JP2005311539A - Bridge device, address management method and network system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform time-out processing of an unused address table from among address tables registered in a bridge device for relaying Ethernet and IEEE1394. <P>SOLUTION: Every time an address table AT is newly registered in which address information in a transmission format of data in the Ethernet and address information in a transmission format of data in the IEEE1394 and registered time data are set, the address table AT is disposed at the head of a linear table in which the registered address tables AT are aligned in a line. Every time an address table AT in the linear table is used, the address table AT is moved to the head of the linear table, the registration time data of the address table AT is updated into the present time. It is sequentially checked from the tail to the head of the linear table whether or not the address table AT is not used over a predetermined time using the registration time data, and the checking is finished if the address table AT used over the predetermined time is found. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a bridge device that relays between Ethernet and IEEE1394, an address management method in the bridge device, and a network system in which Ethernet terminals are connected to each other via the bridge device. .

  One of the high-speed serial bus standards is IEEE1394. IEEE 1394 can connect 63 devices (nodes) to one bus, supports plug-and-play and hot plug, and can be connected in daisy chain or tree form. Is possible. In addition, real-time transmission of moving image parameters and the like is possible by the isochronous transmission method.

  Since IEEE 1394 enables communication between nodes without using a host computer, it can be used to connect digital home appliances such as digital video cameras, digital video decks, and digital television receivers at home. Forms are possible.

  However, in recent years, IEEE 1394b has been announced as a specification of IEEE 1394 (IEEE 1394b) that realizes a transmission speed of a maximum of 400 MB / s and a communication distance of several hundred meters using an optical fiber. From such a high transmission speed and long communication distance, IEEE 1394 can be expected to be used in such a way that computers are connected to each other in a home or a company to construct a LAN (local area network).

  However, since the currently mainstream LAN uses Ethernet, most existing computers have an Ethernet interface but do not have an IEEE 1394 interface. In order to connect terminals having Ethernet interfaces (also referred to as “Ethernet terminals”) such as existing computers using IEEE 1394, it is necessary to use a bridge device that relays between Ethernet and IEEE 1394. I need it.

  FIG. 20 shows a configuration example of a LAN using such a bridge. The Ethernet terminal 51 and the Ethernet terminal 53 are connected to the bridge 55a by Ethernet. The Ethernet terminals 51 and 53 and the bridge 55a are arranged on the first floor of a certain building, for example.

  The Ethernet terminal 52 and the Ethernet terminal 54 are connected to the bridge 55b by Ethernet. The Ethernet terminals 52 and 54 and the bridge 55b are arranged on the second floor of the same building, for example.

  The bridge 55a and the bridge 55b are connected by an IEEE 1394 bus (IEEE 1394b specification).

  One of the roles of the bridges 55a and 55b is that a MAC (Media Access Control) frame or IP packet, which is a data transmission format in Ethernet, and an isochronous packet or asynchronous packet, which is a data transmission format in IEEE1394. Format conversion between them.

  Another role of the bridges 55a and 55b is the MAC address and IP address which are address information in the data transmission format in Ethernet, and the node number and offset address which are address information in the data transmission format in IEEE1394. Address translation between them.

Conventionally, as a method of address conversion in a bridge that relays between Ethernet and IEEE 1394 like the bridges 55a and 55b of FIG. 20, address information in the data transmission format in Ethernet and address information in the data transmission format in IEEE 1394 are used. There has been proposed a table in which address conversion is performed with reference to this table (see, for example, Patent Document 1).
JP 2001-268108 A (paragraph numbers 0007 to 0009, FIG. 1)

  By the way, this Patent Document 1 does not describe a specific structure of this table or a specific address management method using this table. Therefore, taking the LAN shown in FIG. 20 as an example, the structure of this table that can be generally considered is as shown in FIG.

  Address tables Ta51 to Ta54 in FIG. 21A are tables registered in the bridge 55a in FIG. Also, the address tables Tb51 to Tb54 in FIG. 21B are tables registered in the bridge 55b in FIG.

  The structure of these address tables will be described by taking the address tables Ta51 to Ta54 on the bridge 55a side as an example. The top parameter of the address table is the MAC address of the Ethernet terminal connected to the bridge 55a via Ethernet. “EtherMAC1” in the tables Ta51 and Ta52 represents the MAC address “MAC1” of the Ethernet terminal 51 in FIG. 20, and “EtherMAC3” in the tables Ta53 and Ta54 represents the MAC address “MAC3” in the Ethernet terminal 53 in FIG. Yes.

  The third parameter from the top of the address table is the IP address of the Ethernet terminal represented by the top parameter. “EtherIP1” in the tables Ta51 and Ta52 represents the IP address “IP1” of the Ethernet terminal 51 in FIG. 20, and “EtherIP3” in the tables Ta53 and Ta54 represents the IP address “IP3” in the Ethernet terminal 53 in FIG. Yes.

  The second parameter from the bottom of the address table is the MAC address of the Ethernet terminal that is the destination of the MAC frame from the Ethernet terminal represented by the parameter at the top. “Ieee1394MAC2” in the tables Ta51 and Ta53 represents the MAC address “MAC2” of the Ethernet terminal 52 in FIG. 20, and “Ieee1394MAC4” in the tables Ta52 and Ta54 represents the MAC address “MAC4” in the Ethernet terminal 54 in FIG. Yes.

  The fourth parameter from the top of the address table is the IP address of the Ethernet terminal represented by the second parameter from the bottom. “Ieee1394IP2” in the tables Ta51 and Ta53 represents the IP address “IP2” of the Ethernet terminal 52 in FIG. 20, and “Ieee1394IP4” in the tables Ta52 and Ta54 represents the IP address “IP4” in the Ethernet terminal 54 in FIG. Yes.

  The parameter “Ieee1394Node1” in the fourth row from the bottom of the address table represents the node number “Node1” of the bridge 55b in FIG. 20 connected to the bridge 55a via the IEEE1394 bus. (Conversely, the parameter “Ieee1394Node0” in the fourth row from the bottom of the address table Tb51 to Tb54 on the bridge 55b side represents the node number “Node0” of the bridge 55a.)

  The parameters in the third row from the bottom of the address table are the MAC frame with the Ethernet terminal represented by the top MAC address as the transmission source and the Ethernet terminal represented by the second MAC address from the bottom as the transmission destination. When an IP packet with the Ethernet terminal represented by the IP address of the third stage from the top as the transmission source and the destination of the Ethernet terminal represented by the IP address of the fourth stage from the top is sent to the bridge 55a The offset address for transferring the data of the MAC frame or IP packet to the bridge 55b via the IEEE 1394 bus (the offset value in the 64-bit address space by IEEE 1212 to which IEEE 1394 complies). For example, “Ieee1394 proxy offset 21” in the table Ta51 represents the offset address “proxy offset21” when the Ethernet terminal 51 and the Ethernet terminal 52 in FIG. 20 are the transmission source and the transmission destination, respectively.

  The parameter in the second row from the top of the address table is opposite to the parameter in the third row from the bottom, and the Ethernet terminal represented by the uppermost MAC address or the IP address of the third row from the top is the destination and When the Ethernet terminal represented by the second-stage MAC address from the top and the fourth-stage IP address from the top is the transmission source, the offset address is associated with the transmission destination and the transmission source. For example, “Ether proxy offset 11” in the table Ta51 represents an offset address “proxy offset 11” when the Ethernet terminal 52 and the Ethernet terminal 51 in FIG. 20 are the transmission destination and the transmission source, respectively. In the table where the destination MAC address or IP address and the source MAC address IP address are interchanged with each other as in the table Ta51 and the table Tb51, the offset address in the second stage from the top and the offset address in the second stage from the bottom The offset addresses are interchanged with each other.

  The parameter at the bottom of the address table is a registration time parameter indicating the time when the address table is registered (or the last used time when the address table is used after registration).

  Of these parameters in the address table, the IP address is an address used in the TCP / IP network layer, and is a value specific to each Ethernet terminal 51 to 54. The MAC address is an address used in the data link layer of TCP / IP and is acquired by ARP (address resolution protocol).

  The node number is automatically assigned to each of the bridges 55a and 55b when the bridge 55a and the bridge 55b are connected by the IEEE1394 bus. The offset address is determined by the bridges 55a and 55b as a unique value for each address table registered in the bridges 55a and 55b. The registration time data is created by the bridges 55a and 55b when registering or using each address table.

  In this way, 2 × 2 = 4 address tables correspond to the bridges 55a and 55b, corresponding to the two Ethernet terminals 51 and 3 and the Ethernet terminals 52 and 4 being connected to the bridges 55a and 55b, respectively. Are registered respectively. For example, when 100 Ethernet terminals are connected to each of the bridges 55a and 55b, 100 × 100 = 10000 address tables are registered in the bridges 55a and 55b, respectively.

  FIG. 22 is a diagram showing a state of MAC frame transfer in the bridges 55a and 55b, taking as an example the case of transferring a MAC frame from the Ethernet terminal 51 to the Ethernet terminal 52 in FIG.

  When the MAC frame is transferred from the Ethernet terminal 51 to the Ethernet terminal 52, the destination MAC address and the source MAC address in the MAC frame are “MAC2” and “MAC1”, respectively. When this MAC frame is sent from the Ethernet terminal 51 via Ethernet, the bridge 55a uses “MAC2” and “MAC1” as search keys, respectively, from the address tables Ta51 to Ta54 in FIG. The address table in which the parameter at the stage is “MAC2” and the parameter at the top is “MAC1” is searched.

  In this case, the address table Ta51 hits the search. Therefore, the bridge 55a uses the address table Ta51 to change the destination and source MAC addresses “MAC2” and “MAC1”, the node number “Node1” (parameter in the fourth row from the bottom), and the offset address “ The address is converted to “proxy offset 21” (the third parameter from the bottom).

  The bridge 55a converts the address-converted MAC frame into an isochronous packet or an asynchronous packet, which is a data transmission format in IEEE 1394, and transfers the MAC frame to the bridge 55b via the IEEE 1394 bus.

  When the packet is transferred from the bridge 55a, the bridge 55b uses the source node number “Node 0” and the offset address “proxy offset 21” in the packet as search keys to search the address tables Tb51 to Tb54 of FIG. The address table in which the node number in the fourth row from the bottom is “Node 0” and the offset address in the second row from the top is “proxy proxy 21” is searched.

  In this case, the address table Tb51 hits the search. Therefore, the bridge 55b uses this address table Tb51 to set “Node1”, “proxy offset21”, “MAC2” (the top parameter) and “MAC1” (the second parameter from the bottom). ) Is converted to a MAC address having a transmission source.

  The bridge 55b converts the address-converted packet into a MAC frame and transfers the packet to the Ethernet terminal 52 via Ethernet. Thereby, the transfer of the MAC frame from the Ethernet terminal 51 to the Ethernet terminal 52 is completed.

  In FIG. 22, the case of transferring a MAC frame is taken as an example, but when transferring an IP packet from the Ethernet terminal 51 to the Ethernet terminal 52 in FIG. 20 (for example, a table is already created before the MAC address is acquired by ARP). For example, the bridge 55a uses the destination and source IP addresses “IP2” and “IP1” as search keys from the address tables Ta51 to Ta54 in FIG. The address table in which the parameter in the fourth row from the top is “IP2” and the parameter in the third row from the top is “IP1” is searched. Further, the bridge 55b uses the address table Tb51 that has been found in the search, sets “Node1”, “proxy offset21”, “IP2” (the third parameter from the top) as the transmission destination, and “IP1” (upper To the IPC address with the fourth parameter) as the transmission source.

  The registration time data at the bottom of the address table in FIG. 21 is data for performing a timeout process for excluding an unused address table from registration.

  For example, when 100 Ethernet terminals are connected to each of the bridges 55a and 55b, the number of address tables registered in the bridges 55a and 55b is 100 × 100 = 10000, respectively. If the number of registrations increases, the number of address tables to be searched increases as it is. However, the address table actually used (address table corresponding to the Ethernet terminal to which the MAC frame or IP packet is actually transferred) is often a part of the registered address table. Therefore, it is periodically checked whether or not each registered address table has been used for a certain period of time (for example, 2 to 3 minutes), using registration time data. The number of address tables to be searched is suppressed by performing a timeout process.

  As a method of timeout processing, generally, as shown in FIG. 23, a small number (for example, about 10 to 20) of address tables used recently are temporarily stored as a cache, and all the remaining addresses are stored. A method of checking the unused time of the table can be considered.

  However, in this method, since it is necessary to check almost all registered address tables, the processing load increases and the processing time becomes long.

  As another method of timeout processing, for example, what kind of registration time data is currently set in each address table is managed, and unused time is checked by referring to the registered registration time data. A method is also conceivable.

  However, in this method, it is necessary to provide a new table for storing the registration time data currently set in each address table, and to perform processing for associating the table with the address table. Becomes larger and the process becomes complicated.

  In view of the above points, the present invention, when constructing a network system in which Ethernet and IEEE 1394 are relayed by a bridge device, such as the LAN shown in FIG. 20, is unused among the address tables registered in the bridge device. It is an object of the present invention to make it possible to perform the time-out process for excluding the address table from registration in a short time without increasing the processing load / complexity and increasing the circuit scale.

  In order to solve this problem, the present invention performs format conversion between a data transmission format in Ethernet and a data transmission format in IEEE 1394, and addresses information in the data transmission format in Ethernet and IEEE 1394. In a bridge device that performs address conversion between address information in the data transmission format, address information in the data transmission format in Ethernet, address information in the data transmission format in IEEE 1394, and registration time data are parameters. The first processing means for registering the address table set as a new address table, and each time an address table is newly registered, the address table is a linear Each time the second processing means arranged at the head of the linear table to be arranged and the address table already arranged in the linear table are used, the address table is moved to the head of the linear table and the address is moved. The third processing means for updating the registered time data of the table to the current time and the address table located at the end of this linear table are used for a certain period of time toward the address table located at the beginning of this linear table. And fourth processing means for sequentially checking whether or not the address table is used, and ending the check when the address table used within the predetermined time is found.

  In this bridge device, each time an address table in which address information in the data transmission format in Ethernet, address information in the data transmission format in IEEE 1394, and registration time data are set as parameters is registered. However, the registered address table is arranged at the head of the linear table arranged in a straight line.

  Each time an address table already arranged in the linear table is used, the address table is moved to the head of the linear table and the registered time data of the address table is updated to the current time. .

  As a result, the newly registered address table and the most recently used address table (address table with new registration time data) are arranged closer to the head of this linear table. Conversely, an address table that has passed since registration or an unused address table (an address table with old registration time data) is arranged at a position closer to the end of this linear table.

  Then, from the address table located at the end of this linear table to the address table located at the beginning of this linear table (that is, from the address table where the registration time data is old to the new address table) Then, it is sequentially checked using the registration time data whether it has not been used for a certain period of time, and the check is terminated when an address table used within a certain period of time is found.

  In this way, all registered address tables are arranged linearly in the order of registration time data, and by checking the unused time in order from the address table with the oldest registered time data, it is possible to Since it becomes unnecessary to perform the check, the processing load is reduced and the processing time is shortened.

  In addition, it does not manage what registration time data is currently set in each address table, but it simply arranges registered address tables in a straight line, resulting in increased circuit scale and complicated processing. Nor.

  In this bridge device, as an example, a border table that creates a linear table in which a predetermined number of blank tables with no parameters set are arranged in a straight line, and indicates the boundary between the blank table and the registered address table Processing means for setting a pointer at the head of the linear table is further provided. The first processing means sets a parameter in a blank table at the end of the linear table, and the second processing means newly adds an address table. Is registered at the head of the linear table and the border pointer is moved down by one table, and the fourth processing means uses the border pointer of the linear table. From the address table located above Towards the address table located on the head, it is preferable to check in the order using the registration time data or not been used more than a predetermined time.

  In this way, the blank table used for registering the address table and the registered address table are arranged in the same linear table, thereby managing the blank table and reusing an unused address table as a blank table. Can be easily performed.

  Next, the present invention performs format conversion between the data transmission format of Ethernet and the data transmission format of IEEE 1394, and addresses information in the data transmission format of Ethernet and the data transmission format of IEEE 1394. In the address management method in the bridge device that performs address conversion with the address information of the address, the address information in the data transmission format of Ethernet, the address information in the data transmission format of IEEE 1394, and the registration time data are used as parameters. Register the set address table, and each time a new address table is registered, the registered address table is arranged in a straight line. Each time an address table already placed in this linear table is used, the address table is moved to the beginning of this linear table, and the registered time data of that address table is Update to the time, from the address table located at the end of this linear table, to the address table located at the beginning of this linear table, in order to check whether it has been used more than a certain time, using the registered time data, The check is terminated when the address table used within the predetermined time is found.

  In this address management method, each time an address table in which address information in the data transmission format in Ethernet, address information in the data transmission format in IEEE 1394, and registration time data are set as parameters, the address table is registered. Are arranged at the head of a linear table in which registered address tables are arranged in a straight line.

  Each time an address table already arranged in the linear table is used, the address table is moved to the head of the linear table, and the registration time data of the address table is updated to the current time.

  As a result, the newly registered address table and the most recently used address table are arranged closer to the head of the linear table. Conversely, an address table that has passed since registration and an address table that is not used are arranged closer to the end of the linear table.

  Then, from the address table located at the end of this linear table to the address table located at the beginning of this linear table, it is checked in order using the registered time data whether it has been used for a certain period of time. The check is finished when the address table used for is found.

  In this way, all registered address tables are arranged linearly in the order of registration time data, and by checking the unused time in order from the address table with the oldest registered time data, it is possible to Since it becomes unnecessary to perform the check, the processing load is reduced and the processing time is shortened.

  In addition, it does not manage what registration time data is currently set in each address table, but it simply arranges registered address tables in a straight line, resulting in increased circuit scale and complicated processing. Nor.

  Next, the present invention provides a network system in which a plurality of bridge devices that relay Ethernet and IEEE 1394 are connected to an IEEE 1394 bus, and terminals having Ethernet interfaces are connected to each other via these bridge devices. Each bridge device has a first processing means for registering an address table in which address information in the data transmission format in Ethernet, address information in the data transmission format in IEEE 1394, and registration time data are set as parameters. Each time an address table is newly registered, the address table is already arranged in this linear table, with the second processing means arranged at the head of the linear table in which the registered address table is arranged in a straight line. Each time the address table is used, the address table is moved to the top of the linear table, and the third processing means for updating the registered time data of the address table to the current time, and at the end of the linear table From the address table that is located to the address table that is located at the beginning of this linear table, it is checked in order using the registration time data whether it has been used for a certain period of time, and the address table that is used within this certain time And a fourth processing means for ending the check when it is found.

  In this network system, each bridge device registers an address table in which address information in the data transmission format in Ethernet, address information in the data transmission format in IEEE 1394, and registration time data are set as parameters. In addition, the address table is arranged at the head of a linear table in which registered address tables are arranged in a straight line.

  Each time an address table already arranged in the linear table is used, the address table is moved to the head of the linear table and the registered time data of the address table is updated to the current time. .

  As a result, the newly registered address table and the most recently used address table are arranged closer to the head of the linear table. Conversely, an address table that has passed since registration and an address table that is not used are arranged closer to the end of the linear table.

  Then, from the address table located at the end of the linear table to the address table located at the beginning of the linear table, it is checked in order using the registered time data whether it has been used for a certain period of time. The check is finished when the used address table is found.

  In this way, all registered address tables are arranged linearly in the order of registration time data, and by checking the unused time in order from the address table with the oldest registered time data, it is possible to Since it becomes unnecessary to perform the check, the processing load is reduced and the processing time is shortened.

  In addition, it does not manage what registration time data is currently set in each address table, but it simply arranges registered address tables in a straight line, resulting in increased circuit scale and complicated processing. Nor.

  According to the present invention, in a network system in which Ethernet terminals are connected to each other via a bridge device that relays Ethernet and IEEE 1394, an unused address table among address tables registered in the bridge device is excluded from registration. Thus, the time-out process can be performed in a short time without increasing the processing load / complication and increasing the circuit scale.

  Also, when preparing a blank table to be used as an address table, by arranging this blank table and the registered address table in the same linear table, it is possible to manage the blank table and register it because it has been unused for a certain period of time. There is also an effect that the process of reusing the excluded address table as a blank table can be easily performed.

  Hereinafter, the present invention will be specifically described with reference to the drawings. FIG. 1 shows the overall configuration of a LAN to which the present invention is applied. A plurality of Ethernet terminals 2 (1), 2 (3),... Are connected to the bridge 1a by Ethernet. These bridges 1a and Ethernet terminals 2 (1), 2 (3),... Are arranged, for example, on the first floor of a building.

  Further, a plurality of Ethernet terminals 2 (2), 2 (4),... Are connected to the bridge 1b by Ethernet. These bridges 1b and Ethernet terminals 2 (2), 2 (4),... Are arranged on the second floor of the same building, for example. The Ethernet terminals 2 (1), 2 (2), 2 (3), 2 (4),... Are, for example, personal computers or remote routers (WAN routers).

  The bridge 1a and the bridge 1b are connected by an IEEE 1394 bus (IEEE 1394b specification). Further, for example, a plurality of digital AV devices 3 (1), 3 (2),... Such as a digital video camera, a digital video deck, and a digital television receiver are connected to the IEEE 1394 bus in a daisy chain.

  The bridges 1a and 1b are devices that relay Ethernet and IEEE 1394 and have the same configuration. FIG. 2 is a block diagram showing the configuration of these bridges. The bridges 1 a and 1 b include a controller 11, an address management table memory 12, a Node-GUID table memory 13, a blank table memory 14, a hash table memory 15, an Ethernet interface 16, an IEEE 1394 interface 17, and a control logic 18.

  The Ethernet interface 16 is an interface for performing communication via Ethernet. The IEEE 1394 interface 17 is an interface for performing communication via the IEEE 1394 bus. The control logic 18 converts the format between a MAC (Media Access Control) frame or IP packet, which is a data transmission format in Ethernet, and an isochronous packet or asynchronous packet, which is a data transmission format in IEEE 1394. It is a circuit which performs.

  The controller 11 is a circuit that controls the control logic 18 and performs address management using the address management table memory 12, the Node-GUID table memory 13, the blank table memory 14, and the hash table memory 15.

  In the following, address management by the controller 11 will be described by dividing it into processing relating to node numbers, processing relating to offset addresses, time-out processing of address tables, and processing relating to search of address tables.

[Node number processing]
First, processing related to the node number will be described. The Node-GUID table memory 13 in FIG. 2 includes node numbers assigned to the bridges 1a and 1b and the digital AV devices 3 (1), 3 (2),..., And the bridges 1a and 1b and the digital AV. This is a memory for storing one Node-GUID table in which the GUIDs (Global Unigue IDs) of the devices 3 (1), 3 (2),. The GUID indicates a 64-bit area that is a combination of vender_ID and chip_ID on the Configuration ROM in the 64-bit address space according to IEEE1212 conforming to IEEE1394, and is a value unique to each node connected to the IEEE1394 bus. It is.

  FIG. 3 shows an example of the Node-GUID table. The GUIDs of the bridges 1a and 1b are “1245678” and “56789012”, respectively, and the GUIDs of the digital AV devices 3 (1), 3 (2),... Are “12345123”, “9805678”,. In this example, the node numbers “Node 0” and “Node 1” are allocated to the bridges 1a and 1b, respectively, and the node numbers “Node 3” and “Node 4” are assigned to the digital AV devices 3 (1), 3 (2),. ... Are allocated, and these node numbers are set in correspondence with the GUIDs.

  This Node-GUID table is created when a node number is first assigned to the bridges 1a and 1b (when the bridges 1a and 1b are first connected to the IEEE 1394 bus).

  The address management table memory 12 in FIG. 2 is a memory for registering an address table (stores parameters of the address table) and stores pointers for each address table described later. FIG. 4 shows a case where two Ethernet terminals 2 (1) and 2 (3) and two Ethernet terminals 2 (2) and 2 (4) are connected to the bridges 1a and 1b in FIG. The address table registered in the address management table memory 12 is shown.

  Among these, the address tables Ta1 to Ta4 in FIG. 4A are tables registered in the bridge 1a in FIG. Also, the address tables Tb1 to Tb4 in FIG. 4B are tables registered in the bridge 1b in FIG.

  The structure of these address tables will be described by taking the address tables Ta1 to Ta4 on the bridge 1a side as an example. The uppermost parameter of the address table is the MAC address of the Ethernet terminal connected to the bridge 1a via Ethernet. “EtherMAC1” in the tables Ta1 and Ta2 represents the MAC address “MAC1” of the Ethernet terminal 2 (1) in FIG. 1, and “EtherMAC3” in the tables Ta3 and Ta4 is the MAC address of the Ethernet terminal 2 (3) in FIG. “MAC3” is represented.

  The third parameter from the top of the address table is the IP address of the Ethernet terminal represented by the top parameter. “EtherIP1” in the tables Ta1 and Ta2 represents the IP address “IP1” of the Ethernet terminal 2 (1) in FIG. 1, and “EtherIP3” in the tables Ta3 and Ta4 is the IP address of the Ethernet terminal 2 (3) in FIG. “IP3” is represented.

  The second parameter from the bottom of the address table is the MAC address of the Ethernet terminal that is the destination of the MAC frame from the Ethernet terminal represented by the parameter at the top. “Ieee1394MAC2” in the tables Ta1 and Ta3 represents the MAC address “MAC2” of the Ethernet terminal 2 (2) in FIG. 1, and “Ieee1394MAC4” in the tables Ta2 and Ta4 is the MAC address of the Ethernet terminal 2 (4) in FIG. “MAC4” is indicated.

  The fourth parameter from the top of the address table is the IP address of the Ethernet terminal represented by the second parameter from the bottom. “Ieee1394IP2” in the tables Ta1 and Ta3 represents the IP address “IP2” of the Ethernet terminal 2 (2) in FIG. 1, and “Ieee1394IP4” in the tables Ta2 and Ta4 is the IP address of the Ethernet terminal 2 (4) in FIG. “IP4” is represented.

  The fourth parameter from the bottom of the address table represents “56789012” (FIG. 3) which is the GUID of the bridge 1b in the Node-GUID table in the Node-GUID table memory 13 of FIG. (Conversely, the parameter in the fourth row from the bottom of the address table Tb1 to Tb4 on the bridge 1b side represents “1245678” which is the GUID of the bridge 1a.)

  The parameters in the third row from the bottom of the address table are the MAC frame with the Ethernet terminal represented by the top MAC address as the transmission source and the Ethernet terminal represented by the second MAC address from the bottom as the transmission destination. When an IP packet is sent to the bridge 1a with the Ethernet terminal represented by the third-stage IP address from the top as the transmission source and the Ethernet terminal represented by the fourth-stage IP address from the top as the transmission destination The offset address for transferring the data of the MAC frame or IP packet to the bridge 1b via the IEEE 1394 bus (the offset value in the 64-bit address space by IEEE 1212 to which IEEE 1394 complies). For example, “Ieee1394 proxy offset 21” in the table Ta1 represents the offset address “proxy offset 21” when the Ethernet terminal 2 (1) and the Ethernet terminal 2 (2) in FIG. 1 are the transmission source and the transmission destination, respectively.

  The parameter in the second row from the top of the address table is opposite to the parameter in the third row from the bottom, and the Ethernet terminal represented by the uppermost MAC address or the IP address of the third row from the top is the destination and When the Ethernet terminal represented by the second-stage MAC address from the top and the fourth-stage IP address from the top is the transmission source, the offset address is associated with the transmission destination and the transmission source. For example, “Ether proxy offset 11” in the table Ta1 represents an offset address “proxy offset 11” when the Ethernet terminal 2 (2) and the Ethernet terminal 2 (1) in FIG. 1 are the transmission destination and the transmission source, respectively. In the table in which the destination MAC address or IP address and the source MAC address IP address are interchanged with each other as in the table Ta1 and the table Tb1, the second offset address from the top and the third offset from the bottom The offset addresses are interchanged with each other.

  The parameter at the bottom of the address table is a registration time parameter indicating the time when the address table is registered (or the last used time when the address table is used after registration).

  Among these parameters in the address table, the IP address is an address used in the TCP / IP network layer, and is a value unique to each Ethernet terminal. The MAC address is an address used in the parameter link layer of TCP / IP, and is acquired by ARP (Address Resolution Protocol).

  The offset address is created by the controller 11 (FIG. 2) of the bridges 1a and 1b as a unique value for each address table. The controller 11 also creates registration time data when registering or using the address table.

  Thus, when two Ethernet terminals are connected to each of the bridges 1a and 1b, 2 × 2 = 4 address tables are registered in the bridges 1a and 1b, respectively. For example, when 100 Ethernet terminals are connected to each of the bridges 1a and 1b, 100 × 100 = 10000 address tables are registered in the bridges 1a and 1b, respectively.

  FIG. 5 is a flowchart showing a part related to the node number in the address management process executed by the controller 11 of FIG. Among these, the process of FIG. 5 (a) is a process for converting the MAC address (or IP address) of the transmission destination / transmission source, which is address information in the data transmission format in Ethernet, into a node number. The MAC address (or IP address) is converted into a GUID using the address table in the address management table memory 12 (the address table hit in the search) (step S1). Then, the GUID is converted into a node number using the Node-GUID table in the Node-GUID table memory 13 (step S2).

  The process of FIG. 5B is a process for converting the node number and the offset address into the MAC address (or IP address) of the transmission destination / transmission source, contrary to the process of FIG. The node number is converted into a GUID using the Node-GUID table in the Node-GUID table memory 13 (step S11). Then, the GUID and the offset address are converted into a destination / source MAC address (or IP address) using an address table (address table hit in the search) in the address management table memory 12 (step S12). .

  FIG. 6 shows an example in which the MAC frame is transferred from the Ethernet terminal 2 (1) to the Ethernet terminal 2 (2) in FIG. 1, and the processing of FIGS. 5 (a) and 5 (b) in the bridges 1a and 1b is performed. It is a figure which shows the mode of transfer of the included MAC frame.

  When a MAC frame is transferred from the Ethernet terminal 2 (1) to the Ethernet terminal 2 (2), the destination MAC address and the source MAC address in the MAC frame are “MAC2” and “MAC1”, respectively. When the MAC frame is sent from the Ethernet terminal 2 (1) via the Ethernet, the controller 11 of the bridge 1a uses “MAC2” and “MAC1” as search keys from the address table in the address management table memory 12. The address table in which the second parameter from the bottom is “MAC2” and the top parameter is “MAC1” is searched.

  In this case, the address table Ta1 in FIG. Therefore, the controller 11 of the bridge 1a uses the address table Ta1 to set the MAC address “MAC2” and “MAC1” of the transmission destination and the transmission source, the GUID “56789012” (parameter in the fourth row from the bottom), and the offset. The address is converted to the address “proxy offset 21” (the third parameter from the bottom) (step S1 in FIG. 5A). Subsequently, using the Node-GUID table (FIG. 3) in the Node-GUID table memory 13, this GUID “56789012” is converted into the node number “Node1” (step S2 in FIG. 5A).

  Then, the controller 11 of the bridge 1a controls the control logic 18 to convert the address-converted MAC frame into an isochronous packet or an asynchronous packet, which is a data transmission format in IEEE1394, and to IEEE1394. The data is transferred from the interface 16 to the bridge 1b via the IEEE1394 bus.

  When the packet is transferred from the bridge 1a, the controller 11 of the bridge 1b assigns the node number “Node 0” (FIG. 3) of the bridge 1a that is the transmission source in the packet to the Node-GUID in the Node-GUID table memory 13. Using the table, it is converted to the GUID “1245678” of the bridge 1a (step S11 in FIG. 5B). Subsequently, using the GUID “1245678” and the offset address “proxy offset21” as search keys, the GUID in the fourth row from the bottom in the address table in the address management table memory 12 is “1245678” and two rows from the top. The address table whose eye offset address is “proxy offset 21” is searched.

  In this case, the address table Tb1 in FIG. Therefore, the controller 11 of the bridge 1b uses this address table Tb1 to set “Node1”, “proxy offset21”, “MAC2” (the uppermost parameter) as a transmission destination, and “MAC1” (two steps from the bottom The eye parameter) is converted into a MAC address having the transmission source (step S12 in FIG. 5B).

  Then, the controller 11 of the bridge 1b controls the control logic 18 to convert the address-converted packet into a MAC frame and transfer it to the Ethernet terminal 2 (2) via Ethernet. Thereby, the transfer of the MAC frame from the Ethernet terminal 2 (1) to the Ethernet terminal 2 (2) is completed.

  In FIG. 6, the case of transferring the MAC frame is taken as an example, but when the IP packet is transferred from the Ethernet terminal 2 (1) to the Ethernet terminal 2 (2) in FIG. 1 (for example, before the MAC address is acquired by ARP). For example, in the case of a communication in which a table has already been created, the bridge 1a uses the address table Ta1 hit in the search to change the IP address “IP2” and “IP1” of the transmission destination and the transmission source to the GUID. After converting the address to “56789012” and the offset address “proxy offset21”, this Node ID “56789012” is converted to the node number “Node1” using the Node-GUID table. Further, the bridge 1b converts the node number “Node0” of the bridge 1a into the GUID “1245678” of the bridge 1a using the Node-GUID table, and then uses the address table Tb1 hit in the search to generate “Node1” ”And“ proxy offset 21 ”are converted into IP addresses having“ IP2 ”as the transmission destination and“ IP1 ”as the transmission source.

  The process shown in FIG. 5C is executed every time the bridge power is turned on or every time the IEEE1394 bus is reset. It is first determined whether or not the node numbers allocated to the bridges 1a and 1b have changed. (Step S21). If there is no change, the process is terminated as it is. On the other hand, if it has changed, the node number of the Node-GUID table in the Node-GUID table memory 13 is rewritten accordingly (step S22). Then, the process ends.

  FIG. 7 illustrates a state where the Node-GUID table of FIG. 3 is rewritten by the process of FIG. 5C. In this example, when the power of the bridges 1a and 1b is turned on or the bus is reset, the node numbers allocated to the bridges 1a and 1b are changed from “Node 0” and “Node 1” to “Node 1” and “Node 0”, respectively. The node number corresponding to the GUID “1245678” of the bridge 1a is rewritten from “Node 0” to “Node 1”, and the node number corresponding to the GUI ID “56789012” of the bridge 1b is rewritten from “Node 1” to “Node 0”. .

  As described above, in the bridges 1a and 1b, the node number is not set in the address table. Therefore, even if the node number changes when the power is turned on or the bus is reset, it is not necessary to rewrite the address table. Only one node-GUID table in 13 needs to be rewritten. As a result, even when the number of address tables is large, the processing when the node number of the node changes is simplified.

[Process related to determination of offset address]
Next, processing relating to determination of the offset address will be described. FIG. 8 is a flowchart showing processing executed by the controller 11 of FIG. 2 regarding determination of an offset address.

  Of these, the process of FIG. 8A is a process executed when the bridge power is turned on. First, the blank is a table having the same structure as the address table shown in FIG. The number of address tables required from the maximum number of Ethernet terminals that can be connected to the bridges 1a and 1b (for example, when up to 100 Ethernet terminals can be connected to each of the bridges 1a and 1b, As many as 100 × 100 = 10000) are prepared in the blank table memory 14 of FIG. 2 (step S31).

  A unique value that does not overlap each other is set as an offset address in these blank tables (step S32). This offset address determination method may be a method in which a certain value M is used as a reference value, and offset addresses set in each blank table are sequentially increased to M, (M + 1), (M + 2),. Alternatively, a random number generated by a random number generator may be determined as an offset address.

  FIG. 9 shows an example of the blank table in which the offset address is set in this way. FIG. 9A is a blank table on the bridge 1a side, and FIG. 9B is a blank table on the bridge 1b side. . As the parameters in the second row from the top of the set blank table, the offset addresses “proxy offset 11” and “proxy offset 21” are permanently used as “Ether proxy offset 11” and “Ether proxy offset 21”. Also, in the third row from the bottom of the blank table, the offset addresses “proxy offset 21” and “proxy offset 11” of the parameter in the second row from the top of the counterpart bridge are “Ieee 1394 proxy offset 21” and “Ieee 1394 proxy offset 11”, respectively. , Set each time the table is registered.

  As shown in FIG. 8A, after step S32, the process waits until the MAC address of the Ethernet terminal currently connected to the bridges 1a and 1b is acquired by ARP (address resolution protocol) (step S33). When the MAC address is acquired, parameters other than the offset address (as shown in FIG. 4, the MAC address and IP address of the transmission destination / transmission source and the GUID) are set in the blank table in the blank table memory 14. Thus, an address table is created, and the created address table is registered in the address management table memory 12 of FIG. 2 (step S34). And it returns to step S33 and waits until a new MAC address is acquired by ARP.

  The process of FIG. 8B is a process executed in response to a timeout process to be described later, and the address table registered in the address management table memory 12 is excluded from registration by the timeout process because it has not been used for a certain period of time. (Step S41). When a certain address table is excluded from registration, the address table is stored again in the blank table memory 14 of FIG. 2 as a blank table without changing the offset address (step S42). Then, the process returns to step S41 and waits until a new address table is excluded from registration.

  FIG. 10 is a diagram showing the relationship between the blank table and the address table by the processing of FIG. 8 (the linear table will be described later). A predetermined number of blank tables BT in which unique values are set as offset addresses and no other parameters are set are prepared in advance (steps S31 and S32 in FIG. 8A). When registering the address table AT, parameters other than the offset address are set in the blank table BT (steps S33 and S34 in FIG. 8A).

  After that, when a certain registered address table AT is not used for a certain period of time, the address table AT is again set to the blank table without changing the set offset address (second stage from the top). Treated as BT (steps S41 and S42 in FIG. 8B). Therefore, at the time of subsequent address table registration (steps S33 and S34 in FIG. 8A), the address table AT that has not been used for a certain period of time is initially set (second stage from the top). While the address is continuously used, a parameter other than the offset address is newly set and reused as a new address table AT.

  In this way, since the offset addresses set in the predetermined number of blank tables prepared at the beginning are repeatedly used, the offset addresses are not exhausted no matter how many times the address table is registered.

  Then, management of which offset address of the offset address determined once can be reused (management of which offset address is set in each address table that has not been used for a certain period of time) is performed. However, since the address table that has not been used for a certain period of time is simply treated as a blank table again, the processing is not complicated.

[Address table timeout processing]
Next, address table timeout processing will be described. FIG. 11 is a flowchart showing the address table timeout processing executed by the controller 11 of FIG. In this process, it is determined whether or not a new address table has been registered in the address management table memory 12 by the above-described process of FIG. 8A (step S51), and registered in the address management table memory 12. Whether or not any one of the address tables is newly used (step S52) and whether or not a certain period (for example, several seconds) for which timeout should be checked has elapsed (step S53) Repeat until the answer is yes.

  If the answer is yes in step S51, the newly registered address table is placed at the head of the linear table in which the registered address tables are arranged in a straight line (step S54). Then, the process returns to step S51.

  If the answer is yes in step S52, the newly used address table is placed at the head of the linear table, and the registration time data (FIG. 4) in the address table is updated to the current time (step S55). Then, the process returns to step S51.

  If the answer is yes in step S53, it is registered whether or not the address table located at the end of the linear table has been used for a predetermined time (for example, 2 to 3 minutes) from the address table located at the beginning of the linear table. The time data is checked in order, the address table that has not been used for a certain period of time is excluded from registration, and the check is terminated when an address table that has been used within a certain period of time is found (step S56). Then, the process returns to step S51.

  A state in which an unused address table is excluded from registration by the processing of FIG. 11 will be described with reference to FIG. It is a figure shown in relation to the process of above-mentioned FIG. When the address table is registered using the blank table by the process of FIG. 8A, the address table is arranged at the head of the linear table in which the registered address tables are arranged in a straight line (until the linear table until then). The address tables arranged in the table are shifted downward by one table) (step S54 in FIG. 11).

  Each time an address table already arranged in the linear table is used, the address table is arranged at the head of the linear table, and the registered time data of the address table is updated to the current time. (Step S55 in FIG. 11).

  As a result, the newly registered address table and the most recently used address table (address table with new registration time data) are arranged closer to the head of this linear table. Conversely, an address table that has passed since registration or an unused address table (an address table with old registration time data) is arranged at a position closer to the end of this linear table.

  Then, from the address table located at the end of this linear table to the address table located at the beginning of this linear table (that is, from the address table where the registration time data is old to the new address table) , It is checked in order using the registration time data whether it is not used for a certain period of time, address tables that are not used for a certain period of time are excluded from registration (also removed from the linear table) and used within a certain period of time The check is completed when the address table being found is found (step S56 in FIG. 11). The address table excluded from the registration is again treated as a blank table by the process of FIG.

  In this way, all registered address tables are arranged linearly in the order of registration time data, and by checking the unused time in order from the address table with the oldest registered time data, it is possible to Since it becomes unnecessary to perform the check, the processing load is reduced and the processing time is shortened.

  In addition, it does not manage what registration time data is currently set in each address table, but it simply arranges registered address tables in a straight line, resulting in an increase in circuit scale and complexity of processing. Nor.

  Note that the linear table as shown in FIG. 10 is specifically created as a bidirectional list in the C language. The bidirectional list represents the relationship between a plurality of data using a pointer to next data (next) and a pointer to previous data (previous). Here, as shown in FIG. 12, a linear table is created by representing the relationship of all registered address tables with these pointers (next) and (previous). A pointer (previous) of the address table arranged at the head of the linear table and a pointer (next) of the address table arranged at the end of the linear table indicate NULL (zero).

  When a new address table is registered, a registered address table is newly used, or an unused address table is excluded from registration, the arrangement of the address table in the linear table is changed (FIG. 11). In steps S54 to S56, the pointers (next) and (previous) of the address table to be changed are rewritten.

  The pointer for creating this linear table is stored in the address management table memory 12 of FIG. 2 together with the parameters of the address table (FIG. 4) for each address table. FIG. 13 shows parameters and pointers stored in the address management table memory 12 for one address table (the pointer for creating the hash linear table will be described later).

  By creating the linear table as such a bidirectional list, the capacity of the memory (pointer area 12b) for storing the linear table can be reduced, and the arrangement of the address table in the linear table can be easily changed. Can do.

[Process related to address table search]
Next, processing related to address table search will be described. FIG. 14 is a flowchart showing a process executed by the controller 11 of FIG. 2 regarding an address table search. Among these, the process of FIG. 14A is a process executed when the bridge power is turned on. First, the number of bases m × n (m and n are fixed values) in the hash table memory 15 of FIG. A two-dimensional MAC address hash table, a two-dimensional IP address hash table having a base number m × n, and a two-dimensional GUID / offset address hash table having a base number m × n are respectively created (steps). S61).

  Subsequently, a determination is made as to whether or not a new address table has been registered in the address management table memory 12 by the above-described processing of FIG. 8A (step S62). The determination as to whether or not the address table is excluded from registration (step S63) is repeated until the answer becomes yes in any step.

  If YES in step S62, 0 to (m-1), 0 from the MAC address of the transmission source and the transmission destination in the parameters of the newly registered address table (the top row in FIG. 4, the second row from the bottom), respectively. To the hash value of (n-1) by an appropriate method (for example, a method of dividing the number of bytes of the MAC address of the transmission source and the transmission destination by m and n, respectively, and obtaining the remainder), and calculating the hash value of the hash value The address table is stored in the table position of the MAC address hash table corresponding to the combination. Similarly, from the source and destination IP addresses (the third stage from the top and the third stage from the top in FIG. 4) in the parameters of the address table, 0 to (m−1) and 0 to (n−1), respectively. ) And the address table is stored in the table position of the IP address hash table corresponding to the combination of the hash values. Also, hash values from 0 to (m−1) and 0 to (n−1) from the GUID and offset address (fourth row from the bottom in FIG. 4 and second row from the top) in the address table parameters, respectively. Calculate and store the address table in the table position of the GUID / offset address hash table corresponding to the hash value combination. (Step S64). Then, the process proceeds to step S66.

  FIG. 15 shows an example of the MAC address hash table in which the address table is stored in step S64. The horizontal direction (X direction) in the figure represents the hash value of the source MAC address (EtherMAC), and the vertical direction (Y direction) in the figure represents the hash value of the destination MAC address (Ieee 1394 MAC). In this example, the address table AT is stored at the table position corresponding to the combination (1, 2) of hash values calculated from the MAC addresses of the transmission source and the transmission destination.

  Specifically, these hash tables are created as an array of pointers in the C language, and when an address table is stored at a certain table position, the address table is pointed to by the table position pointer. By creating the hash table as such an array of pointers, the capacity of the hash table memory 15 can be reduced, and it is easy to save the address table in the hash table memory 15 and to delete the address table from the hash table memory 15. Can be done.

  If YES in step S63, the address table from which registration has been excluded is deleted from the MAC address hash table, IP address hash table, and GUID / offset address hash table (step S65). Then, the process proceeds to step S66.

  In step S66, it is determined whether or not hash collision occurs at any one of the MAC address hash table, IP address hash table, and GUID / offset address hash table.

  If no, the process returns to step S62. On the other hand, if the answer is yes, a hash linear table in which a plurality of address tables in which hash collision has occurred is arranged in a straight line is created (the hash linear table has already been created and hash collision has occurred. When the address table is newly increased or decreased or replaced, the arrangement of the address table in the hash linear table is changed accordingly (step S67). Then, the process returns to step S62.

  FIG. 16 shows the hash linear table created in step S67. A plurality of address tables AT in which hash collisions occur are arranged in a straight line. More specifically, this hash linear table is also created as a bidirectional list in the C language, similar to the linear table shown in FIG. The pointer of the bidirectional list is also stored in the address management table memory 12 of FIG. 2 together with the parameters of the address table (FIG. 4) for each address table. In FIG. 13, the hash collision occurred in the IP address hash table and the pointer for creating the hash linear table in which a plurality of address tables in which the hash collision occurred in the MAC address hash table are arranged in a straight line. Create a hash linear table in which a hash linear table in which a plurality of address tables are arranged in a straight line and a plurality of address tables in which a hash collision has occurred in the GUID / offset address hash table are arranged in a straight line. A pointer is shown.

  The process of FIG. 14B is a process executed when searching the address table. First, in the linear table of FIG. 10, the first part (for example, 10 to 20 address tables are arranged). ) As a cache, and two search keys (source and destination MAC address, source and destination IP address, GUID and offset address) for the address table in the cache linear table ) To search (step S71). Then, it is determined whether or not the search has been hit (step S72).

  If yes, the process ends. On the other hand, if the answer is no, hash values are calculated from the two search keys in the same manner as in step S64 in FIG. 14A, and hash tables created by the process in FIG. MAC address hash table if it is an address, IP address hash table if the search key is an IP address, and GUID / offset address hash table if the search key is a GUID and offset address) The address table stored in the table position corresponding to the combination is searched (step S73).

  Subsequently, it is determined whether there is only one address table stored in the table position (whether or not hash collision occurs) (step S74). If yes, the process is terminated (the one address table is treated as an address table hit in the search).

  On the other hand, if the answer is no, the two search keys are compared with the address table in the hash linear table created by the process of FIG. 14A for the plurality of address tables in which the hash collision occurs at that position. The search is performed (step S75). Subsequently, the address table hit in the search is moved to the head of the hash linear table (step S76). Then, the process ends.

  One address table hit in the search in the process of FIG. 14 (b) is the same as the address conversion from the MAC address (or IP address) to the GUID and the offset address, as described with reference to FIG. Used for address conversion from offset address to MAC address (or IP address).

  FIG. 17 shows how the address table is searched by the processing of FIG. First, a search is performed with two search keys on the address table in the cache linear table, which is the first part of the linear table in FIG. 10 (step S71 in FIG. 14B). In this way, by performing a search first on a predetermined number of address tables that have been used most recently, there is a higher possibility of hits from those address tables, so the time required for the search is reduced. Is more likely.

  If no hit is found, the address table stored in the table position corresponding to the combination of hash values calculated from the two search keys in the hash table illustrated in FIG. 15 is specified. (Step S73 in FIG. 14B).

  In this way, the address table is stored in a hash table having a table position composed of a combination of two hash values, and the address table is searched from this hash table by a combination of hash values calculated from two search keys. Even if the number of registered tables is large, the time required for searching is shortened.

  Further, a hash linear table is created in which a plurality of address tables in which hash collision has occurred are arranged in a straight line, and hash collision occurs at the table position of the hash table corresponding to a combination of hash values calculated from two search keys. If it has occurred, the address table arranged in the hash linear table is searched with the two search keys, and the address table hit in the search is moved to the head of the hash linear table. (Steps S75 to S76 in FIG. 14B). As a result, it is possible to efficiently search a plurality of address tables in which hash collisions have occurred.

  In order to reduce hash collisions, it is desirable to increase the number of bases of the hash table (increase the values of m and n). However, if the maximum number of registered address tables is large, the number of bases Since hash collisions cannot be reduced without increasing the size of the hash table, the size of the hash table increases. On the other hand, even if hash collisions occur, it is possible to efficiently perform a search using the hash linear table as described above, so it can be said that there is no need to reduce hash collisions so much. Therefore, the number of bases of the hash table may be relatively smaller (for example, about 1/100) than the maximum number of registered address tables.

  Finally, some examples of modification of the present invention will be given. In the above example, as shown in FIG. 10, a predetermined number of blank tables in which only offset addresses are set are prepared in advance, the address tables are registered using the blank tables, and the registered addresses are arranged in a linear table. After the time-out process, the address table excluded from registration is treated as a blank table again.

  However, as another example, as shown in FIG. 18, every time an address table is registered, a blank table in which no parameters are set is created, all parameters are set in the blank table, and the registered addresses are set. The address table excluded from the registration may be deleted after the time-out process is performed by arranging in the linear table.

  Alternatively, it may be as shown in FIG. That is, first, as shown in FIG. 19A, a linear table in which a predetermined number of blank tables BT in which only offset addresses are set (or no parameters are set) is arranged in a straight line is created in advance, and A border pointer PO indicating the boundary between the blank table BT and the registered address table is set at the head of the linear table.

  Thereafter, when registering the address table, as shown in FIG. 19B, parameters (or all parameters) other than the offset address are set in order from the blank table BT at the bottom of the linear table, and one address table AT is set. Each time registration is performed, the address table AT is moved to the head of the linear table and the border pointer BP is moved downward by one table. Each time the registered address table AT is used, the address table AT is moved to the head of the linear table.

  Then, the registration time data is used to determine whether or not the linear table has been used for a certain period of time from the address table AT positioned above the border pointer BP to the address table AT positioned at the head of the linear table. The address table AT that has not been used for a predetermined time is moved again to the blank table BT side of the linear table, and the border pointer BP is moved to the position where the address table used within the predetermined time is found. Move.

  Thus, as shown in the example of FIG. 19, by managing the blank table used for registering the address table and the registered address table in the same linear table, the management of the blank table and the unused address table Can be easily performed as a blank table.

  Also, in the above example, in response to setting the GUID as a parameter in the address table as shown in FIG. 4, a GUID / offset address hash table is created as shown in FIG. A search is performed using two search keys of GUID and offset address using the offset address hash table. However, for example, the node number is set as a parameter in the address table (in this case, a node number / offset address hash table is created, and two search keys for the node number and offset address are created using this node number / offset address hash table) You can make a search with.

  In the above example, as shown in FIG. 17, before searching the hash table, the head part of the linear table in FIG. 10 is used as a cache, and the address table in the cache linear table is used. Searching for. However, for example, when a linear table as shown in FIG. 10 is not created, a few (for example, about 10 to 20) recently used address tables are temporarily stored as a cache, and before searching a hash table, You may make it search with respect to the address table preserve | saved as the cache.

  In the above example, pointers for creating the bidirectional table of the linear table of FIG. 10 and the hash linear table of FIG. 16 are set for each address table together with the parameters of the address table of FIG. It is stored in the address management table memory 12. However, as another example, a memory for storing address table parameters and a memory for storing pointers for each address table may be provided separately. However, it is better to store parameters and pointers together for each address table as in the above example for the first part of the linear table (cache linear table) and the address table arranged in the hash linear table. After performing a search, it is possible to easily and quickly perform an address conversion process using the address table parameters that have been found in the search.

  In the above example, as shown in FIG. 1, two bridges 1a and 1b are connected to the IEEE 1394 bus. However, the present invention is not limited to this, and three or more bridges may be connected to the IEEE 1394 bus.

It is a figure which shows the whole LAN structure to which this invention is applied. It is a block diagram which shows the structural example of the bridge | bridging of FIG. It is a figure which illustrates the Node-GUID table in the Node-GUID table memory of FIG. FIG. 3 is a diagram illustrating an address table in the address management table memory of FIG. 2. It is a flowchart which shows the process which the controller of FIG. 2 performs regarding a node number. It is a figure which illustrates the mode of transfer of the MAC frame in this invention. It is a figure which illustrates the mode of rewriting of a Node-GUID table. It is a flowchart which shows the process which the controller of FIG. 2 performs regarding determination of an offset address. It is a figure which illustrates the blank table which set the offset address. It is a figure which shows the relationship between a blank table and an address table, and a linear table. It is a flowchart which shows the time-out process which the controller of FIG. 2 performs. It is a figure which shows the preparation method of a linear table. It is a figure which shows the parameter and pointer stored in the address management table memory of FIG. 2 about one address table. It is a flowchart which shows the process which the controller of FIG. 2 performs regarding the search of an address table. It is a figure which illustrates the hash table in this invention. It is a figure which illustrates the linear table for hash. It is a figure which shows the mode of a search of the address table in this invention. It is a figure which shows the example of a change of this invention. It is a figure which shows the example of a change of this invention. 1 is a diagram illustrating a LAN in which Ethernet terminals are connected using IEEE1394. FIG. It is a figure which illustrates an address table. It is a figure which illustrates the mode of transfer of a MAC frame. It is a figure which shows an example of a timeout process.

Explanation of symbols

  1a, 1b bridge, 2 (1), 2 (2), 2 (3), 2 (4), ... Ethernet terminal, 3 (1), 3 (2), ... digital AV equipment, 11 controller, 12 address management Table memory, 13 Node-GUID table memory, 14 Blank table memory, 15 Hash table memory, 16 Ethernet interface, 17 IEEE 1394 interface, 18 Control logic

Claims (6)

A format conversion is performed between the data transmission format of the Ethernet and the data transmission format of the IEEE 1394, and between the address information in the data transmission format of the Ethernet and the address information in the data transmission format of the IEEE 1394. In a bridge device that performs address translation,
First processing means for registering an address table in which address information in the data transmission format in Ethernet, address information in the data transmission format in IEEE 1394, and registration time data are set as parameters;
A second processing means for arranging the address table at the head of the linear table in which the registered address tables are arranged in a straight line each time the address table is newly registered;
Each time an address table already arranged in the linear table is used, the address table is moved to the head of the linear table, and the registered time data of the address table is updated to the current time. Processing means;
From the address table located at the end of the linear table to the address table located at the beginning of the linear table, it is checked in order using the registered time data whether it has been used for a certain period of time. And a fourth processing means for ending the check when the address table used in the method is found. The bridge device according to claim 1, wherein
At least processing means for creating a predetermined number of blank tables in which parameters other than the offset address that is address information in the data transmission format in IEEE 1394 are not set;
The first processing means sets a parameter other than an offset address in the blank table. The bridge device according to claim 1, wherein
The first processing means creates a blank table in which no parameter is set when registering an address table, and sets the parameter in the blank table. The bridge device according to claim 1, wherein
A linear table in which a predetermined number of blank tables with no parameters set are arranged in a straight line, and a border pointer indicating a boundary between the blank table and the registered address table is set at the head of the linear table Further comprising processing means,
The first processing means sets a parameter in a blank table at the end of the linear table,
Each time the address table is newly registered, the second processing means arranges the address table at the head of the linear table and moves the border pointer downward by one table,
The fourth processing means determines whether the linear table has been used for a certain period of time from an address table positioned above the border pointer to an address table positioned at the head of the linear table. A bridge device characterized by sequentially checking using registered time data. A format conversion is performed between the data transmission format of the Ethernet and the data transmission format of the IEEE 1394, and between the address information in the data transmission format of the Ethernet and the address information in the data transmission format of the IEEE 1394. In the address management method in the bridge device that performs address translation,
Register an address table in which address information in the data transmission format in Ethernet, address information in the data transmission format in IEEE 1394, and registration time data are set as parameters,
Each time the address table is newly registered, the address table is arranged at the head of the linear table in which the registered address tables are arranged in a straight line.
Each time an address table already arranged in the linear table is used, the address table is moved to the head of the linear table, and the registration time data of the address table is updated to the current time,
From the address table located at the end of the linear table to the address table located at the beginning of the linear table, it is checked in order using the registered time data whether it has been used for a certain period of time. The address management method is characterized in that the check is terminated when the address table used in the process is found. In a network system in which a plurality of bridge devices that relay Ethernet and IEEE 1394 are connected to an IEEE 1394 bus, and terminals having Ethernet interfaces are connected via the bridge device.
Each of the bridge devices
First processing means for registering an address table in which address information in the data transmission format in Ethernet, address information in the data transmission format in IEEE 1394, and registration time data are set as parameters;
A second processing means for arranging the address table at the head of the linear table in which the registered address tables are arranged in a straight line each time the address table is newly registered;
Each time an address table already arranged in the linear table is used, the address table is moved to the head of the linear table, and the registered time data of the address table is updated to the current time. Processing means;
From the address table located at the end of the linear table to the address table located at the beginning of the linear table, it is checked in order using the registered time data whether it has been used for a certain period of time. And a fourth processing means for ending the check when the address table used in the network is found.

Images