WO2004088931A1 - データ通信負荷分散制御プログラム、およびデータ負荷分散制御方法 - Google Patents
データ通信負荷分散制御プログラム、およびデータ負荷分散制御方法 Download PDFInfo
- Publication number
- WO2004088931A1 WO2004088931A1 PCT/JP2003/004148 JP0304148W WO2004088931A1 WO 2004088931 A1 WO2004088931 A1 WO 2004088931A1 JP 0304148 W JP0304148 W JP 0304148W WO 2004088931 A1 WO2004088931 A1 WO 2004088931A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- communication
- data
- computer
- algorithm
- distribution
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/24—Multipath
- H04L45/245—Link aggregation, e.g. trunking
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Definitions
- the present invention relates to a data communication load distribution control program and a data load distribution control method for performing data communication via a multiplexed transmission path, and more particularly to a data communication load distribution control program using trunking and a data load distribution control. About the method. Background art
- Trunking is one of the methods for increasing the bandwidth of communication between a plurality of devices.
- devices are connected by multiple transmission lines.
- Each device logically treats a network interface card (NIC) connected to each of a plurality of transmission paths as one LAN port. As a result, the transmission band between the devices can be improved.
- NIC network interface card
- FIG. 21 is a diagram illustrating a configuration example of a conventional trunking system.
- the conventional trunking system includes a computer 910, a switching hub 920, and other computers 930, 940, 950, and 960.
- the computer 910 has an application 911, a trunking mechanism 912, and NICs 913 to 916.
- the application 911 is a processing function that communicates with the other computer 930, 940, 950, 960 and performs various data processing.
- the trunking mechanism unit 912 performs communication with the switching hub 920 by regarding the plurality of NICs 913 to 916 as one LAN port. Note that the trunking mechanism section 912 has a MAC (Media Access Control) address for a plurality of NICs 913 to 916 and an IP (Internet Protocol) address shared by each NIC 913 to 916. Has the set configuration information 912a.
- MAC Media Access Control
- Each of the NI C913 to 916 is connected to the switcher hub 920 by a separate transmission line, and performs data communication with the switching hub 920.
- the switching hub 920 four of the eight LAN ports are connected to the NICs 913 to 916 of the convenience store 910. The other four LAN ports are connected to the other computers 930, 940, 950, 960.
- the switching hub 9220 has a trunking corresponding section 9221.
- the trunking support unit 921 has configuration information 921a in which information such as an IP address of a connection partner to each LAN port is stored.
- trunking is performed between the computer 910 and the switching hub 920. That is, the computer 910 generates one logical NIC using the multiplexed NICs 913 to 916. The computer 910 transmits and receives data overnight through the logical NIC.
- the switching hub 920 sends the data sent from the computer 910 to the other computer 930, 940, 950, 960. In addition, the switching hub 920 receives the data sent from the other party's computer 930, 940, 9550, 9600, 1 and the ranking corresponding section 9 21 Distribute the data to one of the LAN ports connected to the 910 and send the data to the computer 910 via its LAN port.
- the load distribution control of a plurality of LAN ports can be performed without using the switching hub 920 having the trunking corresponding section 9221.
- a host computer having a plurality of LAN (Local Area Network) adapters notifies a communication partner of the physical address of an arbitrary LAN adapter.
- the communication partner specifies the notified physical address and transmits the data to the host computer.
- data transmitted from a plurality of communication partners can be received by distributing the load to a plurality of LAN adapters (for example, see Patent Document 1).
- Patent Document 1
- connection cannot be established unless both the computer and the switching hub have a trunking function, and the application conditions are limited. I was wearing it.
- it is necessary to set the number of NICs and ports to be used between the switching hub and the computer in accordance with the distribution algorithm, and to define a dedicated MAC address for executing trunking. The settings were complicated and the user made mistakes.
- the NIC to be used is determined by means such as exchanging information on the IP address and the MAC address between the computers.
- the exchange of address information alone limits the distribution algorithm. For example, when there is only one communication partner device, there is a problem that only a specific NIC is used and cannot be distributed. Disclosure of the invention
- the present invention has been made in view of such a point, and an object of the present invention is to provide a computer which can perform flexible trunking with a switching hub having no trunking function. .
- a data communication load distribution control program as shown in FIG. 1 is provided.
- a data communication load distribution control program according to the present invention comprises a data communication load distribution control unit for distributing a communication load on a transmission line in a computer capable of implementing a plurality of communication interfaces 1a, lb, lc, 1d. It implements the function of 1 e.
- the computer 1 executes the following processing based on the data communication load distribution control program.
- the computer 1 transmits the physical addresses of the plurality of communication interfaces 1a, lb, 1c, and Id, and the data transmitted from the communication partner computer 3 to the computer 1 through the plurality of communication interfaces 1a, lb, 1c, and Id.
- a distribution algorithm notification bucket 4 including the distribution algorithm when distributing to 1d is generated (step S 1).
- the computer 1 is connected to a communication phase connected via a network.
- the distributed algorithm notification packet 4 is transmitted to the hand computer (step S2).
- the physical addresses of the plurality of communication interfaces 1a, 1b, 1c, and 1d mounted on the computer 1 are notified to the other computer 3, and the load balancing algorithm can be designated.
- a load distribution data transmission program is a load distribution data transmission program for distributing a communication load of a transmission path to a communication partner computer 1 capable of mounting a plurality of communication interfaces la, 1, lc, and 1d. This implements the function of the transmission unit 3b. Based on this load distribution data transmission program, the combination 3 executes the following processing.
- the computer 3 transmits the physical addresses of the plurality of communication interfaces 1 a, 1 b, 1 c, and Id, and data to be transmitted to the computer 1 to a plurality of communication interfaces la, lb, 1 c, 1
- the physical addresses of the plurality of communication interfaces 1a, lb, 1c, and 1d and the distributed algorithm are stored in the data table 3c.
- the computer 3 determines the communication interface to be the destination of the transmission data 5 to be transmitted to the communication partner computer according to the distributed algorithm. Further, the computer 3 specifies the physical address of the communication interface determined as the destination as the destination, and transmits the transmission data 5.
- the transmission data 5 transmitted from the computer 3 to the communication partner computer is transmitted via the plurality of communication interfaces 1 a, lb, lc, and 1 d using the distributed algorithm specified in the distributed algorithm notification packet 4.
- the load is distributed to the transmission line and transmitted.
- a physical address of the plurality of communication interfaces is provided.
- a distributed algorithm notification packet including a distributed algorithm for distributing data transmitted by the communication partner computer to the computer to the plurality of communication interfaces. And transmitting the distributed algorithm notification packet to the communication partner computer connected via a network to execute the processing.
- FIG. 1 is a conceptual diagram of the invention applied to the embodiment.
- FIG. 2 is a diagram showing a configuration example of a system to which the present invention is applied.
- FIG. 3 is a diagram showing the internal configuration of the trunking mechanism.
- FIG. 4 is a diagram illustrating a hardware configuration example of a server computer.
- FIG. 5 is a flowchart showing a data communication processing procedure.
- FIG. 6 is a conceptual diagram illustrating a distributed algorithm notification process.
- FIG. 7 is a diagram showing a frame configuration for a distribution algorithm notification bucket.
- FIG. 8 is a diagram showing an example of the contents of the data portion of the distribution algorithm notification bucket.
- FIG. 9 is a diagram showing an example of a data structure of a distributed algorithm table on the server computer side.
- FIG. 10 is a diagram showing an example of registration of the distribution algorithm table according to the distribution algorithm notification packet.
- FIG. 11 is a diagram illustrating a load distribution processing state of data transmitted from the server computer.
- FIG. 12 is a diagram showing a load distribution processing state of data transmitted from a client computer.
- FIG. 13 is a diagram illustrating an example of a system configuration according to the second embodiment.
- FIG. 14 is a diagram illustrating an example of a data structure of a distributed algorithm table according to the second embodiment.
- FIG. 15 is a flowchart showing a processing procedure at the time of starting the system according to the second embodiment.
- FIG. 16 is a diagram showing a data transmission procedure when a plurality of server computers are used.
- FIG. 17 is a diagram showing a data transmission situation when a plurality of server computers are used.
- FIG. 18 is a diagram for explaining the principle when one NIC is added.
- FIG. 19 is a flowchart showing the procedure for updating the distributed algorithm table when adding the NIC.
- FIG. 20 is an explanatory diagram of the principle when one NIC is deleted.
- FIG. 21 is a diagram illustrating a configuration example of a conventional trunking system. BEST MODE FOR CARRYING OUT THE INVENTION
- FIG. 1 is a conceptual diagram of the invention applied to the embodiment.
- the present invention is applied to data communication between the convenience 1 and the computer 3.
- the computer 1 has a plurality of communication interfaces 1a, 1b, 1c, 1d.
- the physical addresses of the communication interfaces l a, l b, l c, and 1 d are “a”, “b”, “c”, and “d”, respectively.
- the communication interfaces l a, l b, l c, and 1 d are respectively connected to the switcher hub 2 by individual transmission lines.
- the computer 1 executes the data communication load distribution control program, thereby constructing the function of the data communication load distribution control unit 1e.
- the data communication load distribution controller le has a function of distributing the communication load of the transmission path.
- the data communication load distribution control unit 1 e performs the following processing.
- the data communication load distribution control unit 1e transmits the physical addresses of the plurality of communication interfaces 1a, lb, 1c, and Id, and the data transmitted by the communication partner computer 3 to the computer 1 through the plurality of communication interfaces 1. a, lb, lc, Id A distribution algorithm notification bucket 4 including a distribution algorithm to be used is generated (step S1). Then, the overnight communication load distribution control unit 1e transmits the distribution algorithm notification packet 4 to the communication partner computer 3 connected via the network (step S2).
- the physical addresses of the plurality of communication interfaces 1a, 1b, lc, 1d mounted on the computer 1 are notified to the other computer 3, and the load balancing algorithm can be designated.
- the computer 3 has a communication interface 3 a, and the communication interface 3 a is connected to the switching hub 2.
- the computer 3 executes the distributed load data transmission program to realize the function of the load distribution data transmission unit 3b for distributing the communication load of the transmission path to the computer 1.
- the load balancing data transmitting unit 3b executes the following processing.
- the load distribution data transmission unit 3b transmits the physical addresses of the plurality of communication interfaces 1a, lb, 1c, and 1d and the data transmitted to the computer 1 to the plurality of communication interfaces la and lb. , lc, 1d, the distribution algorithm notification packet 4 including the distribution algorithm is received, the physical address of the plurality of communication interfaces 1a, 1b, lc, 1d and the distribution algorithm are distributed in a table. 3 Store in c. In the example of Fig. 1, the physical addresses "a, b, c, d" and the distributed algorithm "round robin" are registered.
- the load distribution data transmission unit 3b determines a communication interface to be the destination of the transmission data 5 transmitted to the communication partner computer according to the distribution algorithm. Further, the computer 3 specifies the physical address of the determined communication interface as the destination, and transmits the transmission data 5. For example, transmission data 5 is divided into four, and the first data 5a is transmitted with the physical address "a" specified as the destination. Thereby, the data 5a is transferred to the communication interface 1a in the switching hub 2. The second data 5b is transmitted with the physical address “b” specified as the destination. As a result, the data 5b is transferred to the communication interface 1b in the switching hub 2. The third data 5 c is transmitted with the physical address “c” specified as the destination. This gives data 5 c is transferred to the communication interface 1c in the switching hub 2. The fourth data 5 d is transmitted with the physical address “d” specified as the destination. Thereby, the data 5 d is transferred to the communication interface 1 d in the switching hub 2.
- the transmission data 5 transmitted from the computer 3 to the communication partner computer is transmitted by the distribution algorithm specified in the distribution algorithm notification bucket 4 via the transmission paths via the plurality of communication interfaces 1 a, lb, lc, and 1 d.
- the load is distributed to and sent.
- the computer 1 notifies the computer 3 of the distributed algorithm notification packet 4 by notifying the physical addresses of the plurality of communication interfaces 1a, lb, lc, and Id and the distributed algorithm.
- the transmission data 5 is transmitted to the physical address of the communication interface determined according to the mechanism. Therefore, even if the switching hap 2 does not have a trunking function, trunking using a plurality of transmission paths can be performed. As a result, a highly flexible network system can be constructed.
- a communication interface used for communication can be divided for each computer of the communication partner by notifying the individual distributed algorithm to each computer.
- the communication interface of the transmission destination in the computer 3 it is also possible to determine the communication interface of the transmission destination by referring to the network layer in the OSI hierarchical model and higher-level protocols.
- the communication interface of the transmission partner can be determined according to the type of application used for communication.
- a signal unit is particularly called a frame.
- FIG. 2 is a diagram showing a configuration example of a system to which the present invention is applied.
- a plurality of client computers 300, 300a, 300b, and 300c are connected to the server computer 100 via the switching hub 200.
- a WWW (World Wide Web) server 110 In the server computer 100, a WWW (World Wide Web) server 110, a trunking mechanism unit 120, and NICs 131 to 134 are provided.
- a WWW World Wide Web
- the WWW server 110 provides various contents in response to requests from the Web browsers 310, 310a, 310b, 310c in the client computers 300, 300a, 300b, 300c.
- the node name of the server computer 100 is rhostAj.
- the node name of the client computer 300 is rhostBJ
- the node name of the client computer 300a is rhostCJ
- the node name of the client computer 300b is rhostDj
- the node name of the client computer 300c is "hostE". is there.
- the trunking mechanism unit 120 is provided between the WWW server 110 and the NICs 131 to 134. 1.
- the ranking mechanism unit 120 secures a logically high communication band by bundling a plurality of NICs 131 to 134. Then, the trunking mechanism unit 120 generates a packet including data to be transmitted from the WWW server 110 to the client computers 300, 300a, 300b, and 300c, and generates one of the plurality of NICs 131 to 134. Choose one and send.
- the trunking mechanism section 120 has a distributed algorithm table 121.
- the distribution algorithm table 121 defines an algorithm for load distribution at the time of data transmission. Then, the trunking mechanism unit 120 refers to the distributed algorithm table 121 and selects an NIC to be used for packet transmission so that the load of communication via each of the NICs 131 to 134 is equalized. That is, load distribution is performed. The load distribution is performed by, for example, a round robin method. Further, when the server computer 100 is activated, the trunking mechanism unit 120 transmits a distribution algorithm notification packet to each of the client computers 300, 300a, 300b, and 300c.
- the distributed algorithm notification packet specifies the algorithm for selecting the physical address (MAC address) to be specified as the destination when the client computer 300, 300a, 300b, 300c sends a packet to the server computer 100. Information.
- the distribution algorithm to be notified is defined in the distribution algorithm table 121.
- the distributed algorithm notification packet also includes the MAC address of each of the NICs 131 to 134 of the server computer 100.
- NIC 131-: L 34 is connected to switcher hub 200 by a LAN cable or the like.
- the NICs 131 to 134 transmit and receive frames to and from the switching hub 200.
- individual MAC addresses are set in each of the NICs 131 to 134.
- the MAC address of NIC 13 1 is raj
- the MAC address of NIC 132 is “b”
- the MAC address of NIC 133 is “c”
- the MAC address of NIC 134 is “ d ”.
- the switching hub 200 has a plurality of LAN ports 211 to 214 and 221 to 224, and receives a frame input from the LAN port.> A LAN to which a NIC corresponding to the MAC address of the destination of the frame is connected. Forward to port.
- the LAN ports 211 to 214 are connected to the NICs 131 to 134 at the server computer via LAN cables, respectively.
- the LAN ports 221 to 224 are connected to client computers 300, 300a, 300b, and 300c, respectively, via LAN cables.
- the plurality of client computers 300, 300a, 300b, 300c have web browsers 310, 310a, 310b, 310c, respectively.
- the web browsers 310, 310a, 310b, 310c output a content acquisition request provided by the WWW server 110 in response to an operation input from a user.
- a trunking mechanism 320 and an NIC 331 are provided in the client computer 300.
- the trunking mechanism 320 is located between the web browser 310 and the NIC 331. Is provided.
- the trunking mechanism 320 transmits a frame including a content acquisition request output from the web browser 310 via the NIC 331.
- the trunking mechanism section 320 has a distributed algorithm table 321.
- the distribution algorithm table 321 defines an algorithm for distributing the load of data transmitted to the server computer 100 to a plurality of NICs 131-134.
- the distribution algorithm table 3221 is set based on the distribution algorithm notification bucket distributed by broadcasting from the trunking mechanism section 120 of the server computer 100.
- the trunking mechanism section 320 determines a destination MAC address of a frame to be transmitted to the server computer 100 based on the distributed algorithm table 321. For example, the MAC address is determined by the round robin method so that frame transmission is performed equally to each of a plurality of NICs 131 to 134.
- the NIC 331 is connected to the LAN port 211 of the switcher hub 200 via a LAN cable, and transmits and receives frames to and from the switching haptic 200.
- the MAC address of NIC331 is "x".
- FIG. 3 is a diagram showing the internal configuration of the trunking mechanism.
- the trunking mechanism 120 of the server computer 100 includes a distribution algorithm table 121, a distribution algorithm notification unit 122, a transmission data load distribution processing unit 123, and a reception data transfer unit 124. Have.
- the distribution algorithm table 122 defines a load distribution algorithm to be applied to the overnight communication with the client computer 300.
- the distributed algorithm notifying unit 122 refers to the distributed algorithm table 122 to the client computer 300, and transmits the data transmitted from the client computer 300 to the server computer 100. Notify the distribution algorithm for.
- the distribution algorithm is notified by a distribution algorithm notification packet 20.
- the transmission data overnight load distribution processing unit 1 23 refers to the distribution algorithm table 1 2 1, and executes a distribution algorithm according to the client computer 3 0 0 0 of the transmission destination. decide.
- the transmission data load distribution processing unit 123 distributes the transmission data 111 to the plurality of NICs 131 to 134 according to the determined distribution algorithm and transmits the data.
- the reception data transfer section 124 transfers the data received from the client computer 300 to the WWW server 110.
- the reception data transfer unit 124 when the reception data is divided and transferred by distributed processing and is input from a plurality of NICs 131 to 134, concatenates the divided data. Pass it to the WWW server 110.
- the trunking mechanism section 320 of the client computer 300 includes a distributed algorithm table 321, a distributed algorithm acquisition section 322, a transmission data load distribution processing section 323, and a reception data transfer section 32. Has four.
- the distributed algorithm template 3 2 1 defines a load distribution algorithm applied to data communication between the server computer 100 and the server.
- the distributed algorithm acquisition unit 3222 receives the distributed algorithm notification packet 20 sent from the server computer 100, and updates the contents of the distributed algorithm table 3221.
- the transmission data load distribution processing unit 3 2 3 Upon receiving the transmission data 311, the transmission data load distribution processing unit 3 2 3 refers to the distributed algorithm table 3 2 1, and executes a distribution algorithm according to the server computer 100 to be transmitted. To determine. Then, the transmission data load distribution processing unit 3 23 3 sends the destination MAC addresses of the plurality of data constituting the transmission data 311 to the plurality of NICs 13 1 to 13 4 in accordance with the determined distribution algorithm. Distribute and send.
- the receiving data transfer unit 3 2 4 transfers the data received from the server computer 100 to the web browser 310. Note that the receiving data transfer unit 3 2 4 concatenates the divided data when the received data is transferred in a divided manner by distributed processing and transferred via multiple NICs 13 1 to 13 4. And pass it to the web browser 310.
- FIGS. 2 and 3 only the configuration of the client computer 300 is shown in detail, but the other client computers 300 a, 300 b, and 300 c are not shown. Has the same configuration.
- a content acquisition request from the Web browsers 310, 310a, 310b, 310c of a plurality of client computers 300, 300a, 300b, 300c is transmitted to the server via the switching hub 200. Sent to Computer Evening 100. Then, the requested content is distributed by the WWW server 110 in the server computer 100.
- the server computer 100 and the switching hub 200 are connected by a plurality of LAN cables, it is possible to obtain a bandwidth four times as large as that of a case where they are connected by one transmission path. For example, if one transmission path has a bandwidth of 100 Mbps, communication between the server computer 100 and the switching hub 200 at a speed of 400 Mbps becomes possible. As a result, it is possible to withstand the high load accompanying simultaneous access to the server computer 100 from a plurality of client computers 300, 300a, 300b, 300c.
- the data transmitted from the server computer 100 to the client computers 300, 300a, 300b, 300c is distributed to the NICs 131 to 134 in a well-balanced manner by the trunking mechanism unit 120, so that the amount of communication on one line Is not biased. Also, since the destination of the data transmitted from the client computer 300 to the server computer 100 is distributed to the NICs 131 to 134 in a well-balanced manner by the trunking mechanism 320, the communication volume may be unevenly distributed to one line. Absent.
- FIG. 4 is a diagram illustrating an example of a hardware configuration of a server computer.
- the entire server computer 100 is controlled by a CPU (Central Processing Unit) 101.
- the CPU 101 has a random access memory (RAM) 102, a hard disk drive (HDD) 103, a graphics processing unit 104, an input interface 105, and a plurality of NICs 131 via a bus 107. To 134 are connected.
- RAM random access memory
- HDD hard disk drive
- 104 graphics processing unit
- input interface 105 input interface
- NICs 131 via a bus 107.
- To 134 are connected.
- the RAM 102 temporarily stores at least a part of an OS (Operating System) program and an application program to be executed by the CPU 101.
- the RAM 102 stores various data required for processing by the CPU 101. It is.
- the HDD 103 stores an OS and application programs.
- the monitor 11 is connected to the graphic processing device 104.
- the graphics processing device 104 displays an image on the screen of the monitor 11 in accordance with an instruction from the CPU 101.
- a keyboard 12 and a mouse 13 are connected to the input interface 105.
- the input interface 105 transmits signals transmitted from the keyboard 12 and the mouse 13 to the CPU 101 via the bus 107.
- the NICs 131 to 134 are connected to the switching hub 200, respectively.
- FIG. 4 shows the hardware configuration of the server computer 100
- the client computers 300, 300a, 300b, and 300c can be realized with the same hardware configuration.
- the client computers 300, 300a, 300b, and 300c need only have one NIC.
- an example of communication between the server computer 100 and the client computer 300 will be described as a process for performing the communication.
- FIG. 5 is a flowchart showing a data communication processing procedure.
- the processing of the server computer 100 is shown on the left side, and the processing of the client computer 300 is shown on the right side.
- the processing illustrated in FIG. 5 will be described along with step numbers.
- Step S11 The client computer 300 waits for the reception of the distributed algorithm notification bucket 20.
- Step S12 The server computer 100 transmits a distribution algorithm notification packet 20 in a broadcast frame at the time of system startup or the like.
- Step S13 The server computer 100 then waits for a response from the client computer 300.
- Step S14 The client computer 300 determines whether or not a distributed algorithm notification bucket has been received. If a distributed algorithm notification bucket has been received, the process proceeds to step S15. If the distributed algorithm notification packet has not been received, the process proceeds to step S11, and waits for the reception of the distributed algorithm notification packet 20. [Step SI 5] The client computer 300 extracts information on the distribution algorithm from the distribution algorithm notification packet 20 sent from the server computer 100, and takes it into the distribution algorithm table 32 1.
- Step S16 The client computer 300 transmits a response to the distributed computer notification packet to the server computer 100.
- Step S17 The server computer 100 determines whether a response has been received within a predetermined time. If a response is received within the fixed time, the process proceeds to step S18. If no response has been received within a certain period of time, the process proceeds to step S12, and the distributed algorithm notification packet 20 is transmitted again.
- Step S18 The server computer 100 confirms the response content, and adds information including the MAC address of the communication partner to the distributed algorithm table 122.
- Step S 19 The server computer 100 starts communication with the client computer 300.
- Step S200 The client computer 300 starts communication with the server computer 100.
- Step S21 The server computer 100 transmits data according to a distributed algorithm.
- Step S 222 The client computer 300 transmits data according to the distributed algorithm.
- the distribution algorithm notification bucket 1 is transmitted from the server computer 100 to the client computer 300, and the distribution algorithm is notified.
- the client computer 300 recognizes the configuration of the NIC in the server computer 100, it is possible to perform data transmission in which the load is distributed to those NICs.
- FIG. 6 is a conceptual diagram illustrating a distributed algorithm notification process.
- a distributed algorithm notification bucket 2 0 Prior to communication with the client computers 300, 300a, 300b, and 300c, a distributed algorithm notification bucket 2 0 is generated, placed on an Ethernet (registered trademark) frame via an arbitrary NIC, and transmitted by broadcast.
- the distribution algorithm notification packet Unit 20 is passed to each client computer 300, 300a, 300b, 300c. For example, if it is desired to perform reception distribution using the round robin method, the distribution algorithm notification bucket 20 is transmitted so that the data destined for itself transmits the NICs 131 to 134 using the round robin method.
- a response packet is transmitted to the server computer 100 from the client computers 300, 300a, 300b, 300c that have received the distributed algorithm notification packet 20.
- FIG. 7 is a diagram showing a frame configuration for a distributed algorithm notification packet.
- the frame 30 includes a partner MAC address 31, a local MAC address 32, a protocol identifier 33, a command part 34, and a data part 35.
- the destination MAC address 31 is a setting area of a MAC address indicating a frame transmission destination.
- a value rFF-FF-FF-FF-FF'J indicating the broadcast transfer is set.
- the own MAC address 32 is set to the MAC address of the NIC 100 itself.
- the server computer 100 sets its own MAC address 32 to the MAC address of the selected NIC for selecting an arbitrary NIC and transmitting the distributed algorithm notification bucket 20.
- an identifier indicating a distributed algorithm notification bucket is set.
- the client computer 300, 300a, 300b, 300c recognizes that the received packet is the distributed algorithm notification bucket 20 by checking the protocol identifier.
- the type of the information packet is set. Types include publicity, follow-up, delete, and response.
- the packet type “publicity” indicates that it is a publicity (broadcast) of distributed algorithm data.
- a distributed algorithm notification packet 20 in which “public information” is set in the command section 24 is transmitted.
- the frame in which “Public information” is set in the command section 34 is transmitted from the server computer 100 to the client computers 300, 300a, 300b, and 300c.
- the packet type “addition” indicates that this is an instruction to add distributed algorithm data.
- the frame in which “Add” is set in the command section 34 is transmitted from the server computer 100 to the client computers 300, 300a, 300b, and 300c.
- the packet type “delete” indicates that this is an instruction to delete the distributed algorithm data.
- the frame set to “delete” in the command part 34 is transmitted from the server computer 100 to the client computers 300, 300a, 300b, and 300c.
- the packet type “Response” indicates that it is a response to a publicity, addition, or deletion instruction.
- the frame in which “response” is set in the command section 34 is transmitted from the client computer 300, 300a, 300b, 300c to the super computer 100.
- the content of the distributed algorithm day is set.
- the node name of the server computer 100 the node name of the client computer 300, 300a, 300b, 300c, the application name, the distribution algorithm, the MA of the NIC 131-134 of the server computer 100 Information such as the C address is set in the data section 35.
- the server computer 100 has a class configuration composed of a plurality of computers, the node name of the computer on which the NIC is installed is added as additional information to the MAC addresses of the NICs 131 to 134. Is done.
- the content of the data section 35 has a list structure for each of the client computers 300, 300a, 300b, and 300c, and each client computer 300, 300a, 300b, and 300c has only information corresponding to itself. And set it in the distributed algorithm table.
- each NIC 13 1 to 13 4 has a unique network address called a MAC address. Therefore, each of the client computers 300, 300a, 300b, and 300c is connected to the four NICs 131-1 to 34 that were notified from the server computer 100.
- the address information is notified to the client computer 300, 300a, 300b, 300c using an Internet protocol such as ARP (Address Resolution Protocol).
- ARP Address Resolution Protocol
- Only one MAC address can be notified.
- ARP cannot report distributed algorithm information.
- only one NIC can be used at a time when data is transmitted from each of the client computers 300, 300a, 300b, and 300c to the server computer 100.
- NIC can be used at a time when data is transmitted from each of the client computers 300, 300a, 300b, and 300c to the server computer 100.
- Only one of the four received data is used and no trunking is performed. In the first embodiment, even in such a case, load distribution by trunking is possible.
- FIG. 8 is a diagram showing an example of the contents of the data portion of the distributed algorithm notification bucket.
- FIG. 8 shows an example of the data section 35 in a case where the file transfer by the file transfer protocol (ftp) is load-balanced by round robin.
- ftp file transfer protocol
- the own node name, the partner node name, the application name, the algorithm used, and the MAC address of the used NIC are set for each partner node name.
- the own node name is “hostA”.
- the partner node names are "hostB”, “hostC”, “hostD”, and “hostE”.
- the application name is “ftp”.
- the algorithm used is "round robin”.
- the MAC addresses “a”, “b”, “c j,” “d” of the NICs 13 1 to 13 4 are set.
- a distributed algorithm table is constructed for both of the 00 and 00c, and load distribution is performed according to the distributed algorithm table.
- FIG. 9 is a diagram showing an example of a data structure of a distributed algorithm table on the server computer side.
- the distributed algorithm table 121 includes columns of the own node name, the other node name, the application name, the algorithm, the own NIC :, the other NIC, and the flag.
- the information arranged in the horizontal direction in each column is associated with each other to form a record.
- the own node name is the node name of the server computer 100 (for example, IP address).
- the partner node name is the node name (for example, IP address) of the client computer 300, 300a, 300b, 300c of the partner of communication.
- the application name is the name of the application to be distributed (telnet'ftp, etc.).
- the algorithm is a distributed algorithm (for example, round robin-. Fixed N) used to send and receive data for the application specified by the application name.
- the own NIC is the MAC address of the NIC 131 to L34 of the server computer 100. If there are multiple own NICs, they are shown in a list separated by commas.
- the other party's NIC is the MAC address of the other party's client computer 300, 300a, 300b, 300c (if there is more than one, it is stored in a list format).
- the flag indicates that the table information of the record is valid or invalid. If the flag is enabled, it indicates that the exchange of the distributed algorithm has been completed and communication is possible. If the flag is invalid, it indicates that the exchange of the distributed algorithm is incomplete and communication is not possible.
- FIG. 9 shows the distributed algorithm table 122 on the server computer 100 side
- the distributed algorithm table 321 on the client computer 300 side has the same data structure. Since the client computer 300 has only one NIC 331, one MAC address is set for its own NIC, and the NICs 13 1 to 1 3 4 in the server computer 100 are set for the partner NIC. MAC address is set.
- FIG. 10 is a diagram showing an example of registration of the distribution algorithm table according to the distribution algorithm notification bucket.
- a distribution algorithm notification bucket 20 a for public information is transmitted from the server computer 100 to the client computer 300.
- the own node name is rhostAJ
- the partner node name is rhostBj
- the application name is rftpj
- the algorithm is “round robin”
- the own NIC is “a, b, c, d”.
- the client computer 300 Upon receiving the distributed algorithm notification packet 20a, the client computer 300 exchanges its own node name and the remote node name in the distributed algorithm notification bucket 20a. 2 Register in 1. Next, the client computer 300 sets the MAC address “x” of its NIC 331 in its own NIC of the registered record, and sets the flag to “valid”. Then, the client computer 300 sends to the server computer 100 a distributed algorithm notification packet 20b of a response including the MAC address "of its own NIC 31.
- the server computer 100 When the server computer 100 receives the distributed algorithm notification packet 20 b of the response, the server computer 100 stores a record obtained by adding the partner NIC “X” and the flag “valid” to the contents of the distributed algorithm notification packet 20 a. Register in 1 2 1
- FIG. 11 is a diagram illustrating a load distribution processing state of data transmitted from the server computer.
- load distribution is performed according to the round robin distribution algorithm.
- data is allocated to a plurality of NICs 131 to 134 in a predetermined order. For example, NIC 131, NIC 132, NIC 133, NI C 134. After the NIC 134, data is allocated to the NIC 131.
- the transmission data 111 sent to the client computer 300 from the WWW server 110 is passed to the trunking mechanism unit 120.
- the transmission data 111 is transmitted after being divided into six data.
- the divided data is referred to as “D 1”, “D 2”, “D 3 J,” “D 4”, “D 5”, and “D 6”.
- the transmission data 111 is distributed by the trunking mechanism unit 120 to a plurality of NICs 131 to 134 in a round robin manner.
- the data of “D 1” is transmitted to the client computer 300 via the NIC 131.
- the data of “D 2” is transmitted to the client computer 300 via the NIC 132.
- the data of “D3” is transmitted to the client computer—evening 300 via the NIC 133.
- the data of “D4” is transmitted to the client computer 300 via the NIC 134.
- the data of “D5” is transmitted to the client computer 300 via the NIC 131.
- the data of “D 6” is transmitted to the client computer 300 via the NIC 136.
- the transmission data 111 from the server computer 100 destined for the client computer 300 is distributed to the switching hap 200 while distributing the load on the transmission path, and then transmitted to the client computer 300.
- FIG. 12 is a diagram illustrating a state of load distribution processing performed overnight from a client computer.
- load distribution is performed according to the round robin distribution algorithm.
- the transmission data 311 addressed to the WWW server 110 from the client computer 300 is passed to the trunking mechanism section 320.
- the transmission data 31 1 is transmitted after being divided into four pieces of data.
- the divided data are referred to as “D11”, “D12”, “D13”, and “D14”.
- the transmission data 311 is distributed to a plurality of NICs 13 1 to 13 4 in a round-robin manner by the trunking mechanism 3 20.
- the MAC address of the target NIC is set to the destination MAC address of the frame that transmits the data.
- the switching hub 200 refers to the destination MAC address to determine the LAN port to be transmitted, so that each data is transmitted to the allocated NIC.
- the data of “D11” is transmitted to the NIC 131 of the server computer 100.
- the data of “D 1 2” is transmitted to NIC 1 32 of server computer 100.
- the data of “D 13” is transmitted to NIC 133 of server computer 100.
- the data of D 14 j is transmitted to NIC 134 of server computer 100.
- the transmission data 311 from the client computer 300 to the server computer 100 is distributed to a plurality of transmission paths via the switcher hub 200 to distribute the load to the server computer 100. Passed to 0.
- the switching hub 200 does not have a trunking mechanism, the load is distributed to a plurality of transmission paths between the server computer 100 and the switching hub 200.
- the overnight communication is allowed. Therefore, when building a system using trunking, any switching haptic that does not support trunking can be used, and the range of options can be expanded.
- each of the client computers 300, 300a, 300b, and 300c has distributed algorithm information, even if only one-to-one communication is performed between two devices, the Transmission / reception dispersion by ranking can be realized, and bandwidth improvement equivalent to that when using a trunking compatible switch can be realized.
- the load distribution is managed in the client computers 300, 300a, 300b, and 300c, a distribution algorithm according to the application can be specified.
- the switching hub 200 performs the trunking process as in the conventional case, the switching hub functions in the data link layer of the OSI hierarchical model, so that the application cannot be determined, and the application is not performed. Corresponding trunking was not possible.
- the client computer 300 can specify the distributed algorithm according to the application, A highly flexible trunking mechanism can be provided.
- the second embodiment is an example in which the present invention is applied to a system in a case where a plurality of server computers operate in cooperation to form a cluster.
- FIG. 13 is a diagram illustrating an example of a system configuration according to the second embodiment.
- two server computers 100a and 100b are connected to each other by an inter-system communication path 40.
- the inter-system communication path 40 is a communication interface that enables higher-speed communication than a LAN. Note that components other than the server computers 100a and 100b are the same as those in the first embodiment, and therefore are denoted by the same reference numerals as in the first embodiment shown in FIG.
- the server computer 100a has a WWW server 110a, a trunking mechanism 120a, and NICs 131a and 132a.
- the trunking mechanism 120a has a distributed algorithm table 121a.
- the server combination 100b has a trunking mechanism 120b and NICs 131b and 132b.
- the trunking mechanism section 120b has a distributed algorithm table 121b.
- each server computer 100a, 100b has two NICs 131a, 132a, 131b, 132b, respectively.
- NICs 131a, 132a, 131b, 132b are required. It cannot be attached.
- the trunking mechanisms 120a and 120b between the server computer 100a and the server computer 100b transmit control information registered in the distributed algorithm tables 121a and 121b for trunking through the inter-system communication path 40. Sharing.
- the trunking mechanism sections 120a and 120b use NICs 131a, 132a, 131b, and 132b, respectively, for the client computers 300, 300a, 300b, and 300c. It has a distributed algorithm table in which trunking control information is registered so as to perform transmission and reception.
- FIG. 14 shows an example of the data structure of the distributed algorithm table according to the second embodiment.
- FIG. The data structure of the distributed algorithm table 121a is the same as the distributed algorithm table 121 of the first embodiment shown in FIG. 9 except for the own NIC.
- the own NIC includes one of the NI C 131a, 132a, 131, and 132b installed on both server computers 100a and 100b, which is used for data transfer of the corresponding application.
- MAC address is set.
- the mounting computer name is set in parentheses.
- the mounted computer name is the name (identification name) of the server computer that implements each NIC.
- the WWW server 110a wants to distribute data in a round-robin manner with four client computers 300, 300a, 300b, and 300c, the server must be configured at system startup. The following processing is performed between the computers 100a and 100b.
- FIG. 15 is a flowchart showing a processing procedure at the time of starting the system according to the second embodiment. Hereinafter, the processing illustrated in FIG. 15 will be described along the step numbers.
- the trunking mechanism unit 120a of the server computer 100a generates a distributed algorithm table 121a as in the first embodiment.
- the trunking mechanism 120a distributes the distributed algorithm notification packet 20 by broadcast to the distributed algorithm template 121a itself, and the own NIC of the distributed algorithm notification bucket 20 has a server computer 1
- the MAC addresses of NICs 131a, 132a, 131b, and 132b between 00a and the server computer 100b are included.
- the node name of the server computer on which the NICs 131a, 132a, 131b, and 132b are mounted is given to the MAC address.
- Step S32 The trunking mechanism unit 120a determines whether or not the server computer 100a forms a cluster in cooperation with another server computer. If it is composed of a plurality of server consoles, the process proceeds to step S33. If it is configured with one server computer 100a, the process proceeds to Step S36.
- Step S33 The trunking mechanism unit 120a uses a dedicated inter-system communication path. Then, the content of the distributed algorithm table 121a is transmitted to the server computer 100b using the source 40.
- Step S34 The trunking mechanism unit 120b of the server computer 100b generates its own distributed algorithm table 121b based on the contents of the distributed algorithm table 121a received from the server computer 100a. I do.
- Step S35 The trunking mechanism unit 120b transmits a response to the server computer 100a.
- Step S36 The server computer 100a starts communication with the client computers 300, 300a, 300b, and 300c.
- Step S37 the server computer 100b starts communication with the client computers 300, 300a, 300b, and 300c.
- trunking can be performed by a class system using a plurality of server computers 100a and 100b.
- the following data transfer is performed by load distribution by trunking.
- FIG. 16 is a diagram showing a data transmission procedure when a plurality of server computers are used. Hereinafter, the processing illustrated in FIG. 16 will be described along the step numbers.
- Step S41 The trunking mechanism 120b of the server computer 100b keeps waiting for an event from the server computer 100a.
- Step S42 The trunking mechanism unit 120a of the server computer 100a keeps waiting for an event from the WWW server 110a.
- Step S43 When any event occurs, the trunking mechanism unit 120a determines whether the occurred event is a data transmission request from the WWW server 110a. If so, the process proceeds to Step S45. If the event is not an overnight transmission request, the process proceeds to step S44.
- Step S44 The trunking mechanism unit 120a executes a process other than overnight transmission according to the event that has occurred. Thereafter, the process proceeds to step S42, and waits for an event again.
- the trunking mechanism section 120a executes a distributed algorithm table.
- File 121a to determine the NIC to use for transmission Specifically, the trunking mechanism unit 120a refers to the distribution algorithm table 121a, and determines a distribution algorithm from a transmission destination, an application name, and the like. Then, according to the determined distribution algorithm, a NIC to be used for transmission is determined.
- the trunking mechanism unit 120a extracts the device name of the server computer on which the NIC used for transmission is mounted, from the distributed algorithm table 121a.
- Step S47 The trunking mechanism unit 120a determines whether the NIC used is the NIC implemented in the server computer 100a itself. If the NIC is mounted on the server computer 100a, the process proceeds to step S49. NI implemented on other server computers. If so, the process proceeds to Step S48.
- Step S48 The trunking mechanism section 120a transfers the data from the WWW server 110a to the other server computer 100b using the inter-system communication path 40. Thereafter, the processing on the server computer 100a side ends, and the processing shifts to step S50 of the trunking mechanism unit 120b on the server computer 100b side.
- Step S49 The trunking mechanism section 120a transmits the data to the client computer via the NIC determined in step S45.
- Step S50 The trunking mechanism unit 120b of the server computer 100b searches its own distributed algorithm table 121b and uses N for transmission.
- Step S51 The trunking mechanism unit 120b transmits the data to the client computer via the NIC determined in step S50.
- FIG. 17 is a diagram illustrating a data transmission situation when a plurality of server computers are used.
- the web application 110a of the server computer 100a sends the data 111a to the trunking mechanism 120a, the trunking mechanism 1
- 20a divides the transmission data 111a into a plurality of data and distributes the data to a plurality of NICs 131a, 132a, 131b, and 132b.
- 1a is divided into four data, and that each data is “D21J,” “D22”, “D23”, and “D24”.
- the transmission data 111 is sorted by the trunking mechanism unit 120 into a plurality of NICs 131a, 132a, 131b, and 132b in a round robin manner.
- the data “D 11” is transmitted to the client convenience 300 via the NIC 131a.
- the data of “D22” is transmitted to the client computer 300 via the NIC 132a.
- the data of “D23” is sent to the trunking mechanism section 120 b of the server computer 100 b via the inter-system communication path 40, and then distributed by the trunking mechanism section 120 b to the client computer via the NIC 131 b. Sent to 300.
- the data of “D 24” is sent to the trunking unit 120 b of the server computer 100 b via the inter-system communication path 40, and is then distributed to the trunking unit 120 b and passed through the NIC 132 b. Is transmitted to the client computer 300.
- the transmission data 111a destined for the client computers 300, 300a, 300b, 300c from the server computers 100a, 100b is distributed to the switching hub 200 after dispersing the load, and thereafter, Handed over to the client combination 300, 300a, 300b, 300c.
- the client computers 300, 300a, 300b, and 300c when performing distributed processing of data transmitted from the client computers 300, 300a, 300b, and 300c to the server computer 100a, the client computers 300, 300a, 300b, and 300c perform communication before the communication.
- the distribution method is determined based on the distribution algorithm notification bucket transmitted from the server computer 100a.
- data sent from the client computers 300, 300a, 300b, 300c to the server computer 100a is distributed and transmitted to the respective NICs 131a, 132a, 131b, 132b.
- the data transmitted to the NICs 131b and 132b is passed to the WWW server 110a via the inter-system communication path 40 by the trunking mechanism 120b.
- trunking process can be performed using the server computer.
- trunking can be realized by increasing the number of used NICs, without depending on the hardware specifications (upper limit of NIC implementation) of the server computer 100a to be used.
- FIG. 18 is a diagram for explaining the principle when one NIC is added.
- the server computer 100c has a WWW server 110c, a trunking mechanism 120c, NICs 131c, 132c, 133c, 134c, 135, and a NIC addition Z detachment control unit 140. I have.
- the trunking mechanism section 210c has a distributed algorithm table 121c.
- the NIC 135 is a newly added one.
- the administrator of the server computer 100c connects the attached spare NIC 135 to the switching hub 200 with a LAN cable.
- switching trunk 200 is not set for trunking, it can be used for communication immediately after connection without resetting switching hap 200. In other words, if a switcher hub that supports trunking is used, it is necessary to temporarily stop the communication of the already connected NIC to change the setting.
- the server computer 100c needs to add the newly added NIC 135 to the trunking target.
- the NIC addition Z separation controller 140 of the server computer 100c automatically recognizes the addition of the NIC 135 and registers information according to the addition of the NIC 135 in the distributed algorithm table 121c.
- the distributed algorithm table 121c can be updated according to an instruction from an application or an operator without waiting for additional automatic recognition of the NIC 135.
- the NIC addition / disconnection control unit 140 notifies the trunking mechanism unit 120 of the content using the inter-program communication.
- the trunking mechanism unit 120c adds the target NIC to the distribution targets in the distribution algorithm table 121c.
- the trunking mechanism The unit 120c notifies the client computer 300, 300a, 300b, 300c of the update content by a distributed algorithm notification packet. Thereafter, the trunking mechanism unit 120c performs load distribution of data transmission including the NIC 135 at the time of data transmission.
- the addition of the NIC 135 is notified to the client computers 300, 300a, 300b, and 300c by the distribution algorithm notification bucket, and the distribution within the client computers 300, 300a, 300b, and 300c is performed according to the notification.
- the algorithm table is updated. Therefore, even for data transmitted from the client computer 300c, 300a, 300b, 300c to the server computer 100c, the load distribution of data transmission including the NIC 135 is distributed according to the updated distribution algorithm table. Done.
- FIG. 19 is a flowchart showing the procedure for updating the distributed algorithm table when adding a NIC. Hereinafter, the processing illustrated in FIG. 19 will be described along the step numbers.
- Step S61 The trunking mechanism section 320 of the client computer 300 waits for reception of the distributed algorithm notification bucket.
- Step S62 The NIC addition Z disconnection control unit 140 of the server computer 100c waits for an event to occur.
- Step S63 NIC addition
- the separation control unit 140 determines the contents of the I / O interrupt. If so, the process proceeds to Step S65. If it is an I / O interrupt other than the addition of NIC, the process proceeds to step S64.
- Step S64 The NIC addition disconnection control unit 140 executes a process corresponding to the I / interrupt, and thereafter ends the process.
- the NIC addition Z separation control unit 140 extracts information on the added NIC 135 from the IZO table.
- the IZO table is a data table that defines information according to I / O interrupts managed by the OS. It is. Then, the NIC addition Z separation control unit 140 instructs the trunking mechanism unit 120c to add the NIC.
- Step S666 The trunking mechanism unit 120c generates a distribution algorithm notification bucket with the added packet type, and transmits the packet to the client computer 300. As a result, the information of the added NIC 135 is notified to the client computer 300.
- Step S67 The trunking mechanism section 120c enters a response reception waiting state.
- Step S688 The trunking mechanism section 320 of the client computer 300 determines whether or not a distributed algorithm notification bucket has been received. If so, the process proceeds to Step S69. If not, the process proceeds to step S61, and waits for reception of a packet.
- Step S69 The trunking mechanism section 320 determines whether the distributed algorithm notification packet is an instruction to add a newly implemented NIC. This can be determined by whether the bucket type is “added” or not. In the case of an additional instruction of NIC, the process proceeds to step S71. In the case of an instruction other than the NIC addition instruction (for example, deletion of NIC)., The process proceeds to Step S70.
- Step S70 The trunking mechanism section 320 executes a process (a process other than the NIC addition) specified by the distributed algorithm notification packet. Thereafter, the processing proceeds to step S61.
- the trunking mechanism section 320 adds the new NIC specified in the distribution algorithm notification packet to its own distribution algorithm table 321.
- Step S72 The trunking mechanism section 320 transmits a response packet to the server computer 100C.
- Step S73 The trunking mechanism section 120c of the server computer 100c determines whether a response has been received within a predetermined time. If there is a response, the process proceeds to step S74. If there is no response, the process proceeds to step S66. ⁇ [Step S74] After confirming the response, the trunking mechanism unit 120c adds information of the newly added NIC to the distributed algorithm table 122c.
- the trunking mechanism section 120c of the server computer 100c and the trunking mechanism section 320 of the client computer 300 include the newly added NIC in the communication path. Start data communication.
- a new NIC can be added. This operation is performed in parallel with the communication of applications such as the WWW server 110c, so that other NICs 131c, 132c, 13 There is no effect on the communication data on 3c and 134c. Moreover, there is no need to reset the settings and restart the switching hub. Therefore, it is possible to dynamically expand the data transfer bandwidth while maintaining communication.
- FIG. 20 is a diagram for explaining the principle when one NIC is deleted.
- the NIC disconnection control unit 140 is notified of the NIC disconnection by an I / O interrupt.
- the operation input from the application software such as Sano computer 100 c,
- the NIC addition / separation control unit 140 that has received the notification of the NIC separation updates the internal system configuration information, and notifies the content to the trunking mechanism unit 120c using the inter-program communication.
- the trunking mechanism section 120c that has received the notification removes the information of the corresponding NIC135 from the distributed algorithm table 122c so as to exclude the information from the trunking target. Thereafter, transmission dispersion is performed from the server computer 100c side without using the NIC135.
- NIC 135 is removed from the trunking mechanism section 120 C to the client convenience server 300, 300 a, 300 b, and 300 c.
- Send the updated distribution algorithm notification packet (packet type is deleted).
- the trunking mechanism of each of the client computers 300, 300a, 300b, and 300c reads the NIC from the internal distributed algorithm table. Delete 135 information.
- NICs 131c, 132c, 133c, 134c other than the NIC 135 are used as data transmitted from the client computers 300, 300a, 300b, 300c to the server computer 100c.
- NICs 131c, 132c, 133c, 134c other than the NIC 135 are used.
- the connection cable is disconnected from the corresponding LAN port of the switching hub 200, and the deletion of the NIC 135 is completed.
- This operation is performed in parallel with the communication of the application, and does not affect the communication data on other NICs that have been exchanged. Also, since there is no need to reset and restart the settings of the switch hub 200 side, the data transfer bandwidth can be dynamically reduced while the communication is continued. is there.
- the addition and removal of the NIC can be performed without stopping the operation of the server computer 100c, so that the bandwidth adjustment between the server computer 100c and the switching haptic 200 is facilitated.
- NIC can be added to increase the bandwidth while continuing operation.
- a server program describing the processing contents of the functions that the server computer should have, and a client program describing the processing contents of the functions that the client computer should have are provided. You. By executing the server program on a computer, the processing functions of the server computer according to the above embodiment are realized. Further, by executing the client program on a computer, the processing functions of the client computer according to the above-described embodiment are realized.
- Computer-readable recording media include magnetic recording devices, optical disks, magneto-optical recording media, and semiconductor memories.
- Magnetic recording devices include hard disk drives (HDDs), flexible disks (FDs), and magnetic tapes.
- Optical discs include DV D (Digital Versatile Disc).
- DVD-R AM Random Access Memory
- CD-ROM Compact Disc Read Only Memory
- CD-R Recordable
- RW Re Writable
- Magneto-optical recording media include M ⁇ (Magneto-Optical disc).
- the client program can be stored in the storage device of the server computer, and the client program can be transferred from the server computer to the client computer via the network.
- the server computer that executes the server program stores, for example, the server program recorded on a portable recording medium in its own storage device. Then, the server computer reads the server program from its own storage device and executes processing according to the server program. The server computer can also read the server program directly from the portable recording medium and execute processing according to the server program.
- the client computer that executes the client program stores, for example, the client program recorded on the portable recording medium or the client program transferred from the server computer in its own storage device. Then, the client computer reads the client program from its own storage device and executes processing according to the client program. Note that the client computer can also read the client program directly from the portable recording medium and execute processing according to the client program. Further, each time the client program is transferred from the server computer, the client computer can sequentially execute processing according to the received client program.
- the physical address of a plurality of communication interfaces and the distribution algorithm are notified to the communication partner computer by the distribution algorithm notification packet, and the communication interface determined in accordance with the distribution algorithm at the partner computer.
- This enables trunking using multiple transmission paths, even if the device that relays the packet does not have a trunking function.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Small-Scale Networks (AREA)
- Computer And Data Communications (AREA)
Description
Claims
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB038263017A CN100474833C (zh) | 2003-03-31 | 2003-03-31 | 数据通信负荷分散控制、负荷分散数据发送方法及装置 |
JP2004570184A JP3792707B2 (ja) | 2003-03-31 | 2003-03-31 | データ通信負荷分散制御プログラム、およびデータ負荷分散制御方法 |
AU2003221109A AU2003221109A1 (en) | 2003-03-31 | 2003-03-31 | Data communication load distribution control program and data load distribution control method |
DE2003194206 DE10394206T5 (de) | 2003-03-31 | 2003-03-31 | Datenkommunikations-Lastverteilungs-Steuerprogramm und Datenlastverteilungs-Steuerverfahren |
PCT/JP2003/004148 WO2004088931A1 (ja) | 2003-03-31 | 2003-03-31 | データ通信負荷分散制御プログラム、およびデータ負荷分散制御方法 |
US11/231,467 US8068498B2 (en) | 2003-03-31 | 2005-09-21 | Computer readable record medium on which data communication load distribution control program is recorded and data load distribution control method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2003/004148 WO2004088931A1 (ja) | 2003-03-31 | 2003-03-31 | データ通信負荷分散制御プログラム、およびデータ負荷分散制御方法 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US11/231,467 Continuation US8068498B2 (en) | 2003-03-31 | 2005-09-21 | Computer readable record medium on which data communication load distribution control program is recorded and data load distribution control method |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2004088931A1 true WO2004088931A1 (ja) | 2004-10-14 |
Family
ID=33105370
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2003/004148 WO2004088931A1 (ja) | 2003-03-31 | 2003-03-31 | データ通信負荷分散制御プログラム、およびデータ負荷分散制御方法 |
Country Status (6)
Country | Link |
---|---|
US (1) | US8068498B2 (ja) |
JP (1) | JP3792707B2 (ja) |
CN (1) | CN100474833C (ja) |
AU (1) | AU2003221109A1 (ja) |
DE (1) | DE10394206T5 (ja) |
WO (1) | WO2004088931A1 (ja) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2007094520A1 (ja) * | 2006-02-16 | 2007-08-23 | Nec Corporation | ノード、ネットワークシステム、フレーム転送方法及びフレーム転送プログラム |
JP2007243467A (ja) * | 2006-03-07 | 2007-09-20 | Fujitsu Ltd | ブレード型システムおよびコネクション管理方法 |
JPWO2007029297A1 (ja) * | 2005-09-02 | 2009-03-26 | 富士通株式会社 | ネットワークインタフェース制御プログラムおよびネットワークインタフェース制御装置 |
JP2018521393A (ja) * | 2015-05-26 | 2018-08-02 | アイデバイシーズ エルエルシー | サーバフェイルオーバー及びロードバランシングのためのシステム及び方法 |
US11973823B1 (en) * | 2023-01-11 | 2024-04-30 | Dell Products L.P. | Offloading namespace redirection to backup clients in a scale out cluster |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070237084A1 (en) * | 2006-03-29 | 2007-10-11 | Rothman Michael A | Data communication using multiple input/output devices |
US7649892B2 (en) * | 2006-08-30 | 2010-01-19 | Hewlett-Packard Development Company, L.P. | Method and system of network communication receive load balancing |
US7813286B2 (en) * | 2006-08-30 | 2010-10-12 | Hewlett-Packard Development Company, L.P. | Method and system of distributing multicast group join request in computer systems operating with teamed communication ports |
WO2012050071A1 (ja) | 2010-10-14 | 2012-04-19 | 日本電気株式会社 | 通信システム、制御装置、処理規則の設定方法およびプログラム |
US8694618B2 (en) * | 2011-04-13 | 2014-04-08 | Microsoft Corporation | Maximizing data transfer through multiple network devices |
US8627412B2 (en) | 2011-04-14 | 2014-01-07 | Microsoft Corporation | Transparent database connection reconnect |
CN103297354B (zh) * | 2012-03-02 | 2017-05-03 | 日电(中国)有限公司 | 服务器互连系统、服务器和数据转发方法 |
US11050746B2 (en) | 2019-01-29 | 2021-06-29 | Cisco Technology, Inc. | Media access control (MAC) address anonymization based on allocations by network controller elements |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07245619A (ja) * | 1994-03-03 | 1995-09-19 | Hitachi Ltd | Lanシステムの制御方法 |
JPH1174927A (ja) * | 1997-07-08 | 1999-03-16 | Hewlett Packard Co <Hp> | リンクレベルサーバ/スイッチトランキング方法 |
Family Cites Families (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5617417A (en) * | 1994-09-07 | 1997-04-01 | Stratacom, Inc. | Asynchronous transfer mode communication in inverse multiplexing over multiple communication links |
US5608733A (en) * | 1994-11-29 | 1997-03-04 | Valle; Richard | ATM inverse multiplexing |
US6016319A (en) * | 1995-10-31 | 2000-01-18 | Lucent Technologies, Inc. | Communications system for transmission of datagram packets over connection-oriented networks |
US6016310A (en) * | 1997-06-30 | 2000-01-18 | Sun Microsystems, Inc. | Trunking support in a high performance network device |
US6046979A (en) * | 1998-05-04 | 2000-04-04 | Cabletron Systems, Inc. | Method and apparatus for controlling the flow of variable-length packets through a multiport switch |
US6078960A (en) * | 1998-07-03 | 2000-06-20 | Acceleration Software International Corporation | Client-side load-balancing in client server network |
US6952401B1 (en) * | 1999-03-17 | 2005-10-04 | Broadcom Corporation | Method for load balancing in a network switch |
US6490632B1 (en) | 1999-03-18 | 2002-12-03 | 3Com Corporation | High performance load balancing and fail over support of internet protocol exchange traffic over multiple network interface cards |
US6560630B1 (en) * | 1999-03-18 | 2003-05-06 | 3Com Corporation | Receive load balancing and fail over with multiple network interface cards |
US6553028B1 (en) * | 1999-04-30 | 2003-04-22 | Cisco Technology, Inc. | Method and apparatus for multicast switching using a centralized switching engine |
US6874147B1 (en) * | 1999-11-18 | 2005-03-29 | Intel Corporation | Apparatus and method for networking driver protocol enhancement |
JP2001292153A (ja) | 2000-04-05 | 2001-10-19 | Hitachi Cable Ltd | スイッチングハブ |
US20030028644A1 (en) | 2001-08-02 | 2003-02-06 | Patrick Maguire | System and method for load sharing within a core network |
US7290059B2 (en) * | 2001-08-13 | 2007-10-30 | Intel Corporation | Apparatus and method for scalable server load balancing |
JP3898498B2 (ja) * | 2001-12-06 | 2007-03-28 | 富士通株式会社 | サーバ負荷分散システム |
US20040151111A1 (en) * | 2003-01-31 | 2004-08-05 | Yarroll La Monte | Resource pooling in an Internet Protocol-based communication system |
-
2003
- 2003-03-31 DE DE2003194206 patent/DE10394206T5/de not_active Withdrawn
- 2003-03-31 CN CNB038263017A patent/CN100474833C/zh not_active Expired - Fee Related
- 2003-03-31 WO PCT/JP2003/004148 patent/WO2004088931A1/ja active Application Filing
- 2003-03-31 AU AU2003221109A patent/AU2003221109A1/en not_active Abandoned
- 2003-03-31 JP JP2004570184A patent/JP3792707B2/ja not_active Expired - Fee Related
-
2005
- 2005-09-21 US US11/231,467 patent/US8068498B2/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07245619A (ja) * | 1994-03-03 | 1995-09-19 | Hitachi Ltd | Lanシステムの制御方法 |
JPH1174927A (ja) * | 1997-07-08 | 1999-03-16 | Hewlett Packard Co <Hp> | リンクレベルサーバ/スイッチトランキング方法 |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPWO2007029297A1 (ja) * | 2005-09-02 | 2009-03-26 | 富士通株式会社 | ネットワークインタフェース制御プログラムおよびネットワークインタフェース制御装置 |
WO2007094520A1 (ja) * | 2006-02-16 | 2007-08-23 | Nec Corporation | ノード、ネットワークシステム、フレーム転送方法及びフレーム転送プログラム |
JPWO2007094520A1 (ja) * | 2006-02-16 | 2009-07-09 | 日本電気株式会社 | ノード、ネットワークシステム、フレーム転送方法及びフレーム転送プログラム |
US8094584B2 (en) | 2006-02-16 | 2012-01-10 | Nec Corporation | Node, network system, frame transfer method, and frame transfer program |
CN101385284B (zh) * | 2006-02-16 | 2014-07-02 | 日本电气株式会社 | 节点、网络系统、帧传送方法和帧传送程序 |
JP2007243467A (ja) * | 2006-03-07 | 2007-09-20 | Fujitsu Ltd | ブレード型システムおよびコネクション管理方法 |
JP4579850B2 (ja) * | 2006-03-07 | 2010-11-10 | 富士通株式会社 | ブレード型システムおよびコネクション管理方法 |
JP2018521393A (ja) * | 2015-05-26 | 2018-08-02 | アイデバイシーズ エルエルシー | サーバフェイルオーバー及びロードバランシングのためのシステム及び方法 |
US11973823B1 (en) * | 2023-01-11 | 2024-04-30 | Dell Products L.P. | Offloading namespace redirection to backup clients in a scale out cluster |
Also Published As
Publication number | Publication date |
---|---|
US20060013224A1 (en) | 2006-01-19 |
CN100474833C (zh) | 2009-04-01 |
JPWO2004088931A1 (ja) | 2006-07-06 |
DE10394206T5 (de) | 2006-03-30 |
US8068498B2 (en) | 2011-11-29 |
AU2003221109A1 (en) | 2004-10-25 |
JP3792707B2 (ja) | 2006-07-05 |
CN1765084A (zh) | 2006-04-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8068498B2 (en) | Computer readable record medium on which data communication load distribution control program is recorded and data load distribution control method | |
JP4236303B2 (ja) | リンクレベルサーバ/スイッチトランキング方法 | |
CN100544342C (zh) | 存储系统 | |
US7996560B2 (en) | Managing virtual ports in an information processing system | |
US7437477B2 (en) | SCSI-based storage area network having a SCSI router that routes traffic between SCSI and IP networks | |
US7751409B1 (en) | Logical service domains for enabling network mobility | |
US8560627B2 (en) | Virtual switch for use in fibre channel applications | |
US8578053B2 (en) | NAS load balancing system | |
US20030154236A1 (en) | Database Switch enabling a database area network | |
JP2018523932A (ja) | 負荷バランシングコンピュータデバイス、システム、および方法 | |
US7254620B2 (en) | Storage system | |
US20110007666A1 (en) | Digital multimedia network with parameter join mechanism | |
JP2003318949A (ja) | Vpnシステムおよびルータ | |
JP2004192305A (ja) | iSCSIストレージ管理方法及び管理システム | |
JP2011507426A (ja) | 複数のアダプタにわたり複数の仮想ipアドレスを同時にサポートしているホストにおけるフェイルオーバのための方法、システム、およびプログラム | |
JP2006510976A5 (ja) | ||
WO2019144965A1 (zh) | 一种镜像拉取的方法及其系统 | |
JP5034495B2 (ja) | ストレージシステムとプログラム並びに方法 | |
CN110830574B (zh) | 一种基于docker容器实现内网负载均衡的方法 | |
US20070233886A1 (en) | Method and system for a one bit TCP offload | |
CN100414936C (zh) | 平衡网络文件系统服务器多网卡间负载的方法 | |
JP2004048125A (ja) | 接続制御装置、方法及びプログラム | |
JP4309321B2 (ja) | ネットワークシステムの運用管理方法及びストレージ装置 | |
JP3609948B2 (ja) | マルチプロトコルネットワーク管理方法、マルチプロトコルネットワーク管理プロキシサーバシステム、マルチプロトコルアドレス管理サーバシステム、および、マルチプロトコルネットワーク管理システム | |
JPH10135982A (ja) | サーバコンピュータにおけるipアドレス多重化方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NI NO NZ OM PH PL PT RO RU SC SD SE SG SK SL TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LU MC NL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2004570184 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 11231467 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1020057018339 Country of ref document: KR |
|
WWE | Wipo information: entry into national phase |
Ref document number: 20038263017 Country of ref document: CN |
|
WWP | Wipo information: published in national office |
Ref document number: 1020057018339 Country of ref document: KR |
|
WWP | Wipo information: published in national office |
Ref document number: 11231467 Country of ref document: US |
|
RET | De translation (de og part 6b) |
Ref document number: 10394206 Country of ref document: DE Date of ref document: 20060330 Kind code of ref document: P |
|
122 | Ep: pct application non-entry in european phase | ||
REG | Reference to national code |
Ref country code: DE Ref legal event code: 8607 |