TW400473B - Triplet architecture in a multi-port bridge for a local area network including method and apparatus for periodically updating entries in a content addressable memory - Google Patents

Abstract

A multi-port bridge includes a memory and a plurality of ports. Each port includes a receive buffer, a transmit buffer and a ""triplet"" buffer. As a data packet is being received by the receive buffer of a port, a look-up table is utilized to identify the appropriate destination port for the packet. A result of the look-up is a ""triplet"" which includes: a first field containing the identification of the source port, a second field containing the identification of the destination port, and a third field containing a starting address assigned to the incoming packet in the memory. The triplet is placed upon the communication bus a first time. If the destination port is available, the destination port receives the packet simultaneously as the packet is stored in the memory. Otherwise, the triplet is placed on the communication bus a second time after the packet is stored in the memory. The destination port stores the triplet in its triplet buffer. Then, when the destination port is available, the destination port retrieves the packet from the memory for transmission. For packets received during a first period, the number of the port receiving the packet is stored with the address of the node which generated the packet in a first memory and in a second memory. During the first period, the first memory is utilized for directing packets through the bridge. At the end of the first period, a third memory is cleared. For packets received during a second period, the number of the port receiving the packet is stored with the address of the node which generated the packet in the second memory and in the third memory. During the second period, the second memory is utilized for directing packets through the bridge. At the end of the second period, the first memory is cleared. For packets received during a third period, the number of the port receiving the packet is stored with the address of the node which generated the packet in the first memory and in the third memory. During the third period, the third memory is utilized for directing packets through the bridge. At the end of the third period, the second memory is cleared. This sequence is repeated such that no entry more than two periods old is utilized for directing a packet through the multi-port bridge, however, the age of each entry is not tracked.

Description

Printed by the Central Government Bureau of the Ministry of Economic Affairs and Consumers' Cooperative Cooperatives " A7 ____B7 V. Description of the Invention (1) Field of Invention: __ This invention relates to a multi-port bridge in a local area network. More specifically, the present invention relates to a process for controlling data message packets in a port of a multi-port bridge. The present invention is also related to the retrieval of a lookup table, which appropriately indicates a data message packet via a multi-port bridge in a local area network. More specifically, the present invention relates to the field of periodically updating the input in a lookup table, which appropriately indicates a data message packet through a multi-port bridge in a local area network, where the lookup table is configured to imitate content addressable memory (CAM) device's random access memory (RA M) device. BACKGROUND OF THE INVENTION: The nodes of a local area network (LAN) are typically interconnected by a common transmission medium. However, shared transmission media can limit the amount of data traffic that can be accommodated. For example, only one node can successfully transmit data to another node on a shared transmission medium at a time. If two or more nodes try to transmit data at the same time, a collision of data occurs, which easily destroys the data being transmitted. Therefore, the nodes sharing the transmission medium are considered to be in the same collision domain. Multi-port bridges are allowed in the local network by segmenting the area network into multiple collision areas (also known as network segments). Simultaneous communication of local area networks between nodes, each segment has a corresponding transmission medium. Figure 1 illustrates a traditional local area network including a multi-port bridge 10. Multi-port bridge 10 has eight ports A — Η, although the number of ports can be changed. Each port A — 片段 is connected to a segment of the local area network 1 1 — 1 8 9 each piece (^ 先 间 ^ 卄 之 卄 念 ^ Item code fill out this page)

This paper size applies to China National Standards (CNS) Λ 仙 () A7 137 V. Description of the invention (2) Paragraph 1 1 1 1 8 typically includes one or more nodes 1 9 1 3 4 such as workstations, personal computers' Data terminals, file servers, printers, facsimiles, scanners or other traditional digital devices. Each node 1 9-3 4 has a related node that publishes its uniquely identified node. Those nodes 19-3 4 are configured to send data to each other in the form of discontinuous data message packets. Printed by the Industrial and Commercial Corporation, when the local area network operates according to the Ethernet standard, such as the Institute of Electrical and Electronics Engineering CIEEE 802.3 standard, the data is communicated in the form of discontinuous message packets. Figure 2 illustrates the traditional IEEE 802. 3 data message packet 40. Those data message packets 40 include the octet header 4 1, which is usually thrown to synchronize those receivers with the data message packet 4 Q. Title 41 1-byte including the title. After heading 41, those data message packets 40 include the six-byte leader's target address 4 2 which is a six-byte leader's target address 4 2, which is a node message packet for those intended recipients of the data 4 0 node address. Then 'the data message packet 4 0 includes the source address of the 6-byte leader 4 3, the source address of the 6-byte leader 4 3 is the node address of the node that sent those data message packets 4 0 = Following the source address 4 3 is a two-byte length region 4 4. After the length area 4 4 is the data area 4 5. The data area 4 5 may be up to 1 500 bytes long. Finally, those data message packets 40 include a byte frame check area 46 that allows a recipient of the data message packet 40 to determine whether an error has occurred during the transmission of the data message packet 40. When a node (source node) sends data to its local area network (-------------, — ___ — _ ς This paper standard is applicable to Chinese countries (CNS > / to Regulations (ilOxm.d'T ~ B7 B7 Step-Ministry Central Standard ii'-Bureau of Industrial and Commercial Cooperatives Cooperative Association V. Invention Description (3) Internal Section Communication) Nodes other than the same segment (target Nodes) when those data are no longer interfered by the multi-port bridge 10 and communicate directly between those nodes and are treated as internal sector message packets. Therefore, when the multi-channel bridge 10 receives the internal sector message Packets, Multiport Bridge 10 does not bridge message packets (message packets are filtered). When a node (source node) sends a data message packet to another node (target node) located on a different segment (inter-segment communication) The multi-port bridge 10 appropriately forwards those data message packets to the target node. The data message packets received by the multi-port bridge 10 are usually temporarily stored in the multi-node before being transferred to the appropriate target node via the multi-port bridge 1G. Buqiao 10 inside .. However, when the multi-port bridge can When the network's requirements are exceeded, the problem arises. When the message packet is at a higher rate than the multi-port bridge 10 can properly forward the packet, the data is received by the multi-port bridge 10 The road becomes congested. This problem worsens as network users gradually increase their demands on the network. "Therefore," improved technology for controlling data flow in a multi-port bridge for a local area network. "Random access memory (RAM) devices Usually contains a plurality of data storage locations, each data storage location has a corresponding predetermined address. Figure 3 illustrates a conventional random access memory device 50. As illustrated in Figure 3, the random access memory device 5 0 has an address area of eight bits each. Therefore, the random access memory device 50 has 2 5 6 unique addresses. Because the address corresponds to the data storage location in a one-to-one correspondence, there are 2 5 6 Location of the memory in the pen data random access memory device 50 0 This paper size is applicable to China National Echelon Race (CNS) See dry (2iOX29h >#) (" See the notes on the back of the reading and then fill out this page) Clothing- ---- I _ I____ * = " /

I 137 " '—. ....... .- I— ....... " ------ I — V. Description of Invention (4) (District). However, the random access memory device 50 can have other address area lengths, and therefore, there are different numbers of unique addresses and corresponding memory locations. The random access memory device 50 includes an address port 52, a data port 54, and a read / write control input 56. In order to store data in a data store p of the random access memory device 50, the address of the place where those data are to be stored is used for the address 璋 5 2 and those data stored are stored and used for those data ports. 5 4. Then, the read / write control input 56 is conditional on writing those materials within the position described by the address. In order to read data from the random access memory device 50, the read / book write control input 56 is conditional on reading the data, and the address of the location where those data will be read is sent to the address port 52. The data stored in the location specified by the address appears on data port 54. Content addressable memory (C AM) devices usually include most data storage locations, and each data storage location has a predetermined corresponding address. The characteristic of the C AM device is that each data storage location includes a compared character area and a related data area. The compared characters are stored in those compared characters, and the data related to each compared character is stored in the related data area. Data from those related data fields are read by comparing each character being compared with a given frame. When a comparison occurs, the contents of the relevant data area corresponding to the character being compared are read. Fig. 4 illustrates a conventional CAM device 60. »As illustrated in Fig. 4, the CAM device 60 is an address area of eight octets in length. Therefore, the CAM device 60 has 256 unique addresses. Because the address is applicable to the Chinese paper standard (CNS) / \ 4 圯 (2 丨 0 > < 2 people's exchange) for the paper size of the information r—-Ί:-.—-- l «l ----- --—-!--I ^ if fit 1 ~ ¥ (Read 1 and read the notes and fill in this page again) A7 Η 7 V. Description of the invention (5) The one-to-one correspondence of the storage location In the AM device 60, there are storage locations of 256 pieces of data, and each piece of data includes the storage locations of the compared character area and the related data area. However, the CAM device 60 can have other address area lengths, and therefore, the number of unique addresses and corresponding memory locations is different.

-P CAM device 60 includes address port 62, compared character port 6 4, related data port 6 6, a read / write control input 6 8 and a comparison output 6 9. In order to store the compared characters and related data in the CAM device 60, the required 'address' is sent to the address port 62, the desired compared characters are sent to the compared character port 64, and those required related data are Send to those related data port 6 6. Then, the control input 6 8 conditionally pairs the characters to be compared. And those related data are written in the position described by the address. The Central Bureau of Standards and Technology of the Ministry of Shanghai and the People ’s Republic of China consumed the alpha-bismuth to read the relevant data from the CA device 60 and control it. The conditional reading was entered, and a designated frame was sent to the compared character port 64. If the stored compared character matches the specified character of the compared character port 6 4 ', this situation will be indicated by the comparison output 6 9. Relevant data stored in those characters that correspond to the compared characters appear in the related data area port 6 6. Otherwise 'If no characters are compared with the characters sent to the compared characters port 6 4, the comparison output 6 9 will "This will be indicated" Therefore, 'for the relevant data read from the CAM device 60, it is sent to the character port 64 of the compared character and the content previously stored in all the compared character regions. When the comparison is found, the relevant data appears in data port 66. Therefore, the Billion Device 60 is called, --------------. R -__ This paper size applies to the Chinese country # 双 (CNS) Λ4 圯 拈 (2 I 0〆297. f ~) A7 B? 5. Description of the invention (6) Because of the "addressable" feature, CAM devices are typically more expensive to produce and purchase than random access memory devices. However, CAM devices are special Suitable for an application. For example, the CAM device can be used to construct a look-up table to properly use the multi-port bridge instruction data message packet in a local area network (LAN). In particular, the multi-port bridge 10 (Figure 1) Each data message packet 40 is received (Figure 2) and it must be determined whether those data message packets 40 are for internal segment communication or inter-segment communication, and if those data message packets 40 are for inter-segment communication, more Port bridge 10 must establish 40 based on the target address 4 2 contained in those data message packets 40 to determine which ports (target ports) the data message packets 40 will be directed to. This can be done by a query table built in the CAM device 60 (Fig. 4). As usual, the lookup table performs a query phase by performing a learning phase for each received data message packet 40 to store data in the table and by performing a query phase for each received data message packet 40 message packet. To query the data stored in the learning phase of a previous message packet. Jinglai—Ministry History 1? · Bureau Kigong Consumption Hezhudu Printing An example is provided to enhance the understanding of the learning stage. To understand the examples, reference should be made to Figures 1, 2 and 4. In the learning phase, the address of the node (source node) that sends the data message packet 40 is stored in the address area of the CAM device 60 and the number and storage of the port (source port) that will receive the data message packet 40 Executed in the corresponding data area. For example, referring to Figure 1, if node 19 sends a data message packet 40 to node 30 0 ', the data message packet 40 will include the address of node 19 as its source address. Duobuqiao 10 will receive those data from the port record. Applicable to China National Transfer {CNS) (2ΐ〇 × 297,:. Γ) Γ5 The central standard of the Ministry of Distribution 4, the Bureau only eliminates the seal of the Bamboo Society. Install A7 --------_ _B7 V. Description of the invention (7) '~~' 'A. Message packet 4 〇. Therefore, for the learning phase, the address of 'node 丄 9.' Will be stored in the address area of the C.AM device 60 and the number of identification port A will be stored in the corresponding data area. This procedure is repeated for each data message packet 40 received by the multi-port bridge 10. Once the look-up table is constructed, the destination addresses contained in those data message packets are used to properly guide the next data message packet 4 Q 'to the CAM-focused address 垾 6 4 and those of the data port CAM 6 6 The port (target port) to which the message packet will be directed will be indicated. For example, 'node 30+ is the intended recipient. Therefore, those data message packets will include node 30 as its target address. Node 30 is connected to port F °. Therefore, when the address of node 30 is sent to the address 64 of CAM device 60, '6 of those data port CAM device 60 will indicate those data messages. Packet 4 should be directed to port F. Data message packets received before the table is completed 40 are played to all ports A-Η to determine that the appropriate target node received the message packet. In order to reduce costs, the C AM device 60 can be simulated by a random access memory device by known hybrid techniques. Promiscuousness is a mathematical technique that maps into a larger group of smaller groups. For example, in an Ethernet LAN, each node is designated as a six-byte (48-bit) long address. Therefore, there are 248 (about 2.8x1014) possible addresses. Making the random access memory device to store the number of ports associated with each of these possible addresses would require a random access memory device with 248 billion locations. Constructing such a large random access memory device will tend to be impractical and will almost certainly exceed a suitable ______________-This paper size applies to China National Pillow Standard (CNS) Λ4ί.1 ^ δ (}

I Λ7 B7 '; 1 ~ I-.---- V. Description of the invention (8) Cost and space requirements of a local C A M device. Because Ethernet LANs usually have far fewer than 248 nodes, promiscuous can be used to build lookup tables with far fewer than 248 inputs.

-P According to the IEEE 802.3 standard, it correctly reflects the need for a lookup table for the nodes newly added to or moved from the local area network. For this purpose, each node of the local area network periodically transmits a live, message packet, which is sent every 2.5 minutes. It is preferably five minutes during a predetermined time period, and entries in the lookup table that have not been updated are moved from the lookup table. Order-to-Trend: The known techniques used by the Ministry of Economic Affairs and the Jujube Bureau to work with consumer cooperatives to update this look-up table in the Ethernet LAN include the use of a data message packet whenever a corresponding node sends a data message packet. A time stamp is stored in the lookup table along with each registration and update time stamp. In addition, this technique requires keeping each login in the track lookup table at the same age, and removing the login when it becomes more than five minutes old. This technique has the disadvantage of providing time stamps for each login in the table, updating those time stamps, tracking the elapsed time since the last update of each login, and removing logins more than five minutes after the update of each login, It takes a lot of processing overhead. At the same time, additional storage space is needed to store those time stamps. Therefore, in order to periodically update the registration in the CAM device, improved technology is required. What is more needed is an improved technique for periodically updating registrations in a CAM device that is accessed by a random access memory device. Further, the registered data in the used look-up table is updated periodically to properly use the multi-port bridge instruction data in the Ethernet local area network to improve the technology. ------ Ί Ί „This paper size is applicable to China National Standards (CNS) 梠 (210X297 公) Γ_ __157 V. Description of the invention (9) Description of the invention · 'The invention is in the port of the multi-port bridge Method and device for controlling data message packet flow control. Multi-port bridge includes switch engine, a dynamic random access memory and multiple ports, all of which are connected to each other by high-speed communication bus. Switch engine includes memory controller and bus controller. Each is a limited-state machine + a bus controller by collecting requests from those ports in accordance with the appropriate priority and allowing those requests to control access to the message bus. The Kee controller provides the memory and communication bus The interface between the rows. | Billion includes a lookup table that is used to properly indicate data message packets in those ports. Message packet buffering is used temporarily to provide an interface between the on-off engine and an external processor. Stores message packets and postboxes. Each port includes a port controller, a MA C radio transceiver, a receive buffer, a transmit buffer, and a triplet buffer. The message packet received by the LAN segment of the line transceiver is instructed to the communication bus via the receiving buffer, and the segment message packet to be transmitted through the LAN segment is instructed to the radio transceiver by the transmission buffer. The tuple buffer storage memory index, referred to herein as triple A V, is for the message packets being memorized to buffer the stored message packets. Data from nodes (source nodes) in a segment of the local network The message packet is received by a corresponding buffer of the port (source port) of the multi-port bridge. While the message packet is still being received, those look-up tables are used to determine which destination is appropriate based on the destination address Fill in the message packet. The result of the operation query is "* triple", which includes three areas: the first area contains the confirmation of the source port, the second area contains the confirmation of the destination port, and the third paper size applies to the Chinese country ; iH- (CNS) / \ 4 圯 彳 ft (2ΓΟΧ297 公 12 ~ {Note to ^^^ the back of the note ^ matters-^^ 艿 this page)

、 1T Ι.Ϊ7 Ι. 经 7 After the Ministry of Foreign Affairs " · The Bureau only eliminates the combination of bamboo and printed bismuth. V. The description of the invention (10) The area contains reception allocated to the message packet buffer in the memory. The originating address of the message packet "If the source and destination ports are the same, it indicates that the source and destination nodes are on the same segment of the local area network (internal segment communication) * and the message packet is filtered. Otherwise, the Jiyi controller puts the triplet on the communication bus that is monitored by each port. If the destination port in the triple is not considered to be currently busy transmitting or receiving another message packet, the destination port will configure itself to receive the message packet (cut through) from the source port. However, if the triple buffer in the target port considered to be in the triple is almost full, the bus controller of the target port sends a medical congestion request signal to the communication bus. The source port will receive the congestion request and, in contrast, it will discard the received message packet and also send a congestion signal to its relevant segment. The congestion signal will cause the node (source node) that is the source of the message packet to stop transmitting the message packet and attempt to resend the message packet after a waiting period. Once the triplet has been formed, the message packet is loaded from the receive buffer of the source port into the starting address message buffer of the memory identified by the third area of the triplet. Writing a message packet to a buffer The message packet should preferably occur while the remainder of the message packet is still being collected into the receiving buffer of the source port. For this reason, the receiving buffer of each port does not need to be able to store the entire data message packet. In addition, if the target port is configured to cut through, the target port will write the message packet to the message packet buffer. The input cycle loads the message packets in the message packet buffer into its transmission buffer directly and synchronously from the communication bus. Do everything here ____________ _ ή 〇 _____ This paper size is applicable for ¥ 国 ® 家 彳:?: 界 (CNS): 4¾¾ (2ΙΟΧ 297 ^ > 1 ™ ^ (^^ Read the back note ^ ^ Item ^^ 巧 本 ^)

* 1T class A7 H7 V. Description of the invention (11) In operation, the message packet enters the transmission buffer in the target port for immediate transmission to the local network segment related to the target port. Once the entire message packet has been loaded into those message packet buffers, the triplet is again placed on the data line of the communication bus. The target port + then stores the triples in the triple buffer. So 'message packets are stored for transmission by the destination port. Then ’when the destination port is no longer busy, the destination port retrieves the message packets from those message buffers.

• IT When the destination port is receiving message packets from the message packet buffer or directly from the source port into its transmit buffer, the destination port starts sending message packets to the LAN segment associated with the destination port. For this reason, the transmit buffer of each port need not be able to store the entire data message packet. The message packet is then received from the LAN segment by the destination node of the message packet. The limited number of instructions of the bus is used to control the flow of message packets through the multi-port bridge. Compared with the previous system, the production of the multi-port bridge is improved. The present invention includes a method and a device for periodically updating a look-up table input, which is utilized by appropriately using a local area network (area) Network) Multi-Port Bridge indicates data message packets and to ensure that no older than predetermined time input is used to indicate data via Multi-Port Bridge. Each data message packet is sent by one of the individual nodes of the local network and contains the source address 'which is the address of the node message packet that sent those data, and the destination address' which is the data message that you want to receive The address of the packet's node. One or more nodes are contained in each segment of the local area network, and each segment is coupled to a corresponding port of the multi-port bridge. Each node may be a workstation ’personal computer, data terminal, file server, printer,

------- ----- 一 Λ A • The paper scale is applicable to China's National Standard (CNS) Λ4 (2Ι0χ ~ 29Τλ " j ~ V. Description of the invention (彳 2) Scanner, fax or other Traditional digital device. ^ Nodes are identified by a unique node address, and each port is identified by a unique number of ports. {^ &Quot; When a node (source node) sends a data message packet to another node (target node) in the local network, the data message packet is received by one port (source port) of the multi-port bridge. A lookup table is used to identify the location of the target node. The query table is operated for each data message packet for a learning period * where the message is received, and the number of the port (source port) of the packet is related to the source address contained in the data message packet and is stored in the 9-port bridge. The query period contained in the destination address of the other packet (target port) identified by the message packet is used to transmit the data message packet to the node receiving the data message packet. The multi-port bridge then concentrates those data message packets on the indicated (target port). The line lookup table contains the first 100 million layers, the second 100 million layers, and the third 100 million layers. Each memory layer is either a C AM device or a random access memory device emulating a C AM device. A learning period is performed for each data message packet received during a first period of time. During the study period, the source confirmation number is stored in relation to the address of the source node in the first memory storage layer and in the second billion layer. This allows nodes and port addresses to quickly correlate with each other. At the same time, in the query period of each data message packet received in the first period, the input in the first memory layer is used to indicate the data via the multi-port bridge (determine which is the appropriate port for the message packet). At the end of the first period, the third memory layer erases all inputs. During a learning period performed for each data message packet received during a second period of time, the number of source port confirmations and the number of records in the second record ______________- -jF; ___________ This paper standard applies to China橾 2?-(CNS > Λ4 ^ 彳 Μ 2 丨 0〆.) A7 A7 Printed by the Central Government 栉 · ββconsumer Hezhusha ¾ B7 ~ ----— —w_ _ "-Five, Description of the invention (13) The storage layer is stored in association with the address of the source node in the third memory layer. At the same time, in the query period of each data message packet received in the first period, the input in the second memory layer Used as instructional information via a multi-port bridge. At the end of the first period, the second memory layer erases all inputs. During a learning period performed for each data message packet received during a third period, the number of source port confirmations is related to the address of the source node in the first memory storage layer and the third memory layer Of being stored. At the same time, during the query period of each data message packet received in the third period, the input in the third memory layer is used as the data for indicating via the multi-port bridge. At the end of the third period, the second memory layer erases all inputs. This sequence is repeated from the first period as long as the multi-port bridge is indicating a message packet. Therefore, no input more than two cycles before is used to indicate the data message packet via the multi-port bridge. Preferably, the first, second and third periods are each 2.5 minutes. Therefore, no input more than 5 minutes ago was utilized to indicate the data message packet through the multi-port bridge. Unlike previous systems, the age of each entry in the lookup table does not need to be tracked individually. Brief description of the diagram: Figure 1 illustrates a traditional local area network (LAN) including a multi-port bridge 10. Figure 2 illustrates a conventional I Ε Ε Ε 8 0 2 .3 data message packet. FIG. 3 illustrates a conventional random access memory (RAM) device. ______________________ _ 1 R · This paper size is applicable to China National Pillow (CNS) Λ4 圯 怙 (2! Ο κ 2V7 公 # T '--- //' prr --- 1 ---------- f ——Order ----- 0 (1 /, 1 said ii. * t item and then MK-this page) ____; ___ in V. Description of the invention (14) Figure 4 illustrates the traditional content can be Address memory (CAM) device. Figure 5 illustrates a block diagram of a multi-port bridge according to the present invention. Figure 6 illustrates a port of a multi-port bridge according to the present invention. Figure 7 illustrates a ternary according to the present invention. The group includes the first area, which contains the confirmation of the source port, the second area, which contains the confirmation of the target, and the third area, which contains the memory address of the hip. Figure 8 illustrates the switching of the engine and the memory according to the present invention. Summary block diagram of transpose and external processor. Figure 9 illustrates a summary block diagram of an apparatus for periodically updating a lookup table with three lookup subtables according to the present invention. Figure 10 illustrates a periodic update according to the present invention. A timing table of the three lookup subtables illustrated in Figure 9. The main component comparison table 10 Multiport Bridge 11-18 Section 40 Data message packet 4 2 Destination address 4 3 Source address 4 4 Area 4 5 Data area 50 Random access memory device 5 2 Address Port The paper music scale is applicable _Home School Bus (CNS) Λ4 (210x? ≪ J -/ λ) 4: 7 (^^ is called the note of reading '5 ^ Matters " " 邛 This page)

、 1T line passed the Shanghai Ministry of Standards 1 ?. Bureau K Industry Consumers Co., Ltd. Zhuzhu Printing Bismuth A7 B7 V. Invention Description (15) 54-66 Data 璋 56 R / W Control Input 60 Content Addressable Memory Device 64 Comparable character port 102 Communication bus 104-124 Functional block 114 Bus component 116 Mailbox component 118 Memory control component 12 0 Query control component 200 Memory device 2 0 2 Mailbox 2 0 4 Lookup table 206 Message packet buffer 3 0 0 Controller 302 Bus controller 3 0 4 Registers 306, 316 FIFO buffers 3 0 8 Intermediate access control transceiver 3 10 LAN section 3 1 2, 3 1 4 State-limiting machine 3 2 0 address latch 318 transmission buffer 500 external processor This paper standard is applicable to the Ten Countries: CMS (2ΙΟΧ297: .Γ; ί) '' ΐΛ, ι ^ ΐήιϊ · And note & affairs ifi section J / ir; T this p ') line A7 H7 V. Description of the invention (16)' ·. · 15.0 Switch engine reference 1 5 0 6 Resources 5 ◦ 2 memory bus 204A-C sub-table preferred embodiment details Description: The following documents are incorporated by reference: U.S. Patent Application No. 08 / 590'125, filed January 23, 1996 Day; U.S. patent application rights apply consecutively No. 0/3/3 1, 4, 9; applications march on January 11, 1995: US patent application No. 08/947, 081 , Filed on October 8, 1997: US Patent Application No. 08 / 946,866, filed on October 8, 1997; US Patent Application No. 09 / 025,356, filed on 198 February 18th: and US Patent Application No. 09 / 025,355 'filed on February 18, 1998. In the preferred embodiment, the present invention is utilized to indicate data message packets appropriately via a multi-port bridge over an Ethernet LAN. However, it is quite plausible that ‘other devices in an Ethernet local area network, such as a switch or a router’, or devices operating in accordance with another network standard on the network, can take advantage of the present invention. FIG. 5 illustrates a schematic block diagram of a multi-port bridge 100 according to the present invention. The high-speed communication bus 1 0 2 provides the functional blocks 1 04-1 24 which interconnect with each other of the multi-port bridge 100. The communication bus 1 02 should include five command lines and thirty-two data lines, although the other paper standards are very applicable to the Chinese national standard (crs) Λϋ 彳 M 1 A7 ____ B7 V. Description of the invention (17) The bus structure can be used. According to a preferred embodiment, 24 ports of 10 megabits per second 1 0 4 — 108 and two ports of 100 megabits per second 1 1 0-1 1 2 are coupled respectively. The communication buses 102 and can be coupled to individual LAN segments, each LAN segment. One or more nodes. The 24 10 megabits / second ports 1 04-108 are each transmitting and receiving data message packets at a rate of 10 megabits / second. However, two 100 megabits / second Second port. 1 1 0 — 1 1 2 Transmit and receive data message packets at a rate of 100 million bits per second. However, it is clear that other port structures and other numbers of ports with other functional characteristics can be used. -The bus control module 1 1 4 controls the access to the communication bus 1 102 by collecting requests from those ports 104-112 and other components. Based on those requests, the bus control module 1 1 4 allows access to the communication bus 1 102 according to the appropriate priority order. The bus control module 1 1 4 also controls the access to the memory device 2 0 0 by an external processor (MPU) 1 as explained in more detail here. An MP U port and mail box assembly 1 16 provides an interface between the multi-port bridge 100 and an external processor 500 (Figure 8) for operating various functions, as will also be used here Explained in more detail. These functions include loading data into the multi-port bridge 100's register | extracting data from the multi-port bridge 100's register and external processor 5 0 0 and port in multi-port bridge 100 1 0 4_ 1 1 2 to transfer data message packets. The Billion Control Module 1 1 8 provides the interface between the memory device 2000 and the communication bus 1 102 and also provides the memory device 200 and the check ____________________- 9Π -___ Zhang scale is applicable to the Chinese national pillow { CNS) Λ4 圯 Estimation (210X297): (" tread: .ϊτ'Back to the Note Ill # Ill # JAgThis page), order-Gao Jingluo Lang miss! Ϊ́ 彳 .'L; fs5i A 7 _ B7 V. Description of the invention (18) Query the interface between the control components 1 2 0. The memory device 200 includes a post box 202 for exchanging messages between the external processor and the multi-port bridge 100. In addition, the memory device includes a lookup table 204. Those lookup tables 204 include inputs indicating which port of the multi-port bridge 100 is associated with each node of the local area network. Those look-up tables 204 are used to properly indicate among those ports 104-112 that the data was received by the multi-port bridge 10, as will be explained in more detail here. The query control component 120 receives the node's address and the relevant port confirmation from the communication bus 1 102. These addresses and acknowledgments are stored in the inquiry form 204. The query control component 120 helps to use those query tables 204 to indicate the message packets in those ports 104-112 according to the destination address of each message packet. The memory device 200 also includes a message packet buffer 206 for temporarily storing a data message packet guided by the multi-port bridge 100 back. The memory device 200 is preferably an S DRAM device, although other types of memory devices can be used, such as DR A M, S R A M, R A M, or E D 0. In a dynamic memory situation, the Billion Control component 118 makes the memory device 200 revisited. A E-static component 1 2 2 collects data message packets and routes statistics and provides their external processors for performance analysis and network management functions. The timing component 1 24 provides timing signals to the ports 104-112 and the other components of the multi-port bridge 100 1 14-122. Preferably, the main clock signal cycle is 4 OMZ. Other clock signals; L0MHz and 25MHz, originated from the main clock signals. Preferably, those components 1 1 4 to 1.2 4 are the same ones that implement each of the same ---------- _________ _ _ This paper standard is applicable to China National Graduation Standard (CNS) Λ4 (2) 〇 × 29 ·? / ΛΤ; 7 ~~~ ^ 先 ^^ Back to Note ^ 事 ^ 再 mo 衿 衿 本 莨) Order _ gland [_ ΙΠ — 丨-* 1 r,-I,. V. Description of invention (19) Limit state machine, although Those components 1 1 4-1 2 4 can be implemented alternately as one or more processors operating in accordance with a stored software program. Because it can usually perform the necessary operations faster, the state-limiting machine is better. Therefore, the capacity of the multi-port bridge 1000 which achieves a higher message packet operation is achieved. Figure 6 illustrates an essential block diagram of one of the ports 104-112 of the multi-port bridge 100. The port controller 300, including the bus controller 300 and the register 300, provides control to the ports and interfaces between the port and the communication bus 102. The port controller 3 0 0 monitor is instructed to the port and when the port is controlled by the communication bus 102, it also provides commands to the communication bus 1 102 for commands and data communication bus 102. The register 3 0 4 contains information for initializing the status and status messages for the collecting port. The port also includes a triplet F I FO buffer 3 0 6 coupled between the communication bus 102 and the port controller 3 ◦ 0. The triplet buffer 3 0 6 stores the memory index of the data message packet stored in the message packet buffer 2 6 (FIG. 5) of the memory device 2 0 (FIG. 5). To explain). Preferably, the triplet buffer 306 should store 1, 28 triplets, each triplet being a four-byte leader. The port also includes an intermediate access control (MAC) electrical transceiver 308, which takes the LAN segment 310 and / or transmits and receives data message packets from / to the LAN segment 310. More relevant and coupled with the radio transceiver 3 0 8 is a receiving limit state machine 3 1 2 which controls the radio transceiver 3 0 8 when the message packet is received, and a transmitting limit state machine 3 1 4 which is the on-message Controls the radio transceiver 3 0 8 during packet transmission. Receiving limited state machine __________________________ — 99-......... This paper 诙 standard is applicable to Chinese country 彳: ϊ: （(CNS) Λ4 圯 佑) ....... --II II 1 — -I- —] —-:: I--(-Yu Jian 1 'and ·; · 恋 事 # fill out this page) Ding- ° ___B7 V. Description of the invention (2〇) 3 1 2 and transmission The state-limiting machines 3 1 4 are respectively coupled to the bus control components 1 1 4 requesting access to the communication bus 1 102 from it (Fig. 5). . Also. The message packet received by the radio transceiver 3 0 8 from the LAN segment 3 1 0 is directed to the communication bus 1 102 by a receiving F i F0 buffer 3 1 6 and will be transmitted across the area The message packet of the network segment 3 1 0 is directed from the communication bus 1 0 2 to the radio transceiver 3 0 8 by transmitting the FI F0 buffer 3 1 8. Preferably, the receiving buffer 3 1 6 holds 1 2 8 bytes and the transmitting buffer 3 1 8 holds 2 56 bytes. Note that the data message packet of I ΕΈ 8 0 2. 3 can contain up to 1 500 bytes of data in addition to the source address, destination address, and frame inspection area. Therefore, in the preferred embodiment, neither the receiving buffer 316 nor the transmitting buffer 318 can store the maximum capacity of the entire IEEE 802.3 data message packet. Address latches 3 2 0 are also included in the port to latch the addresses from the communication bus 10 2 and provide them to the radio transceiver 3 8. r Step-Ministry Central Standards ΐ · 杓 ·, ί3. 合作 Consumer cooperation.ΗPrinting The flow of 1 0 0 message packets over the port bridge occurs in the following manner. A data message packet, such as the IEEE802.3 data message packet, is sent from a node (source node) in a segment of the local area network and is mapped by a corresponding port 104—112 (source port) of the multi-bridge 100 (Figure 5). Roger that. Those data message packets are received in the buffer 3 1 6 of the receiving source port when they are received from the local area network fragment message packet related to the source port by the radio transceiver 3 0 8 in the source port. After the first twelve bytes, the source address and the destination address of the matching message packet were received, the receiving limit state machine 3 1 2 was coupled to the bus control module by lifting. This paper standard applies to the Chinese national standard ( CNS) AWWM 210X2SI7 Central Standards Bureau of the Ministry of Public Economics and Economics, the Central Bureau of Standards, and only the consumer and investment cooperation Du printed equipment A7 B7 V. Description of the invention (21) 1 1 4 Interrupt request line from the bus control circuit 1 1 4 Request query cycle (Figure 5) . The bus control component 1 1 4 monitors this request and allows each request in the proper priority order. When the request is allowed, the bus control module 1 1 4 identifies the allowed bit type of the bus by placing it on the command line of the communication bus 102, and places a unique bit on the data line of the communication bus 102. Identify the bit type of the source port to notify the source port. The first four bytes of the destination address of the message packet are then placed on the data line of the communication bus 1 102 from the receiving buffer 3 1 6 of the source port, and the corresponding bit_type is placed on the source port's bus The command line of the communication bus 1 0 2 is next to the bus controller 3 0 2. The query control component 12 (Figure 5) receives the first four bytes of the target address. Then, the source 璋 places the last two bytes of the destination address of the message packet and the first two bytes of the source address of the message packet on the data line of the communication bus 102 and places the corresponding bit type on Communication bus 10 2 on command line. The query control component 120 receives these four bytes. Finally, the source port places the last four bytes of the source address of the message packet on the data line of the communication bus 102 and the corresponding bit type on the command line. ≫ The query control module 1 2 0 also receives To these four bytes. Therefore, the destination address and the source address are transferred via the communication bus 102 in the four-byte segment, because it matches the width of the data line (32 bits) of the communication bus 102. However, it is quite plausible that the communication bus 102 may have different numbers of data lines, wherein different numbers of bytes can be transferred at one time. Once the query control module 1 2 0 receives the message address, the destination address and the paper size are applicable to the Chinese national car search (CNS) ΛϋΜ, (2Ι0 × 297 public address.) (^ 1¾¾ ^^ 's note of love 44¾ 本 Ν) 1Τ B7 B7 Printed by the Central Bureau of the Ministry of Economic Affairs and Industry only for consumer goods cooperatives. ^ The paper size is in accordance with the Chinese national standard. (CNS) Λ4 /) 1½ {210χ2 (ί7.: Λ} ;. ~) 22 _) source address, querying the control module 120 will notify the Jiyi control module 118 (Figure 5). The Billion Control Module 1 1 8 then updates the look-up table 2 0 4 by determining the source address of the message packet to be stored in the look-up table 2 0 4 related to the source port of the message packet (Figure 5). It is determined that those look-up tables 204 correctly reflect any changes that may have occurred in the local area network (called a learning period). Messages stored during the learning period are used to indicate subsequent message packets. Once the learning period is completed, the Billion Control Module 1 1 8 uses those lookup tables 2 0 4 to determine which port (target port) is related to the destination address of the message packet. Β As a result of the operation query cycle, the memory control module 1 1 8 becomes % Triples' bit type. Figure 7 illustrates a triple that includes three regions: the first region 400 contains the acknowledgement of the source port, the second region 400 contains the confirmation of the destination message packet, and the third region 400 contains the allocated The starting address of the received message packet in the message packet buffer 206 in the 100 million device 200. The first area 400 and the second area 402 should be one byte leader, respectively, and the third area 404 should be two byte leaders. However, it is clear that the order of the regions of the triples and the size of each region can be changed. If the source port and the destination port identified by the triplet are the same, it indicates that the source and destination nodes are on the same segment of the local area network (intra-segment communication) ', and therefore the message packets need not be bridged. In this case, no further action is taken on the message packet (the message packet is filtered). Otherwise, the 'memory control component 1 1 8 will place the triplet on the data line of the communication bus 1 102 and will indicate one. The initial triplet 〃 is ready ------------ -(1 /) Please read the detailed back note * -4 'Item 4 to fill out this page)

'' —II- | -------- _ '' --II- | -------- _ The central part of the Ministry of Standards must be burned only to eliminate f Cooperative printed plough A7 B7 V. Description of the invention (23) The bit pattern of the good state is on the command line. Each port monitors the communication bus 102. If the port identified as the destination port in the triple is not busy transmitting or receiving all other message packets, the destination port will configure itself to receive (cut through) message packets directly from the source port. Alternatively, if the triple buffer 3 0 6 'in the port identified as the target port in the triple is almost full, the bus controller 3 2 of the target port sends a congestion request signal to high The command line of the bus 102 of the "source port" will receive the congestion request and, accordingly, the received message packet will be discarded and a congestion message packet will also be sent over the relevant segment. Congestion of the message packet will cause the source node (source node) of the message packet to stop sending the message packet and try to resend the message packet after a waiting period. As illustrated in FIG. 7, the triples should have a uniform size β. Therefore, the exact number of triples that can be accommodated by a triple buffer 306 in the port can be obtained from the triple buffer in 璋. The number of 3 0 6 spaces is determined. Therefore, unlike previous arrangements, additional space must be provided in the port to accommodate data message packets of unknown length. However, according to the present invention, when the triple buffer 306 in the port has space to store several triples (for example, ten triples), it is appropriate to generate a congestion request from the destination port of the message packet. . This provides the ability of the target port to store the triples of the message packets in the process being loaded into the message packet buffer 206. The triplet buffer 3 0 6 in each port should be sized by the relevant message packet buffer 2 0 6 in the memory device 200, so that any message before any triple buffer 3 0 6 is full The packet buffer 2 0 6 is almost or completely impossible to fill. _____________-. 26 ^ _____ The standard of this paper is applicable to the Chinese National Standards (CNS) / \ 4 regulations ^ (2 丨 Ox29D > ^) ^ ______P (対 1 Room details ii. £: 1-Jfl refill this K)

、 1T. The central standard of the Ministry of Economic Affairs ^-The Bureau only eliminates the A7 _____B of the Industrial Co., Ltd. V. Description of the invention (24) Once the triad is placed on the communication bus 102, the source port begins a series of Of the message packet from the source port buffer 3 1 6 to the message packet buffer 2 0 6 in the memory device 200, starting from the memory address identified by the third area of the triplet Billion body write cycle. Preferably, the message packet buffer 206 includes a space allocated to each port for storing a message packet to be transmitted by the port. Alternatively, the message packet buffer 206 includes space allocated to each port to store message packets to be received by the port; however, it should be noted that only one non-rain type 'distribution space needs to be provided. Message packets are written in a circular manner into the space allocated to the port in the message packet buffer 206; each new message packet will overwrite the oldest message packet in the allocated space. Partial "message packets should be loaded into those message packet buffers 206 with a predetermined offset from the assigned address. This provides a place to store the header of the message packet, once the message packet is completely loaded into the message packet buffer 206. For example, the header may include the number of acknowledgements assigned to the message packet and the triple of the message packet, and the status of an accepted message packet. The acceptance status indicates whether the entire message packet has been successfully received and loaded into the message packet buffer 206. Because the rest of the message packet will usually include more than 32 bits (the number of data lines in the communication bus 102), multiple memory write cycles usually require the entire message packet to be transferred to the message packet buffer. Within 2 06, since the remainder of the message packet is still being collected into the receiving buffer 3 1 6 of the source port, the message packet should be written into the message packet buffer 2 06. For this reason, the receiving buffer for each port is 3 1 6 _____________________ One paper size applies to China National Pillow Standard (CNS) A4WL (210〆2W.Male A;)

Β7 V. Description of the invention (25). '· It is not necessary to be able to store the entire data message packet. In addition, if the target port is configured to cut through, if the target port is configured to cut through, the target port will be directly from the communication bus in synchronization with the write cycle of the loading message enclosed in the message envelope buffer. The message envelope in the message envelope buffer is loaded into its transmission buffer. In this all-pass operation, the message is sealed into the transmission buffer in the target port for immediate transmission to the LAN segment associated with the target port. Stomach—Once the entire message packet has been loaded into those message packet buffers, the triplet is once again placed on the data line of the communication bus. The target port then stores the triples in the triplet buffer. Therefore, message packets are stored for transmission by the destination port. Then, when the destination port is no longer busy, the destination port retrieves message packets from those message packet buffers. If the target port is no longer busy, the target port will retrieve the message packet from the message packet buffer 2 0 6. This is achieved by the target port requesting access from the bus control module 114 to the communication bus 102. When the request is allowed, the bus control module 1 1 4 places a bit type indicating the bus allowed to be commanded on the command line of the communication bus 1 2 and a bit type indicating the port on the data line. Once the target port takes control of the message bus 102, the target port then places the starting address in the message packet buffer 2 06 by placing the packet (from the third region of the triple of the message packet) on the communication bus. On the data line message packet of row 102 and the bit pattern indicating the memory read and read operation is placed on the command line, a series of read operations above the communication bus 102 are started. Accordingly, the memory control unit 118 accesses the message packets in the message packet buffer 206. Preferably, the target port checks the reception status of the message packet. This paper size applies to the Chinese national standard: t 聿 (CNS) (2〗 0 > < 2 ding public and private) Ministry of Economic Affairs Ministry central standard? 1'- and only the consumer cooperatives printed A 7 _______ Η 7 V. Description of the invention ( 26) If the receiving status indicates that the message packet has not been received, the memory read + fetch operation should be stopped, and no further action on the message packet is taken. In addition, if the cut-through operation is started but unsuccessful, the message packet must be resent from the destination port. For example, if a data collision occurs during transmission on a segment related to the target port, the cut-through operation may not be successful. In this case, the message packet is retrieved from the message packet buffer 206 as described above. When the destination port is receiving message packets from the message packet buffer 206 or directly from the source port (such as the cut-through period) and entering the transmission buffer 3 1 6, the + destination port starts to be controlled by the limit state machine 3 1 4 Send a message packet to the LAN segment associated with the target port. For this reason, the transmit buffer 318 of each port need not be able to store the entire data message packet. The message packet is then received from the LAN segment by the destination node of the message packet. It is quite plausible that without using the message packet buffer 2 06 to store the column message packets, the receiving buffer 3 16 and / or the transmitting buffer 318 of each port can be expanded to store all of each Store message packets. However, this embodiment is less suitable because the performance may be poor, especially if the total storage of temporarily storing data message packets is reduced. In addition, setting a large buffer in each frame will complicate the integration of the multi-port bridge 1000 in a single integrated circuit (the memory device 200 is outside the integrated circuit). Figure 8 shows an example. Explain that the switch of the multi-port bridge 1 00 according to the present invention is activated ---------9Q ^ ____ This paper standard is applicable to China National Pillar Standard (CNS > A4iiL · ^ (2 丨 0X29 以 / ί) — ^ ( ΐ / .ίι read the back note tt-item 4 write this page)

A7 _ B7 V. Description of the invention (27) A block diagram of a machine 50, a memory device 200, and an external processor 500. The on / off engine 150 includes the 'bus control block block 1 1 4', MP port and post box 1 1 6, the memory control block 1 1 8 and the query control block 1 2 0 illustrated in FIG. The memory bus 5 0 2 connects the switching engine 15 0 and the external processor 5 0 0 to the memory device 2 0 0. Preferably, the access of the engine 150 and the external processor 500 to the memory device 2000 is realized by a multiplexer as a part of the memory bus 502, and the The multiplexer is controlled by the switching engine 15. The multi-port bridge 100, which includes a communication bus 102 (Fig. 5), should be implemented as a integrated circuit mounted on a printed circuit board. A memory device 200 and an external processor 500 are also mounted on the printed circuit board 504. As described above with reference to FIGS. 5-7, the bridge and filtering functions of the multi-port bridge are mainly performed by switching the engine 150 and the buffer memory 200. Because turning on and off the engine 150 and port 104—112 is better to connect to most state-limiting machines via communication bus 102, the multi-port bridge 100 provides high-frequency bandwidth for multi-port Bridge 100 indicates a data message packet. Therefore, according to the present invention, an external processor 500 is provided to perform the task of supporting the multi-port bridge 1000 function. These functions include: providing communication ports to enable nodes in the local area network to communicate with nodes in a different local area network or global area network, and enabling nodes in the local area network to communicate with file servers in the local area network; provide The parameters are used to set the initial register of the switching engine: collect data from the local area network to perform network management functions; and provide services for switching the engine. The mailslot interface according to the present invention allows this paper size to be applicable to China, etc.-Standard Cui (CNS) (2 丨 〇 X 297.:.> G 邙 1 read the back note & fill in the VC? Page) ο Order-line warp ¾-Ministry Central Standards ϊ. Bureau only eliminates cooperation Du Yin Bi-Step-Ministry Central Standards Bureau, νη Industrial Consumer Cooperative Cooperatives A7 _____B7 _ 5. Description of the invention (28) Xu external processor 5 0 0 These functions are provided without the need for a large number of pins in the integrated circuit for this interface. Preferably, the external processor 500 is implemented in the form of a reduced instruction set computer (RISC) to improve speed. The external processor 500 may have its own dedicated resources 506, such as a memory for storing software for operating the external processor 5 ◦ 0 and a board for the external processor 500. In addition, when the external processor 500 performs the function of a local area network file server, the resource 506 may include a large-scale file for storing application programs and data files accessible by the external processor 500. Storage device. At the same time, when the external processor 500 performs the function of interconnecting a local area network with a different local area network or a global area network, the resource 506 may include a communication device such as a telephone modem, and the overall service digital network (1 SDN) interface, an interface T 1 media interface or T 3 media that can be accessed by an external processor 500. In addition, multiple external processors 502 can be coupled to the memory bus 502. In such cases I additional resources may be provided to these multiple external processors, such as one or more / mass storage devices and / or one or more communication devices. Figure 9 illustrates a schematic block diagram of a device for periodically updating a lookup table 204 according to the present invention. The lookup table 204 includes three subtables 204 A-C and should be part of a memory device 200 (Figure 5) configured to emulate a C AM device. However, it is quite plausible that the lookup table 204 can be contained in one or more records different from the memory device 200 million. For example, lookup table 204 may be one or more CAM devices. ------------------------ This paper size is applicable to China National Sample (CNS) 8 4 Yue Ku (2I0x2 < ^ >#.) --- 1Λ1§ · : ίϊ1ϊ '而 之 ii. * Fill in the matter again?}

、 1T line _ A7 ___ _ _Η 7 V. Description of the invention (29) Before the address and the corresponding port confirmation are stored in the lookup table 2 0 4, the query control module 1 2 0 should perform the hybrid operation on each address. law. The promiscuous algorithm should remove the most significant bits of each address and keep the least significant bits unchanged, although other promiscuous algorithms can also be used. "Preferably, the sixteen least significant bits. It is held so that there are 2 1 6 (6 5, 5 3 6) possible addresses after being mixed. Therefore, each sub-table 204 A-C should have 216 (65, 5 3 6) billion positions. However, it is quite plausible that the promiscuous algorithm can cause the remaining bits to have different numbers. Therefore, the number of billion positions in each subtable is different. FIG. 10 illustrates a schedule for periodically updating the address and port information stored in each of the three query subtables # 1-3 (204A-C) according to the present invention. Assume that at time 0.0, the multi-port bridge 100 is first opened or reset so that any of the first sub-tables # 1 to 3 (for example, opening or resetting the sequence will set each table input to all zeros). No entry for. During the first 2.5 minute period, the source address in each piece of data contained in the message packet appearing on any port of the multi-port bridge 100 was mixed into 1 by removing the most important 32 bits. 6 bits. An acknowledgement of the port (source port) that received the data message packet is stored in a location on each of the tables # 1 (204A) and # 2 (2_0 4B) designated by the mixed address. Preferably, the 32 address bits moved during the promiscuous algorithm are also stored in the appropriate table along with the port confirmation. The remaining 32 bits are used to verify two of them ___________- 32 -___________ This paper size is applicable to China National Standards (CNS) / \ 4 圯Si A7 __B7 V. The full address in the event of the description of the invention (30) or the mixed address shared by multiple nodes (for example, they have the same least significant 16 address bits). If in any table The mixed source has an input (based on the message packet from the first node), and the bits of the remaining 32 source address message packets are compared to 32 bits stored in the memory location specified by the mixed source address. Meta: If the remaining 32 bits do not match, it indicates that the data message packet sent from the second node has the same least significant bits as the first node. In this case, the linked list is The impact table is formed to store the remaining 32 bits of the second node and the address and port confirmation of the port to which the second node is connected. If the remaining 32 bits match, it indicates that it has only Miscellaneous source address A node (first node) has transmitted a data message packet. In this case, the connection list is not needed. At the same time, during the first 2.5 minutes period, it is included in the received by the multi-port bridge 100 Each data message packet is hashed into sixteen bits. The memory controller 106 uses the hashed destination address to query the data message packet in Table # 1 (2 0 4 A) to be indicated for transmission to the appropriate destination node.璋》 _ If no data is stored in the location specified by the hybrid destination address, it means that those data message packets are left by the node to a data message packet that has not yet been transmitted or a "live" message packet. In this case, Multi-port bridge 100 does not have sufficient information to properly indicate message packets. Therefore, I-message packets should be indicated to each port except the port that received the message packet (message packet is played). If the data is stored in The location specified by the mixed target address, the rest ___________________________ _________ This paper size is applicable to the Chinese national sample (CNS ·) dry (210x2 <mM; ._)-^ ------- packing --- ---- iT --------- 0 (¾Read the note on the back and note the matter later, and then 4, this page)

The central standard bureau of Dibu Ministry only consumes the seal of the Hezhu Society. —— _____ Β7 ____ V. The target residence of the invention description (31). The address bit is determined by the 3__2 address bits stored in the position. Do they match. If no other address bits match ', this means that two or more nodes share the same 16 least significant address bits. When this happens during the learning period, a linked list is set to store the port confirmations for each node that shares 16 least significant address bits. If the remaining address bits match _, the port confirmation is used to properly direct those data message packets to that port for transmission to the target node. If the port specified by the destination address is the same as the port 1 that received the message packet, this means that the source node and the destination node of the message packet are in the same local network segment. Because the destination node receives the message packet directly from the source node, in this case, the multi-port bridge 100 should not point the message packet to the port 0. At the end of the first 2.5 minutes, subtable # 3 (204C) Eliminate all entries. However, if no entry has been made in sub-table # 3 since the start, the sweep of sub-table # 3 may be skipped. Then, in the second 2.5 minute period, the data has been stored in sub-table # 2 (the same way that the data was stored in sub-tables # 1 (204Α) and # 2 (204Β) during the first 2.5 minutes. 204B) and # 3 (204C). Meanwhile, during the second 2.5 minute period, Table # 2 (20413) is utilized to query the period. At the end of the second 2.5 minutes | Subtable # 1 (204Α) erases all entries. Then, during the third 2.5 minute period, the data is stored in sub-tables # 1 (204A) and # 3 (204C) and table three is used for the query period. At the end of the third 2.5 minutes | Sub-sheet _____________________- 34 -________ This paper size applies to China National Standards (CNS) 栝 (2) 0 / 2people々 >#) (ΐΛιι7-rif ] '而 之 Note ^: Revisited reprints of matters ^) —-r J ------- iff nt—

-1T • I.---· A7 ________ ΙΪ7 _ 5. Description of the invention (32) # 2 (2 04B) Eliminate all inputs. During the fourth 2.5 minute period, the data is stored in sub-tables # 1 (204A) and # 2 (2 0 4 B) and Table 3 is used for the query period. At the end of the fourth 2.5 minutes, subtable # 3 deletes all entries. Note that the same steps are performed during the fourth period as during the first period (except that subtable # 3 need not be deleted at the end of the first 2.5 minutes). Therefore, the program loops by repeating the steps of the first, second, and third periods. According to the present invention, the input utilized in any query period does not exceed 5 minutes ago. However, each age entered is not tracked individually (-1.?51 read the back of the note of the city item #fill this page)

.A The Ministry of Economic Affairs of the Central Government Bureau of the People's Republic of China ·· Only printed by Hezhushe ^ The paper size is suitable for flooding China National Standards (CNS) Λ ^ ίΐ ^ (2ΐ〇 χ297Γ > ΐ 1 一-

Claims (1)

6. The scope of patent application method, section I of port I of the multi-port bridge device, the party a. Receive message packets, data area in the future, and message b. Query c. Allocate d. Form the area to identify the coming area It is used to identify the memory e · the first f • the memory g of the packet. The second h · the three i · the message into the target port j. The message A8 B8 C8 D8 • a control in the multi-port bridge It is used to communicate with each other by steps: Including most of the multi-bank bridge method, the message packet in the source port is contained in the memory port of the message packet source port in the table, and the memory device times the ternary information packet. Position; the sub-trip triplet in the interest-storage packet connection is used to transmit the information packet and send the data message packet; the process buses are connected to each other by a memory connection; System (please read the precautions on the back before filling this page) The stapled buffer has an indexed group of one triplet and two from the source destination address in the access memory Information The address and port are related to determine the location of the message to the message, three of which are to identify the internal packet revenue that is targeted to the message bus device, one by one, the source address and the data segment to receive; the message packet destination port: the message packet : The tuple includes the first standard port, and the third packet position: upper: the ternary allocated to the message group on the communication bus and stored in the target port is divided into the buffer from the memory device; And into the segment group buffer related to the destination port of the message packet; allocating the message packet bit 2 _ according to the method of the scope of patent application where the step of querying the destination address is performed when the data area is connected by the source port _ The method of applying for the patent scope is to apply the Chinese paper standard (CNS) A4 specification (210X297 mm) to the paper size of the paper envelope. -36- Α8 Β8 C8 D8 When the district is received by the source port, it will be executed 'll®! III 4. According to the method of the scope of patent application, the transmission message is sealed by the staff of the Central Ministry of Economic Affairs and the Ministry of Economic Affairs. Step-pack of printed material in the target area is performed when the port is received from the apparatus body. It. ·: ··%, *. ≫ t 5. A kind of multi-port bridge is used to connect a plurality of local area network sections to each other. The multi-dock bridge includes: a. Communication bus; b. Coupling A memory device to a communication bus; c. A source port coupled to the communication bus, where the source receives a message packet from a source node in a segment of a local area network associated with the source port, where the source port is receiving A query cycle starts when an address of a destination node of a message packet is formed, and a product of the query period is a triplet, which has a first region for identifying a source 珲, a second region for identifying a destination port of the message packet, and identification in a memory. The third area of the location where the bulk device is assigned to the message packet: and d. The destination port used to transmit the message packet to the destination node in a segment in the local area network associated with the destination port, where if the destination port is not Busy, after the triplet is placed on the communication bus for the first time, the destination port receives the message packet directly from the source port, and if the destination port does not receive the message packet directly from the source port, the triplet After the second is placed on the communication bus, the target port to receive messages from the memory device package seal. According to the scope of the patent application No. 5 Portuguese ^^^ 381, the message packet was released on the first triad and the second triad was quilted (please read the precautions on the back before filling this page) Standards are applicable to Chinese National Standard (〇 奶) 六 4 Specifications (210 > < 297 mm) -37-Printed by A5 B8 C8 DS of Zhengong Consumer Cooperatives, Central Bureau of Standards, Ministry of Economic Affairs VI. Between are stored in the body. 7. According to item 5 of the scope of patent application, the source port includes a receiving buffer for receiving data message packets. 8. The multi-port bridge according to item 7 of the scope of patent application, wherein the data message packet has a maximum expected size, and the receiving buffer is smaller than the maximum expected size. 9 The multi-port bridge according to item 5 of the patent application, wherein the target port contains a transmission buffer for transmitting data message packets. 10. The multi-port bridge according to item 9 of the scope of patent application, wherein the data message packet has a maximum expected size (値) and the receiving buffer is smaller than the maximum expected size. 1 1 A method for controlling the flow of data message packets in a multi-port bridge. The multi-port bridge includes a plurality of ports interconnected to a memory device by a communication bus. In a segment of a local area network, the method includes the following steps: a. The data message packet in the buffer in the source port is received into a received message packet, and the message packet has a destination address'—the source address and the data area, and Message packets are being received from the segment associated with the source port; b. Query the target address in a table to determine the message packet destination port; c _ allocate a location in the memory device to the message packet; d. Form a message packet Triplet, where the triplet includes the first area to identify the source port, the second area to identify the destination port, and the third area to identify the device that is assigned to the message packet in the device. Position; e. Put the triplet on the communication bus for the first time; (Please read the precautions on the back before filling this page) This paper size applies to China National Standards (CNS > A4 size (210X297 mm) -3F- printed by A8 B8 C8 D8 of the Consumer Cooperatives of the Central Ministry of Economic Affairs of the Ministry of Economic Affairs. 6. Scope of patent application f. Load the information packet into the record The location of the memory allocated to the message packet within the billion-body device; and g. Determining whether the destination port can receive message packets directly from the source port, if the destination port can receive message packets directly from the source port The operation of loading the message packet into the memory device synchronously receives the message packet into the transmission buffer in the destination port, and if the destination port cannot directly receive the message packet from the source port ^ After the destination port can receive the message packet The message packet is received from the memory into the target port. 1_2 The method according to item 11 of the scope of patent application 'progressively includes the second step of placing the triplet on the communication bus after the loading step is completed. 13. The method according to item 12 of the scope of patent application 'further includes a period during which the triple is placed on the communication bus for the second time. The step of storing the triplet in the target port. ^ 1 4 _The method according to item 11 of the scope of patent application 'further includes the step of performing transmission of the message packet while the message packet is being received in the transmission buffer of the target port. 1 5 · According to the method in the scope of patent application No. 11 where the message packet also has a data area and where the destination address query step is performed when the data area is received by the source port. 1 6. According to the scope of patent application No. 1 A method according to item 1, in which a step of loading a message packet into a memory device is performed when a data area is received by a source port. 1 7. The method according to item 11 of the scope of patent application 'further (please read the note on the back first) Please fill in this page again on the page) The paper size of the booklet is applicable to the standard 4 (210 father and 297 mm) of China National Standard Feng (milk milk) person. + -39- A8 BS C8 D8 Perform the steps of transmitting message packets when being received from the memory device by the target port. 1 8. — A method of periodically updating the entries in the lookup table to pass through a multi-port bridge with multiple ports as appropriate Indicates a data message packet, each port has a port acknowledgement, where each data message packet is sent by an individual one of the plurality of nodes, each node has a node address, and each node corresponds to an individual one of the plurality of nodes, The method includes the following steps: a. Storing the node address of each node that sends a data message packet about the corresponding port during the first period in the first memory layer and the second memory layer: b. The node address of each node that sends a data message about the port confirmation of the corresponding port during the second period is stored in the second memory layer and the third memory layer; c. The port confirmation about the corresponding port will be sent during the third period The node address of each node of the data message packet is stored in the first memory layer and the third memory layer. Printed by the Consumers' Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs (please H-read the notes on the back before filling this page)-Line 1 9 · According to the method of item 18 of the scope of patent application, further-included in the first period The step of using the first memory layer to indicate the data message packet. 20 • According to the method of item 19 of the scope of patent application, the method further includes the step of using the second memory layer to indicate the data message packet during the first period. 2 1 _According to the method of item 20 of the scope of patent application, the step further includes using a third memory layer during the first period to instruct the data message to encapsulate this material through the wealth management materials (CNS) A view ^ · ( 21GX297mm) -40 ----- The paper size of the printed papers of the Consumer Cooperatives of the Central Government Bureau of the Ministry of Economic Affairs (Chinese National Standard (CNS)) A4 (210X297mm) B8 C8 ______ D8___; __ VI. Application Patent scope,. * Steps. 2 2 · The method according to item 20 of the scope of patent application, wherein the first memory layer is used to instruct the data message package product to halve a triple of each message packet, each triple having a first region containing Corresponds to the source port confirmation of the node sending the message packet, a second area 'contains the acknowledgement of the destination port corresponding to the node intending to receive the message packet' and the third area contains the message allocated to the message packet buffer The starting address of the packet. 2.3. The method according to item 18 of the scope of patent application 'further includes the step of mixing each node address before the storage node address is in each of the first, second and third billion layers. 24. The method according to item 18 of the scope of patent application, wherein the second period starts at the end of the first period. 25. The method according to item 24 of the scope of patent application ', wherein that third period starts at the end of the second period. 26. The method according to item 25 of the scope of patent application 'further includes the step of erasing the first memory layer at the end of the second period. 27. The method according to item 26 of the patent application scope further includes the step of erasing the third memory layer at the end of the second period. .2 8. The method according to item 27 of the scope of patent application 'further includes the step of erasing the first memory layer at the end of the third period. 2 9. The method according to item 28 of the patent application scope further includes repeating steps a'b, c ', and d at the end of the third period. 3 0. — A device that periodically updates the entries in the inquiry form to be used (please read the note on the back, and then fill out this page) to order the printing by the Ministry of Economic Affairs and the Consumer Cooperatives of the Bureau Λ8 __D8 VI. Application The scope of the patent is appropriately indicated by a multi-port bridge with a large number of ports. Each port has a port to confirm. 'Each of these data message packets is issued by an individual of a plurality of nodes. Each node has a node address and Each node corresponds to an individual one of a plurality of nodes, and the device includes: a. A first memory layer for storing, during a first period, a node bit of each node that sends out a data message packet of a port acknowledgement on a corresponding port; Address and the first memory layer is used to store the node address of each node that sends a data message packet about the corresponding port during the third period, wherein a second occurs during the first period and the third period During the time period; b — — the second memory layer is used to store the node bit of each node that sends a data message packet about the corresponding port during the second time period And the second billionth body layer is used to store the node address of each node that sends a data message packet about the corresponding port's port confirmation during the first period; and c. The third billionth body layer is used in the third period Storing the node address of each node that sends a data message packet about the corresponding port during the period, and the third memory layer is used to store each node that sends the data message packet about the port confirmation about the corresponding port during the second period The node address. 3 1 • Device according to item 30 of the scope of patent application * where each node address is mixed before storage. 32. The device according to item 30 of the scope of patent application, wherein the first, second and third memory layers are each a CA device. 33. The device according to item 30 of the scope of patent application, wherein each of the first, second and third memory layers is a RAM device emulating a CAM device. (CNS) {210X297 ^ it)-42.- (Please read the precautions on the back before filling this page) A8 B8 C8 D8 VI. Patent application scope 3.4 · The device according to item 30 of the patent application scope, in which the second memory layer is deleted at the end of the third period. _ 35. The device according to item 34 of the U.S. patent, wherein the second g-memory layer is eliminated at the end of the first period. 36. The device according to item 30 of the scope of patent application, wherein the first memory layer is used to indicate a message packet during the first period by using a multi-port bridge. 37. The device according to item 36 of the scope of patent application, wherein Each message packet is formed into a triplet. Each triplet has a first field containing the confirmation corresponding to the port where the message packet started, the second area contains the confirmation of the port of the intended recipient section of the message packet, and the third The area contains the start address of the message packet allocated in the message packet buffer * 3 8 .. The device according to item 36 of the patent application scope further includes the use of a second billion layer to indicate the data message during the second period Envelope steps. According to the device under the scope of patent application 38, it is further included in the memory layer to instruct the data message packet during the third period of time. (Please read the precautions on the back before filling this page) A device that periodically updates the entries in the lookup table to properly indicate data message packets via a multi-port bridge with a large number of ports, each port has a Tan confirmation, of which Each data message packet is sent by an individual of a plurality of nodes, each node has a node address, and each node corresponds to an individual of a plurality of nodes of the multi-port bridge. The device includes: a. A first memory layer This paper size applies to the Chinese national standard (CNS > A4 specification (210X 297K ^ -43-ABCD) VI. Patent application scope b. The second billionth tomography; c. The third billionth tomography; and d. To be coupled To each of the first, second, and third billion-layer control logic circuits, the control logic circuit is used to store the node bits of each node that will send out a data message about Wei Yingbu ’s confirmation during the first period. The node address is stored in the first and second billion layers, and the node address is stored for each node that sends a data message packet about the corresponding port's port confirmation during the second period The second memory layer and the third memory layer, and the node address of each node for storing a data message packet about the corresponding port during the third period are stored in the first memory layer and the third memory layer Body. 4 1. The device according to item 40 of the scope of patent application, wherein each data message packet contains the address of the node that generated those data message packets. 4 2. The device according to item 40 of the scope of patent application, wherein each The node addresses are mixed before being stored. 4 3. The device according to item 40 of the scope of patent application, in which the first, second and third memory layers are each a CAM device.-Staff of the Central Government Office of the Ministry of Economic Affairs Printed by a consumer cooperative (please read the notes on the back before filling this page) Line 4 4 Device according to item 40 of the scope of patent application, where each of the first, second and third memory layers is a RAM that emulates a CAM device The device. 4 5. The device according to item 40 of the patent application, wherein the second period starts at the end of the first period. 46. The device according to item 40 of the patent application, wherein the third period is at the Begins at the end of the period. This paper size applies to Chinese national standards (CNS > A4 size (2 丨 0X297 mm) -44- Λ8 B8 C8 DS VI. Patent application scope 4 7. Device according to item 46 of patent application scope 'Where the second memory layer is deleted at the end of the third period. 4 8. The device according to item 47 of the patent application scope' wherein the third memory layer is deleted at the end of the first period. 4 9. According to Apparatus No. 40 in the scope of the patent application 'wherein the first memory layer is used to indicate a message packet during the first period by using a multi-port bridge. 0 50. Apparatus in accordance with the scope of the patent application No. 49' in which The packet is formed into a triple. Each triple has a first field containing the confirmation corresponding to the port of the start message packet, a second field containing the confirmation of the port of the intended recipient section of the message packet, and a blue field containing the The start address of the message packet assigned to the message packet buffer. 5 1. The device according to item 49 of the scope of the patent application, in which the second memory module is used to indicate the message packet during the second # segment by using the multi-port bridge (please read the precautions on the back before filling this page) Printed paper wave ruler of the Central Standards Bureau of the Ministry of Economic Affairs, Shellfish Consumer Cooperative, and applicable Chinese National Standard (CNS) A4 (210X297 mm) · 45-