illustrated. Each node is uniquely assigned an identifier (ID) from # 00 to # 255 (FF). Among these, the node with ID = # 12 is set as the monitoring node 100.
Note that the node having the ID “# 00” does not exist in the ARCNET on the standard. In this embodiment, the ID “# 00” is used to detect a burst. A procedure for detecting a burst using a virtual node ID “# 00” will be described below.

  FIG. 2 is a schematic diagram showing a logical connection relationship of the network system of FIG. ARCNET is a token-passing protocol, and the logical connection form between nodes is a ring when viewed in the order in which the tokens circulate. If the network state is stable, the token is transmitted in order through the nodes of this ring. Individual tokens are captured by the monitoring node 100. The capture here is, of course, a process of reading the contents of the packet, not capturing the data and not flowing it over the network. That the token has been captured is hidden from other nodes.

  Assume that a node (# 02 in FIG. 2) that cannot receive a token occurs due to some kind of failure. If the token cannot be received for a certain period, that is, when 840 milliseconds have elapsed in ARCNET, the internal timer (network state monitoring timer) of node # 02 times out. Then, the node # 02 immediately sends a burst signal to the network and requests network reconfiguration (reconfiguration). In short, reconfiguration is the process of resetting the token's circulating route after resetting it. The node that has detected the burst signal immediately starts the reconstruction process.

  FIG. 3 is a functional block diagram showing an embodiment of the monitoring node 100. The monitoring node 100 includes an interface unit 10, a display unit 20, an input / output unit 30, a memory 40, a CPU (Central Processing Unit) 50, and an ARCNET chipset 60. Among these, the interface unit 10 is connected to the ARCNET, and mediates the exchange of packets. The display unit 20 and the input / output unit 30 provide a user interface environment via, for example, a GUI (Graphical User Interface).

  The CPU 50 includes a token capture unit 50a, a monitor processing unit 50b, a specific processing unit 50c, and firmware 50d that supports these functions at a level close to the physical layer as processing functions based on programs stored in the memory. The token capture unit 50a captures a token circulating in the network. The monitor processing unit 50 b monitors the destination node ID recorded in the captured token and writes the result in the management register 40 a of the memory 40. The identification processing unit 50c identifies the node that has transmitted the burst signal from the contents of the management register 40a.

  The ARCNET chipset 60 is an existing chipset mounted on a node device conforming to ARCNET, and includes a token processing unit 60a, a reconstruction processing unit 60b, an internal timer 60c, and a burst signal transmission unit 60d. The token processing unit 60a performs processing related to token transmission / reception. When the network status monitoring timer times out when the token has not been reached, the reconstruction processing unit 60b gives an instruction to the burst signal transmission unit 60d, and transmits the burst signal to the network to request network reconstruction. When a burst signal from another node is detected, the reconstruction processing unit 60b starts a network reconstruction process.

  The internal timer 60c is a timekeeping function provided in a node conforming to ARCNET. In addition to a network state monitoring timer that times out in 840 milliseconds, a network response monitoring timer (timeout in 78 microseconds) and an idle state monitoring timer (86 microseconds) And timeout). Next, the operation of the above configuration will be described.

  FIG. 4 is a diagram showing an example of the contents of the management register 40a when the network is stable. The management register 40a is a 256-bit register, and the ID of a node existing in the network is managed by setting a flag 1 to a bit corresponding to a destination node included in the captured token. The firmware 50d periodically samples the contents of the management register 40a, and records the data if it is in the same state twice in succession.

If the sampling data does not match the previous sampling data, the current time and the changed bit number are recorded. Once the flag 1 is set, the management register 40a maintains that state until a clear instruction is received from the firmware 50d. The firmware 50d samples the register and clears the management register 40a if the same state continues twice.
According to the above procedure, there is a possibility that flag 1 is set in all the bits from bit 0 to bit FF when a burst occurs. In such a case, not all changed bits may be recorded, but only that bit 0 has changed may be recorded.

  FIG. 5 is a diagram showing an example of the contents of the management register 40a immediately after the burst signal is transmitted. When a burst occurs from the state (a) of FIG. 4, a token whose ID is incremented as the reconstruction progresses is transmitted to the network, and each node ID of the management register 40a is shown in (b) to (e). Thus, the flag 1 is gradually raised. If the state (e) is reached, this state is stabilized, that is, the same state is detected twice. Comparing (a) and (e) shows that bit 0 has changed from 0 to 1. Therefore, the firmware 50d records only that bit 0 has changed, and then clears the management register 40a (f).

  Changing bit 0 from 0 to 1 means that a burst has occurred. That is, in the reconfiguration process immediately after the burst, “# 00” is written in the destination ID during the circulation of the token. Since the token does not turn to “# 00” in the normal token circulation, the occurrence of a burst can be detected by detecting the token addressed to the ID “# 00”.

  In addition, when the bit 0 is 0 and the bit X changes from 0 to 1, the firmware 50d records that the node X has joined the network. When bit 0 is 0 and bit X changes from 1 to 0, it is recorded that node X has left the network.

  As described above, in this embodiment, the monitoring node 100 that captures a token is provided, and the destination node ID recorded in the token is always monitored using the management register 40a. Then, the management register 40a is periodically sampled and the change is monitored, and the presence or absence of occurrence of a burst is detected from the monitoring result. Furthermore, it is possible to detect to which node the token has been rotated before the reconstruction occurs. Based on this, the node that transmitted the burst signal, that is, the node that has failed is identified.

  That is, the ID of the transmission source node is not written in the token used in ARCNET. Here is the reason why it is difficult to specify the fault location. In spite of this, according to this embodiment, it is possible to detect the departure of a node, and the system can recognize that a failure has occurred in that node or the node immediately preceding it.

  Therefore, even when the reconstruction process is continuing, the node that has caused the failure can be quickly isolated from the network by notifying the user of the detached node by a communication means other than ARCNET. As a result, it is possible to minimize the downtime of the entire system and to provide a network system with increased resistance to system down.

  In addition, this invention is not limited to the said embodiment, In an implementation stage, a component can be deform | transformed and embodied in the range which does not deviate from the summary. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment.

The figure which shows embodiment of the network system concerning this invention. The schematic diagram which shows the logical connection relation of the network system of FIG. 2 is a functional block diagram showing an embodiment of a monitoring node 100. FIG. The figure which shows an example of the content of the management register | resistor 40a in the state where the network is stable. The figure which shows an example of the content of the management register | resistor 40a immediately after sending out a burst signal.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100 ... Monitoring node, 10 ... Interface part, 20 ... Display part, 30 ... Input / output part, 40 ... Memory, 40a ... Management register, 50 ... CPU, 50a ... Token capture part, 50b ... Monitor processing part, 50c ... Specific process 50d ... firmware, 60 ... ARCNET chipset, 60a ... token processing unit, 60b ... reconstruction processing unit, 60c ... internal timer, 60d ... burst signal sending unit

Claims (3)

In a network system that operates a network formed by multiple nodes using a token-passing protocol,
Comprising a monitoring node for capturing tokens circulating in the network;
The plurality of nodes are:
Restructuring request means for sending a burst signal to the network to request restructuring of the token's circulation route when the state where the token has not been reached continues for a certain period of time;
Reconstructing means for reconstructing after resetting the circulation route upon detecting the burst signal,
The monitoring node is
Monitoring means for monitoring a destination node identifier recorded in the captured token;
A network system comprising: specifying means for specifying a node that has transmitted the burst signal from the monitoring result.   2. The network system according to claim 1, wherein when the burst signal is detected, the reconstructing unit transmits the token to the network while incrementing the destination node identifier.   The network system according to claim 1, wherein the protocol is an ARCNET (Attached Resource Computer Network).

Images