JPH03254245A - Channel identifier communication system - Google Patents

Abstract

PURPOSE:To use an address area efficiently by using a variable length of logic channel number for an address of a packet. CONSTITUTION:The entire length of a fixed length address number is fixed in addition to the case that a node is connected to plural transmission lines in a different direction and a border between a logic node address and a logic channel number is allocated variably. That is, only the border between the logic node address and the logic channel number is moved optionally. Each node detects the border between the logic address and the logic channel number automonously from the address field of the received packet. Thus, the address area of the transfer packet is used efficiently.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a channel identifier communication system using a logical node address and a logical channel number, and in particular, a channel identifier communication method that uses a logical node address and a logical channel number, and in particular, a channel identifier communication method that can exist simultaneously on a ring-shaped or bus-shaped transmission path. When a plurality of nodes with different numbers of communications are connected, a channel identifier communication method is used in which the information packet is transferred by adding a header containing address information that specifies the communication partner.

[Conventional technology]

Conventionally used packet addressing methods include (i) Ethernet compliant with the IEEE 802.3 standard, token bus compliant with the IEEE 802.4 standard, and token ring compliant with the IEEE 802.5 standard. (Li) For example, there is a "channel identifier communication method" described in the specification and drawings of Japanese Patent Application No. 1-54461. ”, and the logical address of the node housing the terminal.

There is a method in which a set of channel numbers for identifying communications being performed simultaneously at the node is stored in the packet (at the beginning).

In particular, the latter is an address (IEEE
Since the logical node address and the logical channel number (about 2 bytes) are used instead of using 6 bytes (standard), redundant address areas can be reduced and addresses can be verified at high speed. In this case, in the latter method, the lengths of the logical node address and the logical channel number are each set to fixed lengths.

[Problem to be solved by the invention]

However, in the method of using the address area of a transfer packet in the channel identifier communication system described in the drawing of Specification 2 above, when there are many communications that can exist simultaneously, if the logical channel number area is small, the logical There are cases where the number of channel numbers cannot be satisfied, and when the number of communications that can exist at the same time is small, there is a problem in that each node has unused logical channel numbers.

If this problem is to be solved within the scope of the channel identifier communication method described above, the logical node address area should be made relatively long, the number of logical channels that can be used with one logical node address should be small, and the number of simultaneous communications should be large. A possible method is to assign multiple logical node addresses to a node. However, with this method, the number of address patterns that must be matched increases when receiving a packet, so
This adds disadvantageous conditions when performing high-speed processing.

SUMMARY OF THE INVENTION An object of the present invention is to provide a channel m531J child communication system that can solve these conventional problems and efficiently use the address area of transfer packets.

[Means for Solving the Problems] In order to achieve the above object, the channel identifier communication system of the present invention provides (a) a boundary between a logical node address and a logical channel number that is variable for each node, and that the logical node address When located in the upper digits of the above fixed length address area, the short logical node address is assigned so as not to overlap with the upper digits of the long logical node address, and the logical node address is assigned to the lower digits of the above fixed length address area. When the logical node address is located, the characteristic is that the short logical node address is assigned so as not to overlap with the lower digits of the long logical node address. (b) Any one node can manage the entire address space as an address management node, and can individually request the number of logical channel numbers required by all other nodes from the address management node. Another feature is that the address management node distributes a logical node address and an available logical channel number area to each requesting node, and notifies the node. Also, (c) any one node manages the entire address space as an address management node, first occupies its own logical node address and available logical channel number area,
The unused address area that was not occupied is notified to the adjacent node, and the adjacent node occupies its own logical node address and available logical channel number area from the unused address area, and provides the unused address area to the adjacent node. Another feature is that all nodes occupy their own logical node addresses and usable logical channel number areas by sequentially performing the process of notifying address areas. (d) An arbitrary l-node manages all address spaces as an address management node, and sends packets indicating unused address space within the communication system and a specific logical channel number length (hereinafter referred to as permission number length). The adjacent node receives the above packet, and when the permission number length written in the packet matches the logical channel number length required by its own node, it uses the logical node address from the address space. By occupying the possible logical channel number area and, if there is a mismatch, transferring the received packet as it is to the adjacent node without performing the occupancy process, the packet is circulated multiple times and all nodes Another feature is that each logical node address and available logical channel number area are occupied.

Furthermore, (e) a method that uses multiple ring-shaped transmission links independently in different directions, and a method that uses the above transmission link configuration and uses two transmission lines with different transmission directions back and forth between two nodes. Another feature is that a different logical node address is assigned to each node depending on the transmission path.

[For production]

In the present invention, the overall length of the fixed-length address area is fixed, even when a node is connected to multiple transmission lines in different directions, and the logical node address and logical channel number are Assign boundaries variably to each node. In other words, only the boundary between the logical node address and logical channel number is moved arbitrarily. Each node autonomously detects the boundary between the logical node address and logical channel number from the address field of the received packet.

The method for allocating logical node addresses and logical channel numbers is as follows: (1) After the address management node receives address space requests from all nodes, it allocates each node and assigns the upper bits of that address space to a logical node address. (■) Sequentially occupying enough address space to satisfy the required number of logical channel numbers starting from any node;
Address space is allocated in order from the node with the largest number of required logical channel numbers by the method of determining the upper bits of the simultaneously occupied address space as the logical node address and (iii) by the address management node sending out an address occupancy permission notification packet. There are ways to occupy it.

As a result, each node connected to a ring-shaped or bus-shaped transmission link can secure a number of logical channel numbers commensurate with the number of simultaneous communications, and can reduce the number of unused logical channel numbers.

〔Example〕

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a configuration diagram of a packet used in the present invention.

As shown in FIG. 1, a packet is composed of a fixed length address area and a user information area. The fixed-length address area stores the logical node address that specifies the packet receiving or releasing optical node as a channel identifier, and the logical channel number specified as the channel identifier to identify multiple communications that exist at the same node at the same time. do. In the packet of the present invention, the entire address area is set to a fixed length, but within that range, the boundary between the logical node address and the logical channel number is moved arbitrarily. In that case, set the short logical node address so that it does not overlap with the beginning part of the long logical node address.

In FIG. 1, the logical node address is located in the upper digits and the logical channel number is located in the lower digits, so a node receiving this packet must first identify the logical node address. For example, (000), (0010)
, (100), (01), the short logical node address (for example, 01
) does not overlap with the upper digits (in this case, OO) of a long logical node address (eg, 0010). Therefore, in addition to these, logical addresses such as (0101) and (OIO) should not be assigned.

Conversely, if the logical channel number is located in the upper digits and the logical node address is located in the lower digits, the node receiving the packet identifies the logical node address using the last bit of the fixed-length address area. In that case, it is necessary to assign the short logical node address so that it does not overlap with the lower digits of the long logical node address. For example, for the above logical node address (01), (0101),
(001).

(101) must not be added.

FIGS. 2(a) to 2(c) are a block diagram of a ring system, an operation time chart, and a diagram of the entire address space in the ring system, respectively, showing the first embodiment of the present invention.

Although not shown in the figure, each node is connected to a plurality of terminals. Further, although a ring-shaped transmission line is shown in the figure, the same applies to a bus-shaped transmission line.

In this embodiment, the packet address length is 8 bits, and the maximum number of logical channel numbers that can be used by one node is 64 (6 bits).

As shown in FIG. 2(a), each node 11-14 currently requires 16, 32, 32, and 64 logical channel numbers, respectively. Among these, node 11 is a node that manages all logical channel numbers and logical node addresses. Further, the transmission direction of packets transferred through the ring is the direction of the arrow.

Each of the nodes 11 to 14 knows only the number of logical channel numbers that it requires, and each of these nodes 11 to 14 notifies the address management node 11 of the number of logical channel numbers. Address management node 1
1 distributes the address space within the ring system to each node and notifies each node 11 to 14 of the available address space.

The address space assignment procedure will be explained with reference to FIGS. 2(b) and 2(c).

(a) First, the node 12 sends out onto the ring an address space request packet a having as information a unique physical address (PA=12) that is not duplicated within the ring system and the number of requested logical channel numbers=32.

In the header of address space request packet a, a unique address within the ring system (11111111) indicating that it is an address space request packet is written. In this way, the address space request packet a sent from the node 12 is received by the address ring node l].

(b) The node 13 sends out onto the ring an address space request packet b having as information a unique physical address (PA=13) that is not duplicated within the ring system and the number of requested logical channel numbers=32.

At this time, a unique address within the ring system (11111111) indicating that the address space request packet is an address space request packet is written in the header of the address space request packet b. In this way, the address space request packet b sent from the node 13 arrives at the adjacent node 12. Since the node 12 recognizes that the packet is an address space request packet by referring to the address, the packet is not received by the node 12 and is relayed as is on the ring.

The address space request packet b sent from the node 13 and relayed by the node 12 is received by the address management node 11.

(c) The node 14 sends out onto the ring an address space request packet C having as information a unique physical address (PA=14) that does not overlap within the ring system and the number of requested logical channel numbers=64. At this time, in the header of the address space request packet C, a unique address within the ring system (1, 1111111) indicating that it is an address space request packet is written. In this way,
Address space request packet C sent from node 14
In the same way as explained in (b) above, note 13 and 1
2 and then received by the address management node 11.

(d) In node 11, the number of requested logical channel numbers of own node 11 = 16, and each node described in the information part of address space request packets a, b, and c sent from node 12, 13, and 14. The address space is distributed based on the number of requested logical channel numbers.

(E) As shown in FIG. 2 (C), first, the address space 00000000 to 00011111 (the number of logical channel numbers is 32) is allocated to the node 12, and the upper 3 bits of that space that are not used as logical channel numbers are allocated to the node 12. (
000) is the logical node address of node 12.

(to) Next, address space 00100000-001
01111 (the number of logical channel numbers is 16) is assigned to the node 11, and of that space, the upper 4 bits (○Ol○) that are not used as the logical channel number are set as the logical node address of the node 11.

(G) Next, address space O1, OOOOO○ ~ 01
111111 (logical channel number 64) to node I4
The upper two bits (01) of that space, which are not used as the logical channel number, are used as the logical node address of the node 14.

Kotsutoki, 00110000~001 ], 1111
The space is a prohibited address space. This means that if this space is distributed as the address space of other nodes, the logical node address of the allocated node will be 001.
Then, a packet addressed to node 11 whose packet address is OOI 0XXXX (where x is an arbitrary value) is
This is to prevent the reception from occurring due to pattern matching of the upper 3 bits. That is, this prevents the address from being mistakenly identified as the logical node address (OOI O) of the node 11.

(H) Next, 10000000 to 100 of the address space
11111 (the number of logical channel numbers is 32) is assigned to the node 13, and the upper three bits (100) of that space that are not used as the logical channel number are set as the logical node address of the node 13.

As a result of the distribution, the address space distributed to each node has the distribution shown in 11N(c).

(v) Next, the node 11 sends out on the ring an address space notification packet d having information about the physical address of each node and the address space distributed to each node for all nodes. In that case, the header of the address space notification packet d contains a unique address within the ring system indicating that it is an address space notification packet <1111
1110). In this way, the address space notification packet d sent from the node 11 is first received by the node 14.

(l) When the node 14 receives the address space notification packet d, the node 14 receives the physical address (P
Based on the allocated address written in pair with A=14), the address space usable by the own node 14 is recognized. Node 14 sends this address space packet d onto the ring again.

In this way, the address space notification packet d is sequentially transferred to nodes 13 and 12, so that node 1
The same processing as the node 14 is performed at nodes 3 and 12 as well. As a result, all nodes 11 to 14 recognize address spaces that can be used by each node.

(w) Once the distribution of address space is completed, normal communication between nodes begins. Of the distributed address space, each node uses the number of bits corresponding to the required number of logical channels as a logical channel number area, and uses the remaining area as the logical node address of that node.

That is, the bit range of the logical node address and logical channel number on the right side of FIG. 2(C) is taken as the address area of the packet transmitted by each node.

(W) The timing for sending the address space notification packet d of the address management node 11 is
1 after a predetermined time has elapsed since the address space request packet was first received. In addition, if an address space request packet is received after sending the address space notification packet d, a new address space is allocated from among the unused address spaces managed by the address management node 11, and the address space is allocated to that node. The address space notification packet d is used for notification. In Figure 2 (C), 1
The address space from 0100000 to 11111111 is an unused address.

Note that, in order to deal with code errors on the transmission ring, a plurality of address space request packets and address space notification packets are continuously transmitted. In that case, other nodes execute the procedure because the procedure is valid only if the consecutively sent packets show the same value, but if the consecutively sent packets have different values, these packets are The receiving node judges these as invalid and does not execute the procedure.

FIGS. 3(a) to 3(d) are a block diagram of a ring system showing an embodiment of M2 of the present invention, an operation time chart, a diagram of the entire address space in the ring system, and a diagram of an example of logical node address calculation. It is.

In this embodiment, the packet address length is 8 bits, and the maximum number of logical channel numbers that can be used by one node is 64 (6 bit length).

As shown in the third [11(a)], each node 21-24 requires logical channel numbers of 16, 32, 32, and 64, respectively.

In this embodiment, the address management node does not allocate address space as in the case of FIG. 2, but any node becomes an address management node and sequentially assigns the address space to adjacent nodes by occupying and securing the necessary address space. Only advertise remaining address space.

That is, each node knows only the number of logical channel numbers it requires. After occupying the necessary address space from any one node in the ring system (for example, node 21, which automatically becomes the address management node), the unused address space is notified to adjacent nodes. By sequentially performing the above processing, each node divides the address space and occupies it.

The address space assignment procedure will be explained in detail with reference to FIGS. 3(b) and 3(C).

(a) First, the node 21 selects an address space (00000000-000
01111) (the number of logical channel numbers is 16), and the logical node address of its own node 21 is (OOOO). Thereafter, the node 21 sends out on the ring an unused address space notification packet e having the remaining address space of the occupied address area as information. In that case, the header of the unused address space notification packet e contains a unique address within the ring system (I I 11zol) indicating that it is an unused address space notification packet.

(b) Next, when the node 24 receives this unused address space notification packet e, it recognizes from the address header of the arriving packet that the packet is an unused address space notification packet e.

The node 24 acquires an address space (10,000
000-10111111) (Number of logical channel numbers 64
), and the logical node address of its own node 24 is (10). Of the unused address space, (
11000000 to 11.111111) are prohibited address spaces, as explained in the first embodiment.

Thereafter, the node 24 sends out on the ring an unused address space notification packet f having as information the remaining address space excluding the area occupied by the node 24 itself. In this case, a unique address within the ring system (11111101) indicating that the packet is an unused address space notification packet is written in the header of the unused address space notification packet f.

(C) Next, when the node 23 receives the unused address space notification packet f sent from the node 24, the node 23 determines the logic required by the node 23 from the unused address space written in the information part of the packet. Address space sufficient to satisfy the number of channel numbers (01000000 to 0
1011111) (logical channel number 32), and the logical node address of its own node 23 is (010).

Note that among the unused address spaces, (01100000
~01111111) becomes a prohibited address space for the same reason as described in the previous embodiment. after that,
The node 23 sends out on the ring an unused address space notification packet g having information about the remaining address space excluding the occupied area. In this case, the header of the unused address space notification packet g contains a unique address (11111101) within the ring system indicating that it is an unused address space notification packet.

(d) Next, when the node 22 receives the unused address space notification packet g sent from the node 23, the node 22 determines the logic required by the node 22 from the unused address space written in the information part of the packet. Address space that satisfies the number of channel numbers (○0100000 to 0
0111111) (logical channel number 32), and the logical node address of its own node 22 is (001).

Thereafter, the node 22 sends out on the ring an unused address space notification packet h having as information the remaining address space excluding the occupied address area. In this case, a unique address within the ring system (11111101) indicating that the unused address space notification packet is an unused address space notification packet is written in the head of the unused address space notification packet h.

(E) When the node 21 receives the unused address space notification packet h sent by the node 22, it recognizes that the address space division has been completed. Thereby, the node 21 sends out the address distribution completion packet i onto the ring. In that case, the header of the address distribution completion packet 1 contains a unique address within the ring system (11111100) indicating that it is an address distribution completion packet.
Describe. Note that the address distribution completion packet l is sent by the node that first occupies the address space and first sends the unused address space notification packet e.

(f) When the distribution of address space is completed in this way, normal communication between nodes begins. Each node uses the number of bits corresponding to the required number of logical channel numbers in the occupied address space as a logical channel number area, and uses the remaining area as the logical node address of each node.

Next, a specific method of calculating the logical node address of each node when using this embodiment will be explained using FIG. 3(d).

Here, assuming an M-bit address space, the number of logical channel numbers required by each node is 2 to the N power (
In FIG. 3(d), M=8). An M-bit final address indication field is provided in the user information area of the unused address space notification packet to indicate unused address space.

(i) First, any node (here, node 21) calculates the number of logical channel numbers required for its own node to N (here, since the number of required channel numbers = 16 = 2'', N = 4
), and the logical node address is set to M-N bits, all O's (here, the number of bits is 8-4=4 bits, 000).

(ii) The final address field at the first node (21) is 00001111. Therefore, in the unused address space notification packet that the first node (21) sends to the adjacent node, the upper M-N bits of the first address indication field are set to the logical Nord address, and the lower bits are set to all l. The content is (00
001 ] 11).

(iii) The adjacent node that received the unused address space notification packet (in this case, the node 22 that received it last) selects N from the number of logical channel numbers (32) required for its own node.
Since 2' = 32, N = 5), calculate the final address indication field + (2 to the power of N): 0010
1111), and the upper M-N bits (8-5=3) are set as the logical node address (001).

(iv) Calculate the final address field (00111111) in the same manner as the procedure (11) above.

(V) In the same way, sequentially select adjacent nodes (here,
Also in the nodes 23 and 24), when the number of logical channel numbers is 32, when N=5.64, N=6 is determined, the same address calculation as above is performed, and the logical node address and final address field are output.

Also in this embodiment, in order to deal with code errors on the transmission ring, each node consecutively sends out multiple identical unused address space notification packets, and each node consecutively performs the same processing multiple times. Repeat times. The node that first sent out the unused address space notification packet e inspects the circulating unused address space notification packets and determines whether the granting procedure is correct or not. If assignment fails, an unused address space notification packet is sent again.

FIGS. 4(a) to 4(d) are a block diagram of a ring system showing a third embodiment of the present invention, an operation time chart, a diagram of all address spaces in the ring system, and a diagram of an example of calculation of logical node addresses. It is.

In the first or second embodiment, the address management node determines the allocation, or the address space is sequentially occupied in the order of connected nodes from an arbitrary node, and notifications are made in the order of connection. However, in this embodiment, the necessary address space is sequentially occupied starting from the node with the largest number of required logical channel numbers, and the remaining nodes are notified.

In this embodiment as well, the packet address length is 8 bits, and the maximum number of logical channel numbers that can be used by one node is 64.
(6-bit length).

As shown in FIG. 4(a), each of the nodes 31 to 34 has a logical channel number of 164! , 32, 32 and 64 are required.

Each node knows only the number of logical channel numbers it requires. An arbitrary node in the ring system (here, node 31) is successively notified of available address spaces from adjacent nodes. Each node divides and occupies the address space by sequentially notifying the required number of channel numbers in the first round, and occupying them in descending order of the required number in the second round.

The address space assignment procedure will be described with reference to FIGS. 4(b) and 4(c).

(B) The node 31 has an address occupancy permission that has information indicating that only nodes with an unused address space and a requested logical channel number of 64 have the address occupancy processing right (hereinafter, occupancy permission number = 64). Send notification packet J onto the ring. That is, first, a notification is circulated to the effect that permission for exclusive use is given only to 64 nodes, which is the maximum number of required logical channel numbers for one node. In that case,
In the header of the address occupation permission notification packet J, a unique address within the ring system (11111011) indicating that it is an address occupation permission notification packet is written.

(b) The node 34 sends the address occupancy permission notification packet J
When the packet J arrives, the packet is recognized as the address occupancy permission notification packet J by referring to the address, and the packet is received. The node 34 uses only the number of logical channel numbers written in the information part of the packet (7) 7 address space (00000000
~00111111) (the number of logical channel numbers is 64), and the logical node address of its own node 34 is set to (00).

Thereafter, the node 34 sends out onto the ring an address occupancy permission notification packet k having the remaining address space excluding the occupied address space area and the occupancy permission number=64 as information. In this case, the header of the address occupancy permission notification packet includes a unique address within the ring system (111
11011).

(c) When the address occupancy permission notification packet k sent from the node 34 arrives, the node 33 refers to the address of the packet, recognizes that the packet is an address occupancy permission notification packet, and discards the packet. Receive. The node 33 recognizes from the occupancy permission number written in the information part of the packet that the node does not have the occupancy right (the number of required logical channel numbers = 32, which is less than 64), and occupies the address space. No processing is performed, and the address occupancy permission notification packet k is sent onto the ring as it is without changing the packet address and information portion.

(iv) Similarly to node 33, node 32 recognizes that it does not have an exclusive right from the exclusive permission number in the address exclusive permission notification packet that has arrived, and sends it out onto the ring as is without changing the information.

(E) When the address occupancy permission notification packet k that has circulated once arrives, the node 31 determines from the address in that packet that the address occupancy permission notification packet with occupancy permission number = 64 has been processed by all nodes. ! ! I and receive the packet.

After this, the node 31 recognizes that it does not have the address exclusive right either, and then has the unused address space written in the address exclusive permission notification packet k and the exclusive permission right number = 32 as information. An address occupancy permission notification packet m is sent onto the ring.

(to) The node 34 sends the address occupancy permission notification packet m
When packet m arrives, refer to the address of the arriving packet m,
The node 34 recognizes that the packet is an address occupation permission notification packet, but determines that its own node 34 has already occupied the address.

As a result, the node 34 sends the address occupancy permission notification packet m without changing the packet address and information part.
Send it out onto the ring.

(g) The node 33 sends the address occupancy permission notification packet m
When the packet arrives, by referring to the address of the arriving packet, it recognizes that the packet is an address occupancy permission notification packet, and receives the packet. Then, the node 33 recognizes that it has the exclusive right from the exclusive right number written in the information part of the packet, and has an address space (01000000) that satisfies the required number of logical channel numbers. ~0101111
1) Occupies (logical channel number 32) and sets the logical node address of its own node to (OIO). Thereafter, the node 33 sends out onto the ring an address occupancy permission notification packet n having as information the remaining address space excluding the occupied address space and the occupancy permission number=32. In this case, a unique address (11111011) within the ring system indicating that the address occupancy permission notification packet is an address occupancy permission notification packet is written in the header of the address occupancy permission notification packet n.

(H) The node 32 sends an address occupancy permission notification packet n
When the packet arrives, by referring to the address of the arriving packet, it recognizes that the packet is an address occupancy permission notification packet, and receives the packet. The node 32 recognizes that it has the exclusive right from the exclusive right number written in the information part of the packet, and creates an address space (01100000 to 01111111) that satisfies the required number of logical channel numbers. )(
It occupies the logical channel number 32), and the logical node address of its own node 32 is (011). Thereafter, the node 32 sends out onto the ring an address occupancy permission notification packet p having the remaining address space excluding the occupied address space and the occupancy permission number=32 as information. In this case, the header of the address occupancy permission notification packet p contains a unique address within the ring system (11111011) indicating that it is an address occupancy permission notification packet.

(S) The node 31 receives the address occupancy permission notification packet p.
When p arrives, it is recognized from the address of the arriving packet p that the address occupancy permission notification packet with the occupancy permission right number=32 has been processed by all nodes, and the packet is received.

After that, when the node 31 recognizes that its own node does not have the address occupation permission right, the node 31 grants the address occupation permission having the unused address space described in the address occupation permission notification packet p and the occupation permission right number = 16 as information. Send notification packet q onto the ring.

(J) Since the nodes 32, 33, and 34 have already completed the address occupation processing, the address occupation permission notification packet q is relayed as is from these nodes and arrives at the node 31 in that state.

(l) The node 31 sends the address occupancy permission notification packet q
When the packet q arrives, by referring to the address of the arriving packet q, it is recognized that the address occupancy permission notification packet with the occupancy permission right number=16 has been processed by all nodes, and the packet q is received. The node 31 is
It recognizes that its own node 31 has the exclusive right from the exclusive right number written in the information part of the packet, and creates an address space (
10000000 to 10001111) (the number of logical channel numbers is 16), and the logical node address of its own node 31 is (1000).

(W) When the above procedure is completed and address space allocation is completed, normal communication between nodes begins. Of the occupied address space, each node 31 to 34 uses the number of bits corresponding to the required number of logical channels as a logical channel number area, and uses the remaining area as the logical node address of each node.

Referring to FIG. 4(d), a specific method of calculating a logical node address at each node when using the method of this embodiment will be described.

First, a common operation flow for each node will be explained, and then FIG. 4(d) will be explained. Here, an address space of M bits (8) is assumed, and the number of logical channel numbers required by each node is 2 to the N power. In the user information area of the address occupancy permission notification packet, an M-bit head address designation field indicating an unused address space and a permission number length field indicating the length of the logical node address currently being queried are prepared.

(i) An arbitrary node (here, node 31) sets the head address indication field to all O's and the permission number length field to M (here, 8).

(ii) Determine N from the number of logical channel numbers required for the own node, and (1) If the permission number length field=N, use the upper M−N bits of the start address indication field as the logical node address of the own node. Further, the start address designation field + (2 to the N power) is set as a new start address designation field. ■Permit number length field≠N
If so, leave each field value as is.

(ni) Transfer the address occupancy permission notification packet to the adjacent node.

(iv) When the address occupancy permission notification packet arrives,
If that node is the node that first sent out the unused address space notification packet, it decrements the address length field by l and returns to (u) above. ■Furthermore,
If the address length field is already 0, then
Address assignment is considered complete. ■If it is another node, return to the above (in) in the same state.

FIG. 4(d) explains the arithmetic processing in FIGS. 4(b) and 4(c). First, the node 31 sends out an address occupancy permission notification packet with the leading address field = oooooooo and the permission number length field = 8, and determines the required number of channel numbers for its own node to be 4 (41).

Since there is no node that requires the same number of channel numbers as the permission number length field, after going around the transmission path twice, the node 31 enters the head address indication field = oooooooo
○ and the permission number length field = 6, and sends out an address occupancy permission notification packet J (42). Next, node 3
4, since N=6, the grant number length field=N
The upper M-N bits (8-6=2 bits) of the first address indication field are set as the logical node address of the own node 34 (00). Then, the start address indication field + (2 to the N power) is set as a new start address indication field (01000000) (43). The address occupancy permission notification packet passes through nodes 32 and 33 and returns to node 31. The node 31 sets the head address indication field=01oooooo○ and the permission number length field=5 and sends out an address occupancy permission notification packet (44). Since the occupancy process has already been completed for the node 34, the signal passes through the node 34 and reaches the node 33. node 33
Since N=5, the logical node address is set to 010, and the start address indication field + (2 of N
(01100000) (45). Next, in node 32, since N=5, in the same way, set the logical node address to -011 and set a new start address indication field (10000000) (46)
. When the address occupancy permission notification packet is transferred to the adjacent node 31, the node 31 sets the first address indication field = 1.
0oooooo and an address occupancy permission notification packet with the permission number length field set to 4 is sent (47). Since the nodes 34, 33, and 32 have already completed the occupancy process, the process passes through these nodes and returns to the node 31 again.

In the node 31, since N=4, the logical node address is set to 1000, and the start address indication field is set to 10010000 (48).

Note that in the above description of the communication procedure and calculation example, the logical node addresses are determined sequentially from the longest logical node addresses, but even if the logical node addresses are determined from the shortest logical node addresses, the same procedure can be used.

In this embodiment as well, in order to deal with code errors on the transmission link, the address management node (referring to the node that first sends out the address occupancy permission notification packet) continuously sends the same address occupancy permission notification packet. Multiple nodes are sent, and each node continuously repeats the same process multiple times. The address management node inspects a plurality of circulated address occupancy permission notification packets and determines whether the granting procedure is correct or not. If granting fails, an address occupancy permission notification packet with the same permission number length is sent again.

Furthermore, in both the second and third embodiments, when adding a node, the logical node address is determined by individually inquiring the address management node. In this case, there is no need to reassign logical node addresses to all nodes.

FIGS. 5(a), 5(b), and 5(c) are configuration diagrams of other transmission links applicable to the present invention, respectively.

The present invention is as shown in FIGS. 2(a), 3(a) and 4(a).
As shown in Fig. 5(a), the channel identifier communication method has been described as being applied to a transmission line in which multiple nodes are connected in a ring shape and packets are transmitted in one direction.
), it can also be applied to multiple ring-shaped transmission links in different directions. Figure 5(a) shows a method in which multiple ring-shaped transmission links in different directions are used and each link is used independently, and Figure 5(b) shows a method in which two transmission paths are overlapped for all nodes. FIG. 5(c) is a diagram of a virtual transmission path for assigning addresses so that addresses are not lost; FIG. It is a method.

In FIG. 5(a), a transmission path 52a connects nodes 5]a to 51n in a link and transmits packets in this order.
This shows a case in which two transmission paths are provided, a transmission path 52b which connects the nodes 51a to 51n in a link and transmits packets in the reverse order. 1 for this system
, when applying the first embodiment, each transmission line 52
A specific node (for example, node 51a) regularly assigns an address to all address space request packets of nodes 51a and 52b.

In that case, as shown in FIG. 5(b), a return transmission path 53 for assigning an address to a node is virtually considered using the two transmission paths 52a and 52b shown in FIG. 5(a). . That is, the transmission line 53 is the transmission line 52a and 52b.
An independent transmission path is formed by using both. Therefore, the node 51a transmits the packet received via the transmission path 52a to the transmission path 52a.
It is relayed by sending it to 2b. Furthermore, packets received in the opposite direction via the transmission path 52b are sent to the node 51a.
The signal is transmitted through the transmission path 52a.

When the second and third embodiments are applied to this transmission path 53, each node 51a to 51n is connected to the transmission path 53.
3 as virtually one transmission path. As a result, when assigning logical node addresses, duplicate addresses are not assigned.

In FIG. 5(a), during normal operation, two transmission lines 52
a and 52b operate independently, but in the event of a failure, the transmission lines are combined between two nodes across the failure point, and 1
operates as one transmission line (loopback). In that case, if there are two transmission lines 52a.

If a node address is arbitrarily assigned in 52b, there is a possibility that the target node will not receive the signal if a loopback state occurs, so addresses must be assigned so that there is no duplication between the two transmission paths. For this purpose, as shown in FIG. 5(b), two transmission lines 52a and 52b are virtually regarded as one transmission line 53 and an address is assigned to them.

FIG. 5(c) shows the node 51a in FIG. 5(a).
This figure shows a loop when a transmission path failure occurs in either or both of the transmission paths 52a and 52b between the transmission paths 52a and 51b. In this case, in the failure loop, the packets received by the nodes 51a and 51b pass from one transmission path 52a (or 52b) to the other transmission path 52b (or 52a), thereby completing the loop without passing through the failure point. can be formed. Although not shown, even if a node has a failure, it is possible to form a loop by passing one transmission line from one transmission line to the other at nodes on both sides of the node without passing through the failed node. I can do it.

For example, if node 51b has a failure, node 51a
and 51c, by sending the packet received on the transmission path 52a (or 52b) to the transmission path 52b (or 52a), it becomes possible to transmit the packet without passing through the faulty node 51b. In the meantime, you can recover from the problem.

As described above, the first embodiment, that is, the method in which the address management node allocates all address spaces and notifies each node in turn, is suitable when each node has already been assigned a physical address, and each It is possible to assign addresses to nodes individually. In addition, in the second embodiment, which is a method in which the address management node exclusively allocates the necessary address space to its own node in order and notifies neighboring nodes of only the remaining unused address space, a physical address is assigned to each node. suitable if not
It is possible to allocate the necessary address space to each node by simply sending the unused address space notification packet and the address distribution completion packet once each. Furthermore, in the third embodiment, the address management node allocates address space in order from the node with the largest required address space to the node with the smallest required address space. It is possible to assign addresses by effectively using the prohibited space.

〔Effect of the invention〕

As explained above, according to the present invention, variable-length logical channel numbers are used for packet addresses, so each node connected to a ring-shaped or bus-shaped transmission link The number of logical channel numbers can be secured, and the number of unused logical channel numbers can be reduced, so the address area can be used efficiently for packets.

It is possible to do so.

[Brief explanation of drawings]

Figure 1 is a configuration diagram of a packet according to the present invention, Figure 2 is a ring system configuration diagram showing a first embodiment of the present invention, an operation time chart, and a diagram of the entire address space in the system. Figure 3 is a diagram of the present invention. A ring system configuration diagram, an operation time chart, a total address space diagram in the ring system, and a logical node address configuration diagram showing a second embodiment of the invention,
FIG. 4 is a block diagram of a ring system showing a third embodiment of the present invention, an operation time chart, a diagram of all address spaces in the ring system, and a diagram of an example of calculation of logical node addresses;
FIG. 5 is a transmission line configuration diagram showing another embodiment of the present invention. 11-14.21-24.31-34: Node, 51
a, b, n: Node, 52a, b, n: Transmission path, d Near address space notification packet, exh: Unused address space notification packet, l Near address distribution completion packet, jp
Near address occupancy permission notice No.■ Figure Address space logic diagram (Part 2) el Address space logic Logical node Atsys Channel number 0000000C Assigned to node 21 [ Disabled space [ Assigned to node 22 ] Assigned to node 23 [ Disabled space [ Assigned to node 24 [ To 00001111 0001000C 00011111 0010000( 00111111 0100000t 0101111 0110000[ 0111111 1000000[ 1n11111 Figure 4 (Part 1) 2 Figure (Part 2) (C Address space logic Figure 5 (Part 1) (a) Figure (Part 2) (b )

Claims (5)

[Claims] (1) Multiple nodes each accommodating multiple terminals,
A fixed length consisting of the logical node address of a node connected to a ring-shaped or bus-shaped transmission link and accommodating a user information area and a transfer destination terminal, and a logical channel number for mutual identification of multiple communications at the transfer destination node. In a communication system that performs communication between the terminals by transferring a packet including an address area onto the transmission link, the boundary between the logical node address and the logical channel number is made variable for each node, and the logical node If the address is located in the upper digits of the above fixed length address area,
The short logical node address is assigned so that it does not overlap with the high-order digits of the long logical node address, and if the logical node address is located in the low-order digit of the fixed-length address area, the short logical node address is assigned the long logical node address. A channel identifier communication method characterized in that the channel identifier is assigned so as not to overlap with the lower digits of the channel identifier. (2) Any one node in the above communication system manages the entire address space as an address management node, and individually requests the number of logical channel numbers required by all other nodes from the address management node. By doing so,
2. The channel identifier communication method according to claim 1, wherein the address management node distributes a logical node address and an available logical channel number area to each requesting node, and notifies the node. (3) Any one node in the above communication system manages the entire address space as an address management node, and first occupies its own logical node address and available logical channel number area, and unoccupies unused areas. The adjacent node notifies the adjacent node of the address area, and the adjacent node occupies its own logical node address and available logical channel number area from the unused address area, and notifies the adjacent node of the unused address area. 2. The channel identifier communication system according to claim 1, wherein all nodes occupy their respective logical node addresses and usable logical channel number areas by sequentially performing the communication. (4) Any one node within the above communication system manages all address spaces as an address management node, and displays unused address spaces within the communication system and specific logical channel number lengths (hereinafter referred to as permission number lengths). When the adjacent node receives the packet and the permission number length written in the packet matches the logical channel number length required by its own node,
By occupying the logical node address and usable logical channel number area from the address space, and when they do not match, the received packet is transferred as is to the adjacent node without performing the occupation process, and the packet is transferred multiple times. 2. The channel identifier communication system according to claim 1, wherein all nodes occupy their respective logical node addresses and usable logical channel number areas by circulating the nodes. (5) The above communication system uses a method in which multiple nodes independently use multiple ring-shaped transmission links in different directions, and the above transmission link configuration is used, and two nodes have two transmission lines with different transmission directions. 5. The channel identifier communication method according to claim 2, wherein a different logical node address is assigned to each node corresponding to a transmission path using a method of looping back and using a different logical node address.