WO2015194072A1

WO2015194072A1 - Relay apparatus, communication apparatus, management apparatus, relay method and relay program

Info

Publication number: WO2015194072A1
Application number: PCT/JP2015/000600
Authority: WO
Inventors: 亮仁小比賀
Original assignee: 日本電気株式会社
Priority date: 2014-06-18
Filing date: 2015-02-10
Publication date: 2015-12-23
Also published as: JPWO2015194072A1

Abstract

A relay apparatus for transferring packets comprises: a connection information storage unit (31) that stores connection information constituted by the numbers for identifying time slots that are sections as divided at least at predetermined time intervals and by the numbers of output ports in the relay apparatus (30); a connection information processing unit (32) that determines whether the connection information corresponding to a time slot to which the current time belongs exists in the connection information storage unit (31); and a packet transfer unit (33) that, if the connection information processing unit (32) determines that the corresponding connection information exists in the connection information storage unit (31), transfers a packet to the output port indicated by the output port number in the corresponding connection information.

Description

Relay device, communication device, management device, relay method, and relay program

The present invention relates to a relay device, a communication device, a management device, a relay method, and a relay program.

Patent Document 1 describes an example of a synchronous scheduling system (first page of Patent Document 1, FIG. 1, etc.). FIG. 28 is a block diagram showing the configuration of the synchronous scheduling system described in Patent Document 1.

As shown in FIG. 28, the synchronous scheduling system described in Patent Document 1 includes a parallel computer 1 including a PE (Processor Element) communication network 2 and a PE 3. The PE 3 includes an inactivation function 4, a condition setting function 5, an allocation function 6, an interrupt generation function 7, and an activation function 8.

The synchronous scheduling system shown in FIG. 28 operates as follows. When the parallel process being executed enters the parallel synchronization wait state, the inactivation function 4 deactivates the parallel process that has entered the wait state and prohibits the parallel process from being allocated.

Allocation function 6 allocates processes of other jobs that can be executed. When the condition set by the condition setting function 5 is satisfied during execution of another job to which a process is assigned, the interrupt generation function 7 generates an interrupt signal for the process currently being executed.

When the interrupt signal is generated, the activation function 8 activates the parallel process that has been waiting for parallel synchronization, and resumes the allocation of the parallel process.

JP-A-9-128351

In the synchronous scheduling system described in Patent Document 1, a problem arises when a computer application communicates with another computer application and network protocol processing overhead occurs.

The reason why the overhead of the network protocol processing occurs is that the network protocol is used when communication is performed between computers. Usually, when communication is performed between computers, a network protocol such as TCP / IP (Transmission Control Protocol / Internet Protocol) is used. The computer realizes communication to an arbitrary computer using a network protocol.

Consider a case where an application that communicates between computers is under the management of synchronous scheduling by the method described in Patent Document 1. When an application is under management, when communication between applications is required, the computer assigns a network protocol to the transmission data and transmits it to the network. Further, the computer receives received data to which a network protocol is assigned from the network.

When assigning a network protocol, a predetermined time is required for calculations such as error correction codes and congestion control. Therefore, when an application of a computer under management of synchronous scheduling communicates with an application of another computer, the communicating application is required to wait for a predetermined time while the network protocol is processed.

In the synchronous scheduling method described in Patent Document 1, when there is an application waiting for a predetermined time, a CPU (Central Processing Unit) is assigned to another parallel application.

That is, when the technique described in Patent Document 1 is used, an application to which a CPU is assigned may be switched from a communicating application to another parallel application while the network protocol is processed. Switching of applications to which CPUs are assigned is a first problem when synchronous communication is performed in the synchronous scheduling system described in Patent Document 1.

As a network with less overhead than a TCP / IP network, there is a network called Infiniband (registered trademark) (Non-Patent Documents 1 and 2). However, in Infiniband also uses a network protocol for performing routing processing. That is, Infiniband is the same as a TCP / IP network in that there is an overhead due to protocol processing.

A second problem in synchronous communication in the synchronous scheduling system described in Patent Document 1 is that a plurality of networks that connect computers so that they can communicate with each other when a computer application performs synchronous communication with another computer application. In the device, data is processed asynchronously.

The reason why data is processed asynchronously is that network devices sequentially process network data input to the queue.

Network devices such as routers and switches operate as follows. For example, in the case of a router, the router arranges packets in a queue in the router in the order of input. Next, the router sequentially extracts packets from the head of the queue (the packet that arrived at the router earliest).

The router checks the network protocol of the extracted packet and checks the packet destination. Next, the router sends the packet to the network to which the destination computer is connected.

Therefore, even if an application starts synchronous communication, synchronous communication is not started unless all the packets stored in the network device before processing are processed.

Therefore, an object of the present invention is to provide a relay device, a communication device, a management device, a relay method, and a relay program that can contribute to synchronous communication without performing network protocol processing.

A relay device according to the present invention is a relay device that forwards packets, and includes at least a number that identifies a time slot that is a section divided by a predetermined time interval, and an output port number in the relay device. A connection information storage unit that stores connection information, a connection information processing unit that determines whether connection information corresponding to a time slot to which the current time belongs in the connection information storage unit, and a connection information processing unit that And a packet transfer unit that transfers a packet to an output port indicated by an output port number in the corresponding connection information when it is determined that corresponding connection information exists in the storage unit.

A communication apparatus according to the present invention is a communication apparatus that performs packet communication, and at least a number that identifies a time slot that is a section divided by a predetermined time interval, and an interface identifier corresponding to the number that identifies the time slot A connection management information storage unit that stores connection management information, and a communication processing unit that determines whether or not the connection management information storage unit includes connection management information corresponding to a time slot to which the current time belongs. When the communication processing unit determines that the corresponding connection management information exists, the packet is transmitted to the interface indicated by the interface identifier included in the corresponding connection management information.

A communication apparatus according to the present invention is a communication apparatus that performs packet communication, and at least a number that identifies a time slot that is a section divided by a predetermined time interval, and a memory map address corresponding to the number that identifies the time slot A connection management information storage unit that stores connection management information, and a communication processing unit that determines whether or not the connection management information storage unit includes connection management information corresponding to a time slot to which the current time belongs. When the communication processing unit determines that the corresponding connection management information exists, the received packet is copied to the buffer area indicated by the memory map address included in the corresponding connection management information.

A management device according to the present invention is a management device capable of communicating with a communication device and a relay device that transfers packets between the communication devices, and is a number that identifies a time slot that is a section divided at least by a predetermined time interval A time management information storage unit that stores the time management information including the time management information, and a transmission request of a packet including the connection source information and the connection destination information is received from the communication device, and the connection source information and the connection destination information are unallocated time management information A synchronization schedule setting unit that allocates the received connection source information and connection destination information, and a synchronization schedule transmission unit that transmits the time management information assigned to the synchronization schedule setting unit to the communication device and the relay device. It is characterized by.

A relay method according to the present invention is a relay method executed in a relay device, and includes at least a number for identifying a time slot which is a section divided by a predetermined time interval and an output port number in the relay device. If the connection information corresponding to the time slot to which the current time belongs is determined in the stored connection information and it is determined that the corresponding connection information exists, The packet is transferred to the output port indicated by the output port number in the corresponding connection information.

A relay program according to the present invention is a relay program executed in a relay device, and at the relay device, a number for identifying a time slot that is a section divided by at least a predetermined time interval, and an output port in the relay device. A storage process for storing connection information composed of a number, a connection information process for determining whether connection information corresponding to a time slot to which the current time belongs is present in the connection information stored in the storage process, and When it is determined by connection information processing that corresponding connection information exists, a transfer process for transferring a packet is executed on the output port indicated by the output port number in the corresponding connection information.

According to the present invention, it is possible to contribute to synchronous communication that does not perform network protocol processing.

It is a block diagram which shows the structural example of 1st Embodiment of the synchronous scheduling system 10 by this invention. 4 is an explanatory diagram showing an example of time slot information stored in a time management unit 120. FIG. It is explanatory drawing which shows the example of the computer connection management information memorize | stored in the computer connection management part 410. FIG. 3 is an explanatory diagram illustrating an example of a routing table stored in a network switch 300. FIG. 6 is an explanatory diagram illustrating an example of network connection management information stored in a network connection management unit 360. FIG. 5 is a flowchart showing an operation of setting a synchronization schedule by a synchronization schedule setting unit 130. 10 is a flowchart showing an operation of converting a synchronization schedule into computer connection management information by a synchronization schedule receiving unit 430. 4 is an explanatory diagram illustrating an example of a routing table stored in a computer 200. FIG. 6 is a flowchart showing a conversion operation of a synchronization schedule into network connection management information by a path conversion unit 320. It is a flowchart which shows operation | movement of the existing interruption process part. It is a flowchart which shows operation | movement of the interruption process part 420 in 1st Embodiment. It is a flowchart which shows operation | movement of the timer execution part 340. It is a block diagram which shows the structural example of 2nd Embodiment of the synchronous scheduling system 10 by this invention. 6 is an explanatory diagram illustrating an example of network route information stored in a network route information storage unit 160. FIG. It is explanatory drawing which shows the example of the network structure comprised from a server and a network switch. 6 is a flowchart showing an operation of generating network route information by the network route information generating unit 150. It is a flowchart which shows operation | movement of the synchronous management server 100 in 2nd Embodiment. It is a block diagram which shows the structural example of 3rd Embodiment of the synchronous scheduling system 10 by this invention. It is explanatory drawing which shows the example of the transfer flag information memorize | stored in the transfer flag memory | storage part 440. It is explanatory drawing which shows the data structure of the packet transferred by the synchronous communication in 3rd Embodiment. It is a flowchart which shows operation | movement of the interruption process part 420 in 3rd Embodiment. It is a flowchart which shows operation | movement of the interruption process part 420 in 3rd Embodiment. 5 is a flowchart showing an operation of synchronization error detection by a synchronization error detection unit 450. It is a block diagram which shows the outline | summary of the relay apparatus by this invention. It is a block diagram which shows the outline | summary of the communication apparatus by this invention. It is a block diagram which shows the other outline | summary of the communication apparatus by this invention. It is a block diagram which shows the outline | summary of the management apparatus by this invention. It is a block diagram which shows the structure of the synchronous scheduling system described in patent document 1. FIG.

Embodiment 1. FIG.
[Description of configuration]
DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, a first embodiment of the invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of a first embodiment of a synchronous scheduling system 10 according to the present invention. A synchronization scheduling system 10 illustrated in FIG. 1 includes a synchronization management server 100, a computer 200, and a network switch 300.

The synchronization management server 100, the computer 200, and the network switch 300 are communicably connected via a network 500 indicated by a straight line in FIG. 1 or a network 600 indicated by a dotted line in FIG.

1, the synchronization management server 100 includes a time synchronization unit 110, a time management unit 120, a synchronization schedule setting unit 130, and a synchronization schedule transmission unit 140. Each component operates generally as follows.

The time synchronization unit 110 has a function of synchronizing the time of another device in the synchronization scheduling system 10 with the time of the synchronization management server 100. The time synchronization unit 110 is connected to a network 500 that is a network different from the network 600 used for transmission / reception of a synchronization schedule.

The time management unit 120 has a function of managing servers that perform synchronous communication via the network 600 at a certain time. The time management unit 120 stores time slot information. In this embodiment, the time slot means the right to perform synchronous communication divided at a predetermined time interval.

FIG. 2 is an explanatory diagram showing an example of time slot information stored in the time management unit 120. As shown in FIG. 2, the time slot information includes a time slot number 111, a connection source IP address 112, and a connection destination IP address 113.

The time slot number 111 is identification information for uniquely identifying time slot information. The time slot number 111 corresponds to each divided time when a certain unit time is divided. If the unit time is 1 second and the number of time slots is 4, the time allocated to one time slot is 0.25 seconds.

When 0: 0: 0 is the start time of the unit time, the first time slot corresponds to the divided time from 0: 0: 0 to 0: 0: 0. The communication device to which the first time slot is assigned communicates between 0: 0: 0 to 0: 0: 0.

The next time slot corresponds to the divided time from 0: 0: 0-25 to 0: 0: 0. The communication device to which the next time slot is assigned communicates between 0: 0: 0-25 and 0: 0: 0: 0. The assignment of communication to time slots is repeated by the number of time slots.

The connection source IP address 112 indicates an IP address used by the application 220 that requested time slot assignment. The connection destination IP address 113 indicates an IP address used by an application communicating with the application 220 that has requested the time slot assignment.

The synchronization schedule setting unit 130 receives a synchronization schedule setting request from the application 220. In addition, when there is an empty time slot stored in the time management unit 120, the synchronization schedule setting unit 130 assigns the received synchronization schedule setting request to the time slot.

In the time slot information shown in FIG. 2, the synchronization schedule setting request from the application 220 is assigned to a single time slot, but the synchronization schedule setting request may be assigned to a plurality of time slots.

In the state where “the time slot is available”, the time slot information stored in the time management unit 120 includes time slot information to which the connection source IP address 112 and the connection destination IP address 113 are not yet assigned. It means that it exists (hereinafter also referred to as “there is an empty slot”). The synchronization schedule setting request sent from the application 220 includes the connection source IP address 112 and the connection destination IP address 113.

The synchronization schedule transmission unit 140 has a function of broadcasting the time slot information set by the synchronization schedule setting unit 130 to the computer 200 and the network switch 300 via the network 600. The transmitted time slot information corresponds to one entry of time slot information stored in the time management unit 120.

As shown in FIG. 1, the computer 200 includes a time synchronization unit 210, an application 220, a computer connection management unit 410, an interrupt processing unit 420, and a synchronization schedule reception unit 430. Each component operates generally as follows.

The time synchronization unit 210 has a function of synchronizing the time of the other device of the synchronous scheduling system 10 with the time of the computer 200. The time synchronization unit 210 is connected to a network 500 that is a network different from the network 600 used for transmission / reception of a synchronization schedule.

Application 220 has a function of performing synchronous communication with an application running on another computer. The application 220 runs on the computer 200. The application 220 performs various processes such as packet creation, transmission, and reception according to communication contents.

The computer connection management unit 410 has a function of storing information necessary for passing an arrived packet to the application 220 or sending a packet from the application 220 at a time corresponding to a certain time slot. FIG. 3 is an explanatory diagram showing an example of computer connection management information stored in the computer connection management unit 410.

The computer connection management information includes a time slot number 411, an interface identifier, an interrupt number 413, and a memory map address 414. An interface 412 shown in FIG. 3 corresponds to an identifier of the interface.

The time slot number 411 indicates the same number as the time slot number 111. The interface identifier indicates, for example, the identifier of the network card at which the packet arrives or is transmitted at the time corresponding to the time slot number 411.

Interrupt number 413 indicates a number notified to the CPU when a packet arrives at the time corresponding to the time slot number 411. The memory map address 414 indicates a memory buffer area used when a packet arriving at the corresponding time slot number 411 is passed to the user program.

Referring to FIG. 3, a packet arriving at time 1 of time slot number 411 is stored in a memory buffer area corresponding to “0x0001-0x000F” indicated by memory map address 414 without performing protocol processing.

The interrupt processing unit 420 has a function of storing the arrived packet in the memory buffer area indicated by the memory map address 414 described in the computer connection management information.

In normal communication processing, the interrupt processing unit 420 stores the arrived packet in a certain memory buffer area and then performs network protocol processing. Next, the interrupt processing unit 420 copies the processed packet to the memory buffer area of the destination application 220.

In this embodiment, network protocol processing may not be performed. In this case, the interrupt processing unit 420 in the present embodiment simply performs processing for storing the transmitted packet in the memory buffer area indicated by the memory map address 414 described in the computer connection management information, and performs network protocol processing. Absent.

The synchronization schedule reception unit 430 has a function of converting the synchronization schedule sent from the synchronization schedule transmission unit 140 of the synchronization management server 100 via the network 600 into computer connection management information and storing it in the computer connection management unit 410. . The sent synchronization schedule is time slot information.

Upon receiving the time slot information, the computer 200 converts the time slot information in accordance with the operation of the computer 200. When the computer 200 receives the time slot information, the synchronization schedule receiving unit 430 converts the received time slot information into computer connection management information.

As illustrated in FIG. 1, the network switch 300 includes a time synchronization unit 310, a path conversion unit 320, a synchronization schedule reception unit 330, a timer execution unit 340, a physical socket connection unit 350, and a network connection management unit 360. Including. Each component operates generally as follows.

The time synchronization unit 310 has a function of synchronizing the time of the other devices in the synchronization scheduling system 10 with the time of the network switch 300. The time synchronization unit 310 is connected to a network 500 that is a network different from the network 600 used for transmission / reception of a synchronization schedule. That is, the network 500 is used for time synchronization processing by the time synchronization unit 110, the time synchronization unit 210, and the time synchronization unit 310.

The synchronization schedule reception unit 330 has a function of converting the synchronization schedule sent from the synchronization schedule transmission unit 140 of the synchronization management server 100 into network connection management information and storing it in the network connection management unit 360. When converting into network connection management information, the synchronization schedule receiving unit 330 uses the path conversion unit 320.

The path conversion unit 320 has a function of converting the synchronization schedule sent from the synchronization schedule transmission unit 140 of the synchronization management server 100 into network connection management information stored in the network connection management unit 360. The sent synchronization schedule is time slot information.

The route conversion unit 320 checks the routing table stored in the network switch 300. At the time of confirmation, the path conversion unit 320 confirms to which network the connection source IP address 112 and the connection destination IP address 113 included in the time slot information belong. In addition, the path conversion unit 320 acquires the port number of the port of the network switch 300 used for input / output to the confirmed network.

FIG. 4 is an explanatory diagram showing an example of a routing table stored in the network switch 300. The routing table shown in FIG. 4 includes a destination network, a subnet mask, a next hop, and a port number.

Upon receiving the time slot information (synchronization schedule), the network switch 300 converts the time slot information in accordance with the operation of the network switch 300. The network switch 300 converts the received time slot information into network connection management information. Hereinafter, an example of conversion of the synchronization schedule into network connection management information by the path conversion unit 320 will be described.

In the example, the time slot number 111 of the received synchronization schedule is 1, the connection source IP address 112 is “192.168.1.2”, and the connection destination IP address 113 is “192.168.2.3”. To do.

As shown in FIG. 4, all the subnet masks in the routing table are set to “255.255.255.0”. The route conversion unit 320 confirms from the subnet mask that the IP address indicated by the connection source IP address 112 belongs to “192.168.1.0” of the destination network. Similarly, the path conversion unit 320 confirms that the IP address indicated by the connection destination IP address 113 belongs to “192.168.2.0” of the destination network.

Next, the path conversion unit 320 indicates that the port number of the port of the network switch 300 used for input / output to the network whose destination network is “192.168.1.0” is “S0” from the routing table. Confirm. Similarly, the path conversion unit 320 confirms that the port number of the port of the network switch 300 used for input / output to the network whose destination network is “192.168.2.0” is “S1”. .

Therefore, in the converted network connection management information, the time slot number 361 is 1, the first port number 362 is “S0”, and the second port number 363 is “S1”. This network connection management information corresponds to the first network connection management information shown in FIG.

When the communication in the synchronization schedule corresponding to the received time slot information is assumed to be one-way communication from the computer to another computer, the path conversion unit 320 displays the time slot information as the time slot information. You may convert into the network connection management information containing the number 361 and one port number. The reason is that, since the packet arrives at the network switch 300 from only one computer at the time corresponding to the time slot number 361, the network switch 300 performs only the process of transmitting the received packet from the port corresponding to the port number. This is because it may be carried out.

The network connection management unit 360 has a function of managing a combination of physical ports of the network switch 300 connected for each time slot. FIG. 5 is an explanatory diagram illustrating an example of network connection management information stored in the network connection management unit 360. The network connection management information includes a time slot number 361, a first port number 362, and a second port number 363.

The timer execution unit 340 has a function of discarding the staying packets according to the set timer. Also, the timer execution unit 340 uses the physical socket connection unit 350 according to the set timer, and the port indicated by the first port number 362 and the port indicated by the second port number 363 described in the network connection management information And a function of connecting the two.

The physical socket connecting unit 350 has a function of connecting the port indicated by the first port number 362 and the port indicated by the second port number 363 described in the network connection management information.

The inter-physical socket connection unit 350 also has a normal function of acquiring a destination IP address from a network protocol included in a received packet and determining a physical port number as a transmission destination according to a routing table.

For example, if the destination IP address accompanying the arrived packet is “192.168.1.1”, the inter-socket connection unit 350 sets the physical port used for transmission to “S0” from the routing table shown in FIG. Decided.

In the present embodiment, the inter-physical socket connection unit 350 has a time corresponding to the time slot number 361, a port indicated by the first port number 362 described in the network connection management information, or the second port number. Transfer to the port indicated by 363.

Note that the network switch 300 in this embodiment may include a centralized buffer. When the network switch 300 includes a centralized buffer, a packet that arrives at the network switch 300 is output from an input port (for example, the port indicated by the first port number 362) to the centralized buffer and temporarily stored in the centralized buffer. After being saved, the packet is sent out from the output port (for example, the port indicated by the second port number 363).

Further, the buffer provided in the network switch 300 is not limited to the centralized buffer, and may be a buffer provided for each port.

More specifically, the physical socket connection unit 350 operates as follows. That is, the timer execution unit 340 as the connection information processing unit confirms the current time and confirms network connection management information corresponding to the current time.

When the timer execution unit 340 determines that the network connection management information corresponding to the current time exists, the inter-physical socket connection unit 350 determines the input port number (for example, the first port number 362) in the corresponding network connection management information. And whether the packet to be transferred matches the port number (for example, the first port number 362) input to the network switch 300.

If they match, the physical socket connection unit 350 transfers the packet to the output port indicated by the output port number (for example, the second port number 363) in the corresponding network connection management information.

In this case, the buffer provided in the network switch 300 described above for temporarily storing the received packet is preferably provided in the physical socket connecting unit 350, but the buffer is installed at the physical socket connecting unit 350. It is not limited to. A storage unit (buffer) may be provided at a location different from the physical socket connecting unit 350.

By operating as described above, the packet arriving at the port indicated by the first port number 362 for the time corresponding to the time slot number 361 is transmitted to the port indicated by the second port number 363. A packet that arrives at the port indicated by the second port number 363 is transmitted to the port indicated by the first port number 362. The physical port from which the packet has arrived is known information of the network switch 300.

In the operation of the physical socket connection unit 350, it is confirmed whether the input port number in the network connection management information matches the port number input to the network switch 300 for the packet to be transferred. The operation of transferring the packet to the output port indicated by the output port number in the corresponding network connection management information may be performed by another component (for example, a packet transfer unit (not shown)).

Each component included in the synchronization management server 100, the computer 200, and the network switch 300 operates so that an application using the time slot can exclusively use the network path at a predetermined time. Therefore, data is transmitted and received between applications in a plurality of devices connected via a network without adding a protocol necessary for routing to the packet.

In the synchronization scheduling system 10 according to the present embodiment, the synchronization management server 100, the computer 200, and the network switch 300 are realized by a CPU that executes processing according to a program, for example. The synchronization management server 100, the computer 200, and the network switch 300 may be realized by hardware.

Also, the synchronization schedule setting unit 130, the synchronization schedule transmission unit 140, the application 220, the interrupt processing unit 420, the synchronization schedule reception unit 430, the path conversion unit 320, the synchronization schedule reception unit 330, the timer execution unit 340, and the physical socket connection unit 350. Is realized, for example, by a CPU that executes processing according to program control.

In addition, the time management unit 120, the computer connection management unit 410, and the network connection management unit 360 are realized by, for example, a RAM (Random Access Memory).

The time synchronization unit 110, the time synchronization unit 210, and the time synchronization unit 310 are realized by a CPU that operates in accordance with software for executing processing based on, for example, NTP (Network Time Protocol) or PTP (Precision Time Protocol). .

[Description of operation]
The operation of the synchronous scheduling system 10 of this embodiment will be described below with reference to FIGS. 6 to 7 and FIGS. 9 to 12.

The operation of the synchronization schedule setting by the synchronization schedule setting unit 130 of the synchronization management server 100 will be described. FIG. 6 is a flowchart showing an operation of setting a synchronization schedule by the synchronization schedule setting unit 130.

The application 220 notifies the synchronization schedule setting unit 130 of its own IP address (connection source IP address 112) and the communication partner's IP address (connection destination IP address 113) as a synchronization schedule request (step S001). .

Next, the synchronization schedule setting unit 130 searches the time management unit 120 to check whether there is an empty slot (step S002). If there is an empty slot (YES in step S003), the synchronization schedule setting unit 130 registers the synchronization schedule request in a time slot that is an empty slot stored in the time management unit 120 (step S004).

The registered synchronization schedule list is a list of time slot information stored in the time management unit 120 in which the time slot number 111, the connection source IP address 112, and the connection destination IP address 113 are combined into one set.

Next, the synchronization schedule transmission unit 140 broadcasts the registered time slot information to the computer 200 and the network switch 300 (step S005). After broadcasting, the synchronization management server 100 ends the process.

If there is no empty slot (NO in step S003), the synchronization schedule transmission unit 140 notifies the application 220 of the failure of the synchronization schedule (step S006). After the notification, the synchronization management server 100 ends the process.

Next, an operation in which the synchronization schedule receiving unit 430 converts the synchronization schedule sent to the computer 200 into computer connection management information will be described. FIG. 7 is a flowchart showing the conversion operation of the synchronization schedule into the computer connection management information by the synchronization schedule receiving unit 430.

The synchronization schedule reception unit 430 searches the routing table for the identifier of the interface used for communication with the IP address indicated by the connection destination IP address 113 included in the time slot information sent from the synchronization schedule transmission unit 140. (Step S011).

FIG. 8 is an explanatory diagram showing an example of a routing table stored in the computer 200. The routing table shown in FIG. 8 includes a destination network, a subnet mask, a gateway address, and an interface.

The synchronization schedule receiving unit 430 confirms the destination network including the IP address indicated by the connection destination IP address 113 from the routing table. Next, the synchronization schedule receiving unit 430 selects an interface used for transmitting the packet to the destination network.

Consider a case where the IP address indicated by the connection destination IP address 113 is “192.168.1.1”. The synchronization schedule receiving unit 430 knows that the destination network including the IP address “192.168.1.1” is “192.168.1.0” from the subnet mask. Further, the synchronization schedule receiving unit 430 recognizes that the interface used for transmission to the confirmed destination network is “Eth0” from the routing table.

Next, the synchronization schedule receiving unit 430 searches for the interrupt number 413 used in the selected interface 412 (step S012). The correspondence information between the interface identifier and the interrupt number 413 is, for example, held in advance in an OS (Operating System) 400.

For example, when the OS 400 is Linux (registered trademark), the OS 400 can know the correspondence between the interface identifier and the interrupt number 413 by referring to the / proc / interrupt file.

Next, the OS 400 secures a buffer area for a user program that handles packets (step S013). The memory map address 414 of the buffer area is notified to the OS 400 when the user program issues a packet reception system call to the OS 400.

For example, consider a case where a user program issues a Read system call and receives a packet. When the Read system call is used, the user program prepares a buffer area and notifies the OS 400 of the prepared buffer area pointer as the second argument of the Read system call.

In the above case, the OS 400 copies the contents of the buffer area of the OS 400 in which the received packet data is stored in the buffer area specified by the user program and returns it to the user program. Through the processing of the OS 400, the user program can receive the transmitted packet data.

When an existing Read system call is used, a memory copy occurs when data is transferred from the OS 400 to the user program. In order to avoid the occurrence of memory copy, the OS 400 may prepare a new system call and directly reference the buffer area in which received packet data is stored to the user program.

Next, the synchronization schedule receiving unit 430 creates computer connection management information from the interface identifier, interrupt number 413, and memory map address 414 obtained in steps S011 to S013 (step S014). The synchronization schedule receiving unit 430 stores the created computer connection management information in the computer connection management unit 410. After storing, the synchronization schedule receiving unit 430 ends the process.

Next, an operation in which the path conversion unit 320 converts the synchronization schedule sent to the network switch 300 into network connection management information will be described. FIG. 9 is a flowchart showing the conversion operation of the synchronization schedule into the network connection management information by the path conversion unit 320.

The synchronization schedule reception unit 330 searches the routing table for the port number of the port used for communication with the IP address indicated by the connection destination IP address 113 included in the time slot information sent from the synchronization schedule transmission unit 140. (Step S021).

When performing a search, the synchronization schedule receiving unit 330 uses the route conversion unit 320. The method for converting the synchronization schedule into the network connection management information using the path conversion unit 320 is as described above.

Next, the synchronization schedule receiving unit 330 sends the port number of the port used for communication with the IP address indicated by the connection source IP address 112 included in the time slot information sent from the synchronization schedule transmitting unit 140 to the routing table. (Step S022). The search method is the same as the search method in step S021.

Next, the synchronization schedule receiving unit 330 creates network connection management information from the two port numbers obtained in step S021 and step S022 (step S023). The synchronization schedule receiving unit 330 stores the created network connection management information in the network connection management unit 360. After storing, the synchronization schedule receiving unit 330 ends the process.

Next, the operation of the interrupt processing unit 420 of the computer 200 will be described. FIG. 11 is a flowchart illustrating the operation of the interrupt processing unit 420 according to the first embodiment. Before describing the operation of the interrupt processing unit 420 in this embodiment, the difference between the processing of this embodiment and the existing processing will be described with reference to the flowchart showing the operation of the existing interrupt processing unit shown in FIG. .

When the packet arrives at the computer 200, an interrupt signal is sent to the CPU (step S101). The interrupt signal is an electrical signal sent from the network card to the CPU.

The network card copies packet data that has arrived in its own buffer to the memory buffer area designated by the OS 400. The memory buffer area where the packet data is copied is called a DMA (Direct Memory Access) buffer. Next, the network card transmits an interrupt signal to the CPU.

The OS 400 that has received the interrupt signal checks the interrupt number information and selects a program to be executed (step S102). Next, the OS 400 executes an interrupt program (step S103).

Next, the OS 400 takes out the packet data from the DMA buffer (step S104), and executes a network protocol process (step S105). Next, the OS 400 copies the packet to the user program buffer (step S106). After copying, the computer 200 ends the process.

The user program buffer corresponds to the memory buffer area specified by the second argument of the read Read system call. As the computer 200 operates as described above, the user program receives the received packet data.

Next, the operation of the interrupt processing unit 420 in the present embodiment will be described with reference to FIG. From step S111 to step S113 shown in FIG. 11, the interrupt processing unit 420 operates in the same manner as the existing interrupt processing unit. The operation from the step following step S113 becomes the operation of the interrupt processing unit 420 in the present embodiment.

The interrupt processing unit 420 acquires the current time (step S114). Next, the interrupt processing unit 420 confirms whether there is computer connection management information corresponding to the acquired time in the computer connection management unit 410 (step S115).

If the computer connection management information exists (YES in step S116), the interrupt processing unit 420 sets the copy destination memory address to the memory map address 414 described in the computer connection management information (step S120). Next, the interrupt processing unit 420 copies data from the DMA buffer to the buffer area corresponding to the memory map address 414 (step S121). After copying, the interrupt processing unit 420 ends the process.

When the computer connection management information does not exist (NO in step S116), the OS 400 takes out the packet from the DMA buffer in the same manner as the normal interrupt process (step S117). Next, after executing network protocol processing (step S118), the OS 400 copies the packet to the memory buffer area specified by the user program (step S119). After copying, the OS 400 ends the process.

Next, the operation of the timer execution unit 340 of the network switch 300 will be described. FIG. 12 is a flowchart showing the operation of the timer execution unit 340.

Suppose the timer is set in advance. For example, if the unit time is 1 second, the time slot is 4 times, and 0: 0: 0 is the start time of the unit time, the timer executes time confirmation every 0.25 seconds from 0: 0 minutes 0 seconds. Is set as follows.

The timer execution unit 340 confirms the current time. Next, the timer execution unit 340 determines whether the network connection management information corresponding to the confirmed current time exists in the network connection management unit 360 (step S131).

The timer execution unit 340 extracts the first port number 362 and the second port number 363 from the confirmed network connection management information. The timer execution unit 340 makes the extracted information unique routing information in the time slot corresponding to the current time.

The routing information set in step S132 becomes the routing information of the connection part 350 between physical sockets. The physical socket connecting unit 350 transfers the packet to the output port indicated by the output port number (for example, the second port number 363) in the network connection management information corresponding to the current time (step S132).

More specifically, the operation from step S131 to step S132 is described as follows. That is, the timer execution unit 340 confirms the current time and confirms network connection management information corresponding to the current time.

The physical socket connection unit 350 connects the port indicated by the extracted first port number 362 and the port indicated by the second port number 363.

Next, the timer execution unit 340 discards packets accumulated in the network queue that do not correspond to the routing information set in step S132 (step S133).

More specifically, in step S133, the timer execution unit 340 operates as follows. As described above, the inter-physical socket connection unit 350 includes the input port number (for example, the first port number 362) in the network connection management information corresponding to the current time, and the port for which the packet to be transferred is input to the network switch 300. It is confirmed whether the number (for example, the first port number 362) matches. If they match, the physical socket connection unit 350 transfers the packet to the output port indicated by the output port number (for example, the second port number 363) in the corresponding network connection management information.

Next, the timer execution unit 340 receives a packet that is not matched by the confirmation of the physical socket connection unit 350, that is, other than the input port number (for example, the first port number 362) in the network connection management information corresponding to the current time. The packet received from the port indicated by the port number is deleted from the stored buffer (network queue). The packet discarding period by the timer execution unit 340 is arbitrary. After discarding the packet, the timer execution unit 340 ends the process.

As described above, the operation of the timer execution unit 340 has been described as a preferred specific example. However, in the present embodiment, the operation of step S133 is not an essential operation. If time synchronization and time slot information are shared among all devices of the synchronization management server 100, the computer 200, and the network switch 300 constituting the synchronization scheduling system 10 of the present embodiment, it corresponds to a predetermined time slot. This is because during the time, the network switch 300 does not receive a packet from an input port that does not correspond to the only routing information in the time slot.

Through the above processing, the network switch 300 starts transferring a packet corresponding to the routing information set in step S132.

[Description of effects]
Next, the effect of this embodiment will be described. In the present embodiment, since applications that communicate at a certain time are limited, data is directly transmitted and received between applications without using protocol processing. Therefore, when two or more applications communicate with each other between computers, the applications can communicate without referring to the header of the packet, so that overhead due to network protocol processing does not occur.

Furthermore, in this embodiment, the packets processed by the network switch within the time corresponding to the predetermined time slot are limited to only the packets processed by the application to which the predetermined time slot is assigned. Data is processed synchronously by a plurality of network devices that can be connected.

Embodiment 2. FIG.
[Description of configuration]
Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 13 is a block diagram showing a configuration example of the second embodiment of the synchronous scheduling system 10 according to the present invention.

Referring to FIG. 13, the present embodiment is different from the first embodiment in that the synchronization management server 100 includes a network route information generation unit 150 and a network route information storage unit 160. In addition, since the network route information generation unit 150 and the network route information storage unit 160 are included, the operation of the synchronization schedule setting unit 130 is different from that of the first embodiment.

[Description of operation]
The network route information generation unit 150 and the network route information storage unit 160 operate as outlined below. The network route information generation unit 150 generates route information 163 of a packet route used when a certain server communicates with another server. The route information 163 is information indicating a combination of network switches through which a packet passes.

As described in the first embodiment, the application 220 sends a synchronization schedule request to the synchronization schedule setting unit 130 with its own IP address (connection source IP address 112) and the other party's IP address (connection destination IP address 113). ).

Based on the notified information, the routing table stored in the computer 200, and the routing table stored in the network switch 300, the network route information generation unit 150 determines a combination of the network switches 300 through which the packet passes. decide.

The network route information storage unit 160 stores the network route information determined by the network route information generation unit 150. FIG. 14 is an explanatory diagram illustrating an example of the network route information stored in the network route information storage unit 160.

As shown in FIG. 14, the network route information describes a combination of a connection source server 161 in which the connection source IP address 112 is described, a connection destination server 162 in which the connection destination IP address 113 is described, and the network switch 300. Route information 163. The route information 163 is a combination of the network switches 300 used when packets are transmitted and received between the connection source server 161 and the connection destination server 162.

Note that information assigned to the connection source server 161, the connection destination server 162, and the route information 163 in the network route information is an IP address. In the case of the example shown in FIG. 14, the names of the respective devices are assigned for the sake of simplicity of explanation.

As an example, a method for determining the route information 163 when the server 101 is connected to the server 104 by communication is shown. FIG. 15 is an explanatory diagram illustrating an example of a network configuration including a server and a network switch. Hereinafter, a method for determining the route information 163 will be described with reference to FIG.

Referring to FIG. 15, the server 101 is connected to the router A (hereinafter, the network switch 300 is referred to as router A). Similarly, the server 102 is connected to the router C, the server 103 is connected to the router B, and the server 104 is connected to the router D. The routers are connected in a grid pattern.

Here, route information 163 of a route used when the server 101 communicates with the server 104 is created as follows. The network path information generation unit 150 confirms the routing table of the server 101 and identifies the router to which the server 101 is next connected for communication. As shown in FIG. 15, the next router for communication connection is router A.

Next, the network route information generation unit 150 confirms the routing table of the router A and identifies the router to which the router A is next connected for communication. As shown in FIG. 15, the next router for communication connection is router B.

The route information 163 is obtained by repeating the operation of specifying the connection destination until the next connection destination becomes the server 104. As illustrated in FIG. 14, route information 163 of a route used when the server 101 communicates with the server 104 is a combination of router A, router B, and router D.

The network route information generation unit 150 is realized by a CPU that executes processing according to program control, for example. The network path information storage unit 160 is realized by a RAM, for example.

Hereinafter, the operation of the synchronous scheduling system 10 of this embodiment will be described with reference to FIGS.

The operation of network route information generation by the network route information generation unit 150 of the synchronization management server 100 will be described. FIG. 16 is a flowchart showing an operation of generating network route information by the network route information generating unit 150.

The network route information generation unit 150 confirms the routing table of the connection source server 161 and acquires the IP address of the router to which the connection source server 161 is connected (step S201).

If the IP address of the router exists (YES in step S202), the network route information generation unit 150 checks the routing table of the router that holds the acquired IP address of the router. Next, the network route information generation unit 150 acquires the IP address of the router to which the router confirmed in step S202 is connected (step S203).

When the IP address of the router exists (YES in step S204), the network route information generation unit 150 repeats the process of step S203. When the IP address of the router does not exist (NO in step S202 or NO in step S204), the network route information generation unit 150 sets the acquired combination of router IP addresses as route information 163.

Next, the network route information generation unit 150 creates network route information (step S205). After creating the network route information, the network route information generation unit 150 ends the process.

For example, referring to the routing table stored in the computer 200 shown in FIG. 8, the IP address of the router to which the connection source server 161 is connected is “192.168.1.1” described in the gateway address. It corresponds to.

That is, when the IP address belonging to the network corresponding to “192.168.1.0” or the IP address belonging to the network corresponding to “192.168.2.0” becomes the connection destination IP address, There is no routed router. That is, nothing is stored as the route information 163.

When an IP address belonging to a network other than the network corresponding to “192.168.1.0” or “192.168.2.0” is the IP address of the connection destination, “192.168.1.1” The router having the IP address is recorded as the first connection destination router.

Referring to the routing table stored in the router A shown in FIG. 4, the IP address corresponding to the next hop becomes the IP address of the router to which the router A communicates next.

For example, when the IP address belonging to the network corresponding to “192.168.2.0” is the IP address of the connection destination, the IP address of the next router for communication connection is “192.168.2.1”. become.

Note that the route information included in the routing tables shown in FIGS. 4 and 8 does not correspond to the network configuration shown in FIG. Through the above processing, the route information 163 is specified.

FIG. 17 is a flowchart showing the operation of the synchronization management server 100 in the second embodiment. The network path information generation unit 150 first inputs the information of the connection destination IP address 113 and the information of the connection source IP address 112, and generates network path information (step S211). The generation method is as shown in the flowchart of FIG.

Next, the synchronization schedule transmission unit 140 sends time slot information to each router of the network path, the server holding the IP address indicated by the connection destination IP address 113, and the server holding the IP address indicated by the connection source IP address 112. Multicast (step S212).

Next, the synchronization schedule transmission unit 140 checks whether replies have arrived from all multicast destinations within the set timeout period (step S213). The timeout time is set in advance.

There are the following methods for setting the timeout time, for example. Assume that the unit time is 1 second, there are four time slots, and the start time of the unit time is 0: 0: 0. When the current time is 0:00, 0, 50 and the fourth time slot is set, the timeout time for confirming the reply to the multicast can be set to 0.25 seconds.

In the above example, the start time of the fourth time slot is 0: 0: 0. Therefore, if no synchronization schedule is set by the start time, synchronous communication is not started. Therefore, in order to start synchronous communication at the start time, a timeout setting method in which the timeout time is assigned to a time slot is given as an example.

If all are received within the timeout period (YES in step S214), the synchronization management server 100 ends the process as it is. If not all can be received within the timeout period (NO in step S214), the synchronization schedule transmission unit 140 notifies the application that the synchronization schedule has failed (step S215). After the notification, the synchronization management server 100 ends the process.

[Description of effects]
Next, the effect of this embodiment will be described. In this embodiment, a route used for communication within a time corresponding to a predetermined time slot is specified in advance, a synchronization schedule is transmitted to the router, the connection destination computer, and the connection source computer in the specified route, and a time-out occurs If there is no reply to the synchronization schedule transmission within the time, it is configured that the synchronization scheduling is regarded as a failure, so that errors in setting the synchronization schedule can be reduced. In the first embodiment, since the synchronization schedule is broadcast, it cannot be determined whether or not the synchronization schedule is correctly set in the computer or the router. However, when using the synchronization scheduling system 10 of this embodiment, the user can You can determine whether the schedule is set correctly.

Embodiment 3. FIG.
[Description of configuration]
Next, a third embodiment of the present invention will be described with reference to the drawings. FIG. 18 is a block diagram showing a configuration example of the third embodiment of the synchronous scheduling system 10 according to the present invention.

Referring to FIG. 18, the present embodiment is different from the first embodiment in that the synchronization management server 100 includes a synchronization error notification unit 170, the computer 200 includes a transfer flag storage unit 440, and a synchronization error detection unit. 450 is different. Since the transfer flag storage unit 440 and the synchronization error detection unit 450 are included, the operation of the interrupt processing unit 420 is different from that of the first embodiment.

[Description of operation]
The transfer flag storage unit 440, the synchronization error detection unit 450, and the synchronization error notification unit 170 generally operate as follows. The synchronization error detection unit 450 has a function of detecting a communication error during synchronous communication. The synchronization error detection unit 450 uses a transfer flag 442 described later when detecting a communication error.

When a synchronization communication error is detected using the transfer flag 442, the synchronization error detection unit 450 notifies the synchronization error notification unit 170 that a synchronization communication error has occurred.

The transfer flag storage unit 440 is a storage area for storing a flag used at the time of transfer. FIG. 19 is an explanatory diagram illustrating an example of transfer flag information stored in the transfer flag storage unit 440.

As shown in FIG. 19, the transfer flag information includes a time slot number 441 and a transfer flag 442. The transfer flag 442 includes a start flag 443 and an end flag 444.

When the synchronous communication is started at the time corresponding to the predetermined time slot, the start flag 443 of the transfer flag 442 is stored in the transfer flag storage unit 440. When the synchronous communication ends within the time corresponding to the time slot, the end flag 444 of the transfer flag 442 is stored in the transfer flag storage unit 440.

By storing a set of the start flag 443 and the end flag 444 in the transfer flag storage unit 440, it is considered that the synchronous communication of the corresponding time slot has been successful. If the end flag 444 is not stored, the synchronous communication is considered to have failed.

FIG. 20 is an explanatory diagram showing a data structure of a packet transferred by synchronous communication in the third embodiment. As shown in FIG. 20, in synchronous communication in the synchronous scheduling system 10 of this embodiment, it is required to add a time slot number 441 and a transfer flag 442 to header information in addition to user data (payload). The transfer flag 442 added to the header information includes a start flag 443, an in-transfer flag, and an end flag 444.

In the example shown in FIG. 20, the start flag 443 is set to “10”, the transferring flag is “00”, and the end flag 444 is set to “11”.

In addition, when the length of the entire packet, the length of the time slot number storage location, and the length of the transfer flag storage location are all set to a fixed length, the position of the time slot number 441 and the transfer flag 442 in a packet Is uniquely identified. In the example shown in FIG. 20, the length of the entire packet is fixed to 1500 bytes, the length of the storage location of the time slot number is 16 bits, and the length of the storage location of the transfer flag is fixed to 2 bits.

Note that, as long as the time slot number 441 and the transfer flag 442 in the packet can be extracted, the fixed-length packet may not be used as in the example shown in FIG.

The synchronization error notification unit 170 has a function of notifying the user that the synchronous communication has failed. The failure of the synchronous communication is sent from the synchronization error detection unit 450.

The synchronization error notification unit 170 and the synchronization error detection unit 450 are realized by a CPU that executes processing according to program control, for example. The transfer flag storage unit 440 is realized by a RAM, for example.

Hereinafter, the operation of the synchronous scheduling system 10 of the present embodiment will be described with reference to FIGS.

The operation of recording the synchronous communication start flag 443 by the changed interrupt processing unit 420 will be described. FIG. 21 is a flowchart illustrating the operation of the interrupt processing unit 420 according to the third embodiment. FIG. 21 shows the operation of the interrupt processing unit 420 when recording a synchronous communication start flag.

The difference from the first embodiment is the operation of the interrupt processing unit 420 when there is a corresponding time slot in the time management unit 120 in step S306. If there is a time slot (YES in step S306), either synchronous communication is started or communication is in progress.

When the synchronous communication is started and the connection destination computer 200 receives the packet for the first time, a start flag 443 is set in the packet. The interrupt processing unit 420 reads the set start flag 443 (step S310).

If the start flag 443 is present in the packet (YES in step S311), synchronous communication is started. When synchronous communication is started, the synchronous error detection unit 450 executes synchronous communication error detection processing (step S312).

That is, when synchronous communication of the time slot corresponding to the current time is started, the transfer flag information stored in the transfer flag storage unit 440 immediately before the transfer flag information corresponding to the current time includes a start flag. 443 and the end flag 444 should be set.

If the end flag 444 is not set with reference to the previous transfer flag information, this corresponds to a state in which the previous synchronous communication has not yet ended and the next synchronous communication has started. That is, the previous synchronous communication has failed. When the previous synchronization communication has failed, the synchronization error detection unit 450 notifies the synchronization error notification unit 170 of the failure of the synchronization communication.

When the end flag 444 is set with reference to the previous transfer flag information, this corresponds to the state where the synchronous communication is completed within the time corresponding to the time slot in which there is synchronization. When the synchronous communication is completed, the synchronization error detection unit 450 notifies the synchronization error notification unit 170 of nothing.

The synchronization error detection unit 450 records the time slot number 441 and the start flag 443 corresponding to the started synchronous communication in the transfer flag storage unit 440 after the notification is executed (step S313).

If the start flag 443 does not exist (NO in step S311), this corresponds to a state in which a packet is being transferred, so the synchronization error detection unit 450 does not execute processing. The processing in the steps after step S314 is the same as the processing in the steps after step S120 in the flowchart shown in FIG.

FIG. 22 is a flowchart showing the operation of the interrupt processing unit 420 in the third embodiment. FIG. 22 shows the operation of the interrupt processing unit 420 when recording a synchronous communication end flag.

The difference from the first embodiment is the operation of the interrupt processing unit 420 when there is a corresponding time slot in the time management unit 120 in step S326. If there is a time slot (YES in step S326), either synchronous communication is started or communication is in progress.

When the synchronous communication ends, an end flag 444 is added to the packet. In the present embodiment, the interrupt processing unit 420 checks whether or not the end flag 444 is set in the packet (step S330).

When the end flag 444 exists (YES in step S331), the synchronization error detecting unit 450 records the end flag 444 in the transfer flag information of the time slot number 441 corresponding to the current time (step S332). When the end flag 444 does not exist (NO in step S331), it is considered that synchronous communication is being executed, and the interrupt processing unit 420 executes the processes after step S333.

FIG. 23 is a flowchart showing the operation of synchronization error detection by the synchronization error detection unit 450. The synchronization error detection unit 450 checks the transfer flag storage unit 440, and acquires transfer flag information corresponding to the previous time slot number of the time slot number 441 corresponding to the current time (step S341).

Next, it is confirmed whether or not the end flag 444 is recorded in the acquired previous transfer flag information (step S342).

If the end flag 444 exists (YES in step S343), the synchronous communication has ended within the time corresponding to the time slot, so the synchronization error detection unit 450 ends the synchronization error detection operation as it is.

If the end flag 444 does not exist (NO in step S343), the synchronization error detection unit 450 notifies the synchronization management server 100 of synchronization failure because the synchronous communication has not ended within the time corresponding to the time slot (step S343). S344). After notifying the synchronization failure, the synchronization error detection unit 450 ends the synchronization error detection operation.

[Description of effects]
Next, the effect of this embodiment will be described. In this embodiment, a start flag and an end flag are used in synchronous communication, and the synchronization corresponding to the next time slot is set before the end flag is set in the transfer flag information corresponding to the time slot number corresponding to synchronous communication. When the communication start flag arrives, the synchronous communication corresponding to the previous time slot is considered not to be completed within the time. Therefore, when using the synchronous scheduling system 10 of this embodiment, the user can detect the error of synchronous communication.

As mentioned above, although the more preferable embodiment of this invention was described, this invention is not limited to the said embodiment.

Further, as described above, the computer 200 can also perform normal communication processing for performing network protocol processing. When the computer 200 performs only transmission processing, the computer connection management information stored in the computer connection management unit 410 may be composed of a time slot number 411 and an interface identifier.

Next, the outline of the present invention will be described. FIG. 24 is a block diagram showing an outline of a relay device according to the present invention. The relay device 30 according to the present invention is a relay device for transferring a packet, and at least a number (for example, a time slot number 361) for identifying a time slot that is a section divided by a predetermined time interval, and the relay device 30 The connection information storage unit 31 (for example, the network connection management unit 360) that stores connection information configured with the output port number (for example, the second port number 363) and the connection information storage unit 31 have the current time. The connection information processing unit 32 (for example, timer execution unit 340) that determines whether or not connection information corresponding to the time slot exists, and the connection information processing unit 32 has connection information corresponding to the connection information storage unit 31. If it is determined, the packet transfer unit 33 transfers the packet to the output port indicated by the output port number in the corresponding connection information. Having.

With such a configuration, the relay device can contribute to synchronous communication that does not perform network protocol processing.

The connection information further includes an input port number (for example, the first port number 362) in the relay device 30, and the packet transfer unit 33 determines that the corresponding connection information exists. The input port number in the corresponding connection information confirms whether the packet to be transferred matches the port number input to the relay device 30, and if it matches, the packet is assigned to the output port in the corresponding connection information. You may transfer to the output port which a number shows.

Further, the relay device 30 receives the number identifying the time slot, the connection source information corresponding to the number identifying the time slot, and the connection destination information from the management device capable of communicating with the relay device 30, and the connection source information Based on the connection destination information, the port number of the input port and the port number of the output port in the relay device 30 may be obtained.

FIG. 25 is a block diagram showing an outline of a communication apparatus according to the present invention. A communication device 40 according to the present invention is a communication device that performs packet communication, and includes a number (for example, a time slot number 411) that identifies a time slot that is a section divided at least by a predetermined time interval, and a time slot. The connection management information storage unit 41 (for example, the computer connection management unit 410) that stores connection management information including the identifier of the interface (for example, the interface 412) corresponding to the identification number, and the connection management information storage unit 41 The communication processing unit 42 (for example, an interrupt processing unit 420) that determines whether or not connection management information corresponding to the time slot to which the time belongs exists, and the communication processing unit 42 determines that the corresponding connection management information exists. The packet is sent to the interface indicated by the interface identifier included in the relevant connection management information. Sending to.

With such a configuration, the communication device can contribute to synchronous communication that does not perform network protocol processing.

FIG. 26 is a block diagram showing another outline of the communication apparatus according to the present invention. A communication device 50 according to the present invention is a communication device that performs packet communication, and includes a number (for example, a time slot number 411) for identifying a time slot that is a section divided at least by a predetermined time interval, and a time slot. A connection management information storage unit 51 (for example, a computer connection management unit 410) that stores connection management information including a memory map address (for example, a memory map address 414) corresponding to an identification number, and a connection management information storage unit 51 A communication processing unit 52 (for example, an interrupt processing unit 420) that determines whether or not there is connection management information corresponding to the time slot to which the current time belongs, and the communication processing unit 52 has the corresponding connection management information. If it is determined, the data is received in the buffer area indicated by the memory map address included in the corresponding connection management information. To copy the packet.

The communication device 50 further includes a synchronization error detection unit (for example, a synchronization error detection unit 450), and the communication processing unit 52 has a start flag (for example, a start flag 443) in the packet when the packet is received. If the start flag is included, it is possible to cause the synchronization error detection unit to execute the synchronization error detection process.

FIG. 27 is a block diagram showing an outline of the management apparatus according to the present invention. The management device 60 according to the present invention is a management device capable of communicating with a communication device and a relay device that transfers packets between the communication devices, and identifies a time slot that is a section divided at least by a predetermined time interval. A time management information storage unit 61 (for example, time management unit 120) that stores time management information including a number (for example, time slot number 111), and connection source information (for example, connection source IP address 112) from the communication device. Synchronization that accepts a transmission request of a packet including connection destination information (for example, connection destination IP address 113) and assigns the received connection source information and connection destination information to time management information to which connection source information and connection destination information are not assigned. Schedule setting unit 62 (for example, synchronization schedule setting unit 130) and time management assigned to the synchronization schedule setting unit 62 Broadcast a synchronization schedule transmission unit 63 to be transmitted to the communication apparatus and the relay apparatus (e.g., synchronization schedule transmission unit 140) and a.

With such a configuration, the management apparatus can contribute to synchronous communication that does not perform network protocol processing.

The management device 60 also includes a route information storage unit (for example, a network route information storage unit 160) that stores packet communication route information, and the synchronization schedule transmission unit 63 performs time management when transmitting time management information. The time management information may be transmitted to the device included in the communication path information corresponding to the information.

Although the present invention has been described with reference to the embodiments and examples, the present invention is not limited to the above embodiments and examples. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

This application claims priority based on Japanese Patent Application No. 2014-125669 filed on June 18, 2014, the entire disclosure of which is incorporated herein.

Industrial applicability

The present invention is suitably applied to communication between telephone exchange servers that require real-time communication and use of a stock price distribution network. The present invention is also suitably applied to real-time communication applications used for remote surgery.

A to D Router 1 Parallel computer 2 PE communication network 3 PE
4 Inactivation function 5 Condition setting function 6 Allocation function 7 Interrupt generation function 8 Activation function 10 Synchronous scheduling system 30 Relay device 31 Connection information storage unit 32 Connection information processing unit 33

Packet transfer unit

40, 50

Communication device

41, 51 Connection Management

information storage unit

42, 52 Communication processing unit 60 Management device 61 Time management information storage unit 62 Synchronization schedule setting unit 63 Synchronization schedule transmission unit 100 Synchronization management server 101 to 104 Server 110 Time synchronization unit 111 Time slot number 112 Connection source IP address 113 Connection destination IP address 120 Time management unit 130 Synchronization schedule setting unit 140 Synchronization schedule transmission unit 150 Network route information generation unit 160 Network route information storage unit 161 Connection source server 162 Connection destination server 163 Route information 170 Synchronization Error notification unit 200 computer 210 time synchronization unit 220 application 300 network switch 310 time synchronization unit 320 path conversion unit 330 synchronization schedule reception unit 340 timer execution unit 350 physical socket connection unit 360 network connection management unit 361 time slot number 362 first port Number 363 Second port number 400 OS
410 Computer connection management unit 411 Time slot number 412 Interface 413 Interrupt number 414 Memory map address 420 Interrupt processing unit 430 Synchronization schedule reception unit 440 Transfer flag storage unit 441 Time slot number 442 Transfer flag 443 Start flag 444 End flag 450 Synchronization

error detection unit

500, 600 networks

Claims

A relay device for forwarding packets,
A connection information storage unit for storing connection information composed of a number identifying a time slot that is a section divided at least by a predetermined time interval, and an output port number in the relay device;
A connection information processing unit for determining whether or not connection information corresponding to a time slot to which a current time belongs exists in the connection information storage unit;
A packet transfer unit that transfers a packet to an output port indicated by an output port number in the corresponding connection information when the connection information processing unit determines that the connection information corresponding to the connection information storage unit exists; Have relay device.
The connection information further includes an input port number in the relay device,
The packet transfer unit, when the connection information processing unit determines that the corresponding connection information exists, the input port number in the corresponding connection information is the port of the port where the packet to be transferred is input to the relay device. The relay apparatus according to claim 1, wherein whether the number matches the number is checked, and if the number matches, the packet is transferred to an output port indicated by the number of the output port in the corresponding connection information.
The connection source information and the connection destination information corresponding to the number for identifying the time slot and the number for identifying the time slot are received from the management device that can communicate with the relay device, and the connection source information and the connection destination information are received. The relay apparatus according to claim 2, wherein a port number of an input port and a port number of an output port are obtained based on the relay apparatus.
A communication device that performs packet communication,
A connection management information storage unit that stores connection management information including a number that identifies a time slot that is a section divided at least by a predetermined time interval, and an identifier of an interface corresponding to the number that identifies the time slot;
The connection management information storage unit includes a communication processing unit that determines whether or not connection management information corresponding to a time slot to which the current time belongs exists,
A communication device that, when the communication processing unit determines that corresponding connection management information exists, transmits a packet to an interface indicated by an interface identifier included in the corresponding connection management information.
A communication device that performs packet communication,
A connection management information storage unit for storing connection management information including a number for identifying a time slot that is a section divided at least by a predetermined time interval, and a memory map address corresponding to the number for identifying the time slot;
The connection management information storage unit includes a communication processing unit that determines whether or not connection management information corresponding to a time slot to which the current time belongs exists,
A communication device that copies a received packet to a buffer area indicated by a memory map address included in the corresponding connection management information when the communication processing unit determines that the corresponding connection management information exists.
A synchronization error detection unit;
When the packet is received, the communication processing unit checks whether the packet includes a start flag. If the packet includes a start flag, the communication processing unit executes a synchronization error detection process on the synchronization error detection unit. The communication device according to claim 5.
A management device capable of communicating with a communication device and a relay device that transfers packets between the communication devices,
A time management information storage unit for storing time management information including a number for identifying a time slot that is a section divided at least by a predetermined time interval;
A transmission request for a packet including connection source information and connection destination information is received from a communication device, and the received connection source information and connection destination information are added to the time management information to which connection source information and connection destination information are not allocated. A synchronization schedule setting section to be assigned;
A management apparatus comprising: a synchronization schedule transmission unit that transmits the time management information assigned to the synchronization schedule setting unit to the communication device and the relay device.
A path information storage unit for storing packet communication path information;
The management device according to claim 7, wherein the synchronization schedule transmission unit transmits the time management information to a device included in the communication path information corresponding to the time management information when transmitting the time management information.
A relay method executed in the relay device,
Storing connection information composed of a number identifying a time slot that is a section divided at least by a predetermined time interval, and an output port number in the relay device;
Determining whether there is connection information corresponding to the time slot to which the current time belongs in the stored connection information;
A relay method for transferring a packet to an output port indicated by an output port number in the corresponding connection information when it is determined that the corresponding connection information exists.
A relay program executed in the relay device,
In the relay device,
A storage process for storing connection information composed of a number for identifying a time slot that is a section divided by at least a predetermined time interval and an output port number in the relay device;
A connection information process for determining whether or not connection information corresponding to a time slot to which a current time belongs exists in the connection information stored in the storage process; and A relay program for causing a transfer process to transfer a packet to an output port indicated by an output port number in the corresponding connection information when it is determined.