JP6011109B2 - Information processing system and information processing method - Google Patents

Description

  The present invention relates to an information processing apparatus and an information processing method, and more particularly to an information processing apparatus and an information processing method that perform a desired process only on a processing target frame.

  In recent years, the communication rate of data flowing through a network has been increasing year by year. If the communication rate continues to rise, a problem has been pointed out that processing for analyzing and extracting input from the network (hereinafter referred to as “analysis extraction processing”) cannot be performed in real time. An example of the analysis and extraction processing here is the case of stock trading. (1) Extract only the information of the specified stock from the time series data of stocks and stock prices input from the network. (2) Smoothing ( For example, a process of averaging the numerical values of a plurality of adjacent points).

  As a countermeasure to such a problem, one of the methods for speeding up the processing is to offload a part of the analysis and extraction processing currently performed by the application (AP) of the server to the hardware accelerator. A method is mentioned. As an example of a hardware accelerator, there is a network interface card (NIC: Network Interface Card) equipped with a reconfigurable device that can write a process specified by the user himself (hereinafter referred to as “user specified process”). Can be mentioned. An example of a reconfigurable device is an FPGA (Field Programmable Gate Array).

  When a NIC (FPGA-NIC) equipped with an FPGA is used as a hardware accelerator, a part of processing performed in the server AP is offloaded. That is, the FPGA-NIC is incorporated in the front stage of the server, and a circuit for executing a part of the processing is programmed in the FPGA. Since the offloaded processing is executed by a dedicated circuit, the processing amount per unit time executed by the server AP increases, so that the processing speed can be increased.

  The user selects an offload target process from the processes performed by the server AP and describes the process contents. Next, the user performs operations such as logic synthesis, conversion to a net list, and placement and routing for the described processing content, and then writes the processing content, specifically, information to be programmed into the FPGA into the FPGA-NIC.

  However, the actual FPGA-NIC in which the processing contents are written does not always operate as expected. This is because there is a possibility that a behavior different from the expectation may be exhibited due to a description mistake of processing contents by the user or other design problems. Therefore, it is necessary to perform verification work to confirm whether or not the operation works as expected, to find out the cause of the behavior different from the expectation, and to reproduce the situation where the behavior different from the expectation occurs.

  In general, verification using an actual machine is more difficult than debugging and verification in software. Therefore, in order to develop a system that performs offloading using an accelerator, it is effective to perform software simulation of offloading processing before writing the content of the offloading processing to the actual machine. Furthermore, it is desirable to perform not only the offload process but also a simulation including an AP executed on the server.

  In a normal server, an OS (Operating System) in the server receives an input frame from the network, analyzes the destination and request contents, and processes them. The processing contents of the OS include, for example, replying to an input frame request and delivering the input frame to an AP designated as a destination. Input frames from the network include IP (Internet Protocol) packets such as TCP (Transmission Control Protocol) and UDP (User Datagram Protocol), ping packets, and ARP (Address Resolution Protocol) frames. In this specification, a packet and a frame are not particularly distinguished and are collectively referred to as “frame”.

  It is difficult for the user to change the OS kernel of the server and the transmission / reception interface of the application to a different one from the normal time only for the simulation. Also, the environment of the server OS and application varies depending on the server. For this reason, it is troublesome to change the setting every time at the start and end of the simulation or to perform different settings depending on the server environment. Therefore, it is desirable that the simulation environment can be constructed without modifying the OS, and the OS kernel and application transmission / reception interfaces need not be changed.

  An example of a network protocol stack constructed without modifying the OS is described in Non-Patent Document 1. According to the method of Non-Patent Document 1, a new network protocol development environment can be constructed without modifying the OS. FIG. 26 is a block diagram illustrating a configuration example of the information processing apparatus of Non-Patent Document 1. The user level protocol stack 400 of Non-Patent Document 1 acquires a copy of the input packet from the OS 200, performs protocol processing, and then transmits the packet to the application designated as the destination. In this specification, “the destination is the application” means that the socket pair (IP address, port number) assigned to the application matches the destination information of the packet. Hereinafter, a frame whose destination is an application is referred to as an “application (AP) addressed frame”. Note that “user level” means including user-defined processing.

  Hereinafter, the configuration of the OS 200 will be described in detail. The OS 200 includes a packet copy unit 240, an OS filter unit 220, and an OS main function 230.

  The packet copy unit 240 acquires only the packet addressed to the AP from the input frame. The packet copy unit 240 determines whether the network input frame is a packet of the target protocol, acquires a copy of the packet when the condition is satisfied, and transfers it to the user level protocol stack 400. The determination and copying of the input frame can be realized by using the OS packet capture function.

  The OS filter unit 220 filters the input frame. Then, the OS filter unit 220 determines whether the destination of the input frame is an application, and when it is destined for the application, discards the packet. This is to prevent the same packet from the OS 200 from being sent to the application via the user level protocol stack 400 and the OS filter unit 220.

The operation when a packet addressed to AP is input to the system of Non-Patent Document 1 from the network will be described with reference to FIG.
(Step 1) The OS filter unit 220 discards the AP-addressed packet in the input frame.
(Step 2) The packet copy unit 240 acquires a copy of the packet addressed to the AP from the input frame.
(Step 3) The user level protocol stack 400 performs protocol processing on the frame acquired by the packet copy unit 240.
(Step 4) The user level protocol stack 400 transmits a packet to the AP.
(Step 5) The OS main function 230 analyzes the header of the packet transmitted from the user level protocol stack 400 and transfers it to the designated destination (application).

  Patent Document 1 discloses a simulation system, method, and program inserted between terminal devices for network connection testing of devices such as printers. According to the method of Patent Document 1, it is possible to construct a large-scale complex virtual network between a terminal and a device and perform a highly accurate connection test.

  In order to construct a simulation environment for NIC offload processing, it is necessary to acquire raw network traffic as an input. The “live network traffic” here includes not only the frames addressed to the AP but also, for example, an Internet Control Message Protocol (ICMP) frame used for checking whether the server is alive, an ARP frame used for address resolution, and the like (hereinafter referred to as “live network traffic”). These frames are referred to as “non-application (non-AP) addressed frames”). Also in the method of Patent Document 1, as in the case of Non-Patent Document 1, when raw network traffic is acquired using the frame copy function of the OS, the same frame is generated twice for both the AP-addressed frame and the non-AP-addressed frame. To do.

  Therefore, it is necessary to simulate the operation of the NIC so that the AP-addressed frame reaches the AP via the NIC simulator, and the AP's AP-addressed frame processing and the OS non-AP-addressed frame processing do not overlap. .

JP2009-164706 (6th page, FIG. 1)

E. Ely et al., Alpine: A User-Level Infrastructure for Network Protocol Development, In Proc. Of the third conference on USENIX Symposium on Internet Technologies and Systems, (USA), USENIX Association, 2001 March, Volume 3, p. 15

  In order to build a simulation environment for NIC offload processing, it is necessary to acquire all network traffic as input. However, when network traffic is acquired using the OS frame copy function as in the techniques of Patent Document 1 and Non-Patent Document 1, two types of flows occur for both AP-addressed frames and non-AP-addressed frames. If processing is also performed on the duplicated AP-addressed frame and non-AP-addressed frame, problems due to duplicate reception of the same frame by the AP and increase in network traffic due to duplicate transmission of response frames by the OS occur. In the latter case, in the worst case, a problem such as disconnection occurs.

  Therefore, in order to perform not only the NIC offload processing but also the overall operation simulation, the AP-addressed frame reaches the AP via the NIC simulator, and the AP's AP-addressed frame processing and OS non-AP It is necessary to avoid duplication in the destination frame processing.

It is necessary not only to perform a simulation but also to perform an actual operation so that the same frame does not occur twice.
(Object of invention)
An object of the present invention is to provide an information processing apparatus and an information processing method capable of performing desired processing on a frame to be processed without generating the same packet twice.

  The information processing apparatus according to the present invention includes: a filter management unit that sets a distribution condition for an input frame input from the outside and a first filter condition for a frame to be output to the outside; A filter that outputs the output frame to the outside according to the first filter condition, a processing unit that performs the first process on the input frame from the distribution unit, outputs the output frame to the filter unit as an output frame, and the first filter condition And a section.

  The information processing method of the present invention sets a sorting condition for an input frame input from the outside and a first filter condition for a frame to be output to the outside, and outputs the input frame to the inside or the outside according to the sorting condition. A first process is performed on the frame, the frame is output as an output frame, and the output frame is output to the outside or discarded according to the first filter condition.

  The information processing apparatus and the information processing method of the present invention have an effect that desired processing can be performed on a processing target frame without generating the same packet twice.

It is a block diagram which shows the structure of the information processing apparatus of the 1st Embodiment of this invention. It is a block diagram which shows the internal structure of OS of the 1st Embodiment of this invention. It is an example of the setting content of the OS filter part of the 1st Embodiment of this invention. It is a figure which shows the setting operation | movement of the OS filter part of the 1st Embodiment of this invention. It is a figure which shows the processing operation with respect to the input frame addressed to AP of OS of the 1st Embodiment of this invention. It is a figure which shows the processing operation with respect to the non-AP addressed input frame of OS of the 1st Embodiment of this invention. It is a block diagram which shows the internal structure of the NIC simulator of the 1st Embodiment of this invention. It is a block diagram which shows the structure of the process part of the NIC simulator of the 1st Embodiment of this invention. It is a figure which shows operation | movement of the process part of the NIC simulator of the 1st Embodiment of this invention. It is an example of the setting content of the user filter of the NIC simulator of the 1st Embodiment of this invention. It is a block diagram which shows the internal structure of the filter management part of the NIC simulator of the 1st Embodiment of this invention. It is an example of the setting content of the filter management part of the NIC simulator of the 1st Embodiment of this invention. It is a figure which shows operation | movement of the filter management part of the NIC simulator of the 1st Embodiment of this invention. It is a figure which shows the setting operation | movement of a user filter part of the NIC simulator of the 1st Embodiment of this invention. It is a figure which shows the processing operation with respect to the input frame addressed to AP of the NIC simulator of the 1st Embodiment of this invention. It is a figure which shows the processing operation with respect to the non-AP addressed input frame of the NIC simulator of the 1st Embodiment of this invention. It is a figure which shows the setting operation | movement by the filter management part of the 1st Embodiment of this invention. It is a figure which shows the whole operation | movement of the process with respect to the input frame addressed to AP of the 1st Embodiment of this invention. It is a figure which shows the whole operation | movement of the process with respect to the non-AP addressed input frame of the 1st Embodiment of this invention. It is a block diagram which shows the essential structure of the information processing system of the 1st Embodiment of this invention. It is a block diagram which shows the structure of the information processing system of the 2nd Embodiment of this invention. It is a block diagram which shows the internal structure of the filter management part of the NIC simulator of the 2nd Embodiment of this invention. It is a figure which shows filter setting operation | movement of the information processing system of the 2nd Embodiment of this invention. It is a figure which shows the whole operation | movement of the process with respect to the input frame addressed to AP of the information processing system of the 2nd Embodiment of this invention. It is a figure which shows the whole operation | movement of the process with respect to the non-AP addressed input frame of the information processing system of the 2nd Embodiment of this invention. FIG. 11 is a block diagram illustrating a configuration example of an information processing apparatus described in Non-Patent Document 1. It is a figure which shows the processing operation with respect to the input packet addressed to AP of the information processing apparatus described in the nonpatent literature 1.

(First embodiment)
Next, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the information processing apparatus according to the first embodiment of this invention.

  As shown in FIG. 1, the information processing apparatus according to the first embodiment for carrying out the present invention includes an OS 200, a NIC simulator 100, and an application executed under the management of the OS 200.

  The NIC simulator 100 is a simulator that is executed as a user application under the management of the OS 200. The NIC simulator 100 acquires a copy of the input frame from the network as an input, and performs a software simulation of logic that allows the user to rewrite the process. Details of the NIC simulator 100 will be described later.

  With the configuration of the OS 200 and the NIC simulator 100 described above, the AP-addressed frame is transferred to the application via the NIC simulator 100. For one non-AP addressed frame, the OS 200 performs a process designated for each frame. By performing the processing classified according to the frame destination as described above, it is possible to perform the software simulation of the NIC including the processing unit that executes the user-defined processing without modifying the OS kernel and the transmission / reception interface of the application. .

  Hereinafter, the internal configuration and operation of the OS 200 will be described in order. FIG. 2 is a block diagram showing the internal configuration of the OS 200. The OS 200 includes a frame copy unit 210, an OS filter unit 220, and an OS main function 230.

  The frame copy unit 210 acquires a copy of the network input frame and outputs it to the NIC simulator 100. The frame copy unit 210 can be realized by using an OS function called packet capture.

  The OS filter unit 220 determines whether to pass or discard the input frame, and passes only a frame that satisfies the condition. Specific conditions are set such that the AP-addressed frame is discarded and the non-AP-addressed frame is allowed to pass based on the setting contents of the filter management unit 110 in the NIC simulator 100 described later. The OS filter unit 220 can be realized using an OS function generally called a firewall or a packet filter.

  The contents of the OS filter unit 220 are set as shown in FIG. 3, for example. In other words, the designated protocol to be filtered, the destination port number designated for each designated protocol, and the designated operation for the frame corresponding to the designated protocol and destination port number are set.

  Examples of the designated protocol include TCP and UDP. Examples of the designated operation include passing (ACCEPT) and discarding (DROP). In the setting example of FIG. 3, the OS filter unit 220 is set to discard the frame addressed to the 11111 port by the TCP protocol and the frame addressed to the 22222 port by the UDP protocol, and to pass all frames of other protocols.

The operation of each unit of the OS 200 will be described with reference to FIGS. First, the operation when setting the OS filter unit 220 will be described with reference to FIG. The operation when setting the OS filter unit 220 is as follows.
(1) The filter management unit 110 of the NIC simulator 100 sets a frame to be passed and a frame to be discarded to the OS filter unit 220. Specifically, the OS filter unit 220 is configured to discard the AP-addressed frame and pass the non-AP-addressed frame.

Next, the operation when the OS 200 processes the AP-addressed frame will be described with reference to FIG. The operation when processing a frame addressed to AP is as follows.
(1) The frame copy unit 210 copies an input frame from the network.
(2) The OS filter unit 220 discards the AP-addressed frame based on the setting by the filter management unit 110.
(3) The frame copied by the frame copy unit 210 is input to the NIC simulator 100. The frames input at this time include not only AP-addressed frames but also non-AP-addressed frames.
(4) The OS main function 230 receives the frame output from the NIC simulator 100 and analyzes the contents of the frame.
(5) The OS main function 230 transfers the AP-addressed frame to the application.

The operation at the time of processing the frame addressed to the non-AP to which the OS 200 is input will be described using FIG. The operation when processing an input frame addressed to a non-AP is as follows.
(1) The frame copy unit 210 copies an input frame from the network.
(2) The OS filter unit 220 passes the non-AP-addressed frame based on the setting by the filter management unit 110.
(3) The OS main function 230 performs a process corresponding to the content of the frame on the frame that has passed through the OS filter unit 220.
(4) The frame copied by the frame copy unit 210 is input to the NIC simulator 100. As described above, the frames input at this time include not only non-AP-addressed frames but also AP-addressed frames.

  As described above, the OS 200 transfers a copy of the AP-addressed frame to the NIC simulator 100 on the one hand, and blocks it on the OS filter unit 220 on the other hand. The OS 200 processes the non-AP addressed frame.

  Next, the internal configuration and operation of the NIC simulator 100 will be described in order. FIG. 7 is a block diagram showing an internal configuration of the NIC simulator 100. The NIC simulator 100 includes a frame reception unit 120, a processing unit 130, a user filter unit 140, a frame transmission unit 150, and a filter management unit 110.

  The frame receiving unit 120 receives an input frame from the frame copy unit 210 of the OS 200. The frame receiving unit 120 may buffer the received frame.

  The processing unit 130 performs a software simulation for user-defined processing using the frame received from the frame receiving unit 120 as an input frame. Alternatively, the processing unit 130 may be a hardware simulator in which an operation is described in a hardware description language.

  The target of the software simulation by the processing unit 130 is user-defined processing that is scheduled to be executed in the NIC. User-defined processing in the NIC may be executed by hardware or software.

  The content of the user-defined process may be a fixed process or a changeable process. The contents of the user-defined process are set in a device such as an FPGA that can write the circuit configuration, and a memory that stores a program for the user-defined process, which are installed in the NIC after confirming the normal operation of the process by simulation. The When the processing can be changed, a device such as an FPGA that can rewrite the circuit configuration, a programmable memory, or the like is mounted on the NIC.

  FIG. 8 is a block diagram illustrating an internal configuration of the processing unit 130. The processing unit 130 includes an input / output interface (Interface. I / F) 131 and a verification logic unit 132. The verification logic unit 132 is a functional block that can be rewritten by the user. The user describes the processing to be realized, that is, the contents of the user-defined processing in the verification logic unit 132. The input / output I / F 131 represents input / output connection specifications of the functional block.

The operation of the processing unit 130 will be described with reference to FIG.
(1) The input / output I / F 131 receives an input frame from the frame receiving unit 120. The input / output I / F 131 converts the input frame into an input signal sequence that satisfies the connection protocol (protocol) of the input / output signal of the verification logic unit 132. The input / output I / F 131 may buffer the input frame.
(2) The input / output I / F 131 passes the input signal sequence to the verification logic unit 132.
(3) The input / output I / F 131 receives the output signal sequence from the verification logic unit 132.
(4) The input / output I / F 131 converts the output signal sequence into an output frame, and transmits the output frame to the user filter unit. The input / output I / F may buffer the output signal string.

  In this way, the user can verify whether the user-defined process described by himself / herself outputs as expected.

  Returning to the description of FIG. 7, the user filter unit 140 determines whether or not to pass the output frame of the processing unit 130 based on the setting of the filter management unit 110 described later, and passes only the frames that satisfy the condition.

  The contents of the user filter unit 140 are set as shown in FIG. 10, for example. In other words, the designated protocol to be filtered, the destination port number designated for each designated protocol, and the designated operation for the frame corresponding to the designated protocol and destination port number are set.

  Examples of protocols include TCP and UDP. Examples of the designated operation include passing, (ACCEPT) and discarding (DROP). In the setting example of FIG. 10, the user filter unit 140 is set to pass a frame addressed to the 11111 port in the TCP protocol and a frame addressed to the 22222 port in the UDP protocol, and discard all frames of other protocols. .

  The frame transmission unit 150 receives the output frame that has passed through the user filter unit 140 and transmits it to the application.

  The filter management unit 110 performs setting so that the OS filter unit 220 discards the AP-addressed frame and the user filter unit 140 passes the frame. The internal configuration of the filter management unit 110 is shown in FIG. The filter management unit 110 includes a setting information storage unit 111, a user filter setting unit 112, and an OS filter setting unit 113.

  The setting information storage unit 111 holds a list of application communication ports. The setting information storage unit 111 is updated when the application is started / finished. The stored content of the setting information storage unit 111 may be a text file.

  The setting information storage unit 111 holds contents as shown in FIG. 12, for example. That is, the setting information storage unit 111 holds, for example, an application protocol and a port number. In the example of FIG. 12, the port number 11111 is assigned to the TCP protocol and the number 22222 is assigned to the UDP protocol as the port number of the application.

  The filter management unit 110 sets the frames to these destinations so that the OS filter unit 220 discards the frames and the user filter unit 140 passes them.

The operation of the filter management unit 110 will be described with reference to FIG.
(1) The setting information storage unit 111 is transmitted to the user filter setting unit 112.
(2) The user filter setting unit 112 sets the user filter unit 140 so as to pass only frames addressed to the communication port of the application held in the setting information storage unit 111.
(3) The OS filter setting unit 113 performs settings opposite to those of the user filter setting unit 112 for passing and discarding frames. That is, the OS filter unit 220 is set to discard the frame addressed to the communication port of the application described in the setting information storage unit 111.

The operation of each part of the NIC simulator 100 will be described with reference to FIGS. First, the operation when setting the user filter unit 140 will be described with reference to FIG. The operation when setting the user filter unit 140 is as follows.
(1) Based on the content stored in the setting information storage unit 111, the filter management unit 110 sets a frame to be passed and a frame to be discarded to the user filter unit 140. Specifically, the user filter unit 140 is set to pass the AP-addressed frame and discard the non-AP-addressed frame.

Next, the operation when processing the AP-addressed frame input to the NIC simulator 100 will be described with reference to FIG. The operation when processing the input frame addressed to the AP is as follows.
(1) The frame reception unit 120 receives a frame copied by the frame copy unit 210 of the OS 200 as an input frame.
(2) The processing unit 130 receives the input frame received from the frame receiving unit 120 as input and performs software simulation of user-defined processing.
(3) Based on the settings of the filter management unit 110, the user filter unit 140 passes the AP-addressed frame that is the output frame of the processing unit 130.
(4) The frame transmission unit 150 receives the AP-addressed frame that has passed through the user filter unit 140, and transmits it to the application.
(5) The frame transmitted from the frame transmission unit 150 to the application is transferred to the application via the OS main function 230.

The operation when the NIC simulator 100 processes a non-AP addressed frame will be described with reference to FIG. The operation when processing an input frame addressed to a non-AP is as follows.
(1) The frame reception unit 120 receives a frame copied by the frame copy unit 210 of the OS 200 as an input frame.
(2) The processing unit 130 receives the input frame received from the frame receiving unit 120 as input and performs software simulation of user-defined processing.
(3) The user filter unit 140 discards the non-AP addressed frame that is the output frame of the processing unit 130 based on the setting of the filter management unit 110.

  As described above, the NIC simulator 100 blocks the non-AP addressed frame by the user filter unit 140. The frame addressed to the AP is transmitted to the application.

  Next, the overall operation of this embodiment will be described in detail with reference to FIGS. 17, 18, and 19.

First, the filter setting operation by the filter management unit 110 will be described with reference to FIG.
(1) The filter management unit 110 updates the list of the setting information storage unit 111 in the filter management unit 110 when the application is started and ended. Then, the filter management unit 110 performs settings for the user filter unit 140 and the OS filter unit 220 based on the contents held in the setting information storage unit 111.

Next, an overall operation for processing a frame addressed to an AP input from the network will be described with reference to FIG.
(1) The OS filter unit 220 discards the AP-addressed input frame.
(2) The frame copy unit 210 generates a copy of the input frame.
(3) The frame receiving unit 120 acquires the copied input frame and passes it to the processing unit 130.
(4) The processing unit 130 receives the input frame received from the frame receiving unit 120 as input and performs software simulation of user-defined processing.
(5) Based on the setting of the filter management unit 110, the user filter unit 140 passes the AP-addressed frame that is the output frame of the processing unit 130.
(6) The frame transmission unit 150 transmits the frame that has passed through the user filter unit 140 to the application.
(7) The OS 200 processes the frame transmitted from the NIC simulator 100.
(8) The OS 200 analyzes the contents of the frame and transfers the frame to the application.

Further, an overall operation for processing a non-AP addressed frame input from the network will be described with reference to FIG.
(1) The OS filter unit 220 passes non-AP-addressed input frames.
(2) The OS main function 230 performs a process corresponding to the content of the frame on the frame that has passed through the OS filter unit 220.
(3) The frame copy unit 210 generates a copy of the input frame.
(4) The frame receiving unit 120 acquires the copied input frame and passes it to the processing unit 130.
(5) The processing unit 130 receives the input frame received from the frame receiving unit 120 as input and performs software simulation of user-defined processing.
(6) Based on the setting of the filter management unit 110, the user filter unit 140 discards the non-AP addressed frame that is the output of the processing unit 130.

  The effect of this embodiment will be described. In the present embodiment, an AP-addressed frame and a non-AP-addressed frame are input to the NIC simulator. The frame addressed to the AP is delivered to the application after predetermined processing is performed. The non-AP addressed frame is not output from the NIC simulator and is processed by the OS 200. In this way, the same frame does not occur twice for the application and the OS. Therefore, there is an effect that the software simulation of the NIC and the server including the NIC can be performed without causing a problem caused by the duplicated frame.

  Further, in the present embodiment, both the AP-addressed frame and the non-AP-addressed frame are input to the NIC simulator, and the processing unit can confirm the operation in a state where both are input. Even in an actual NIC, both an AP-addressed frame and a non-AP-addressed frame are input from the network. Therefore, the operation of the processing unit confirmed by the NIC simulator is guaranteed to be executed in the same manner even in an actual NIC. Therefore, the NIC simulator can be used as a simulator before a predetermined process is actually performed by the NIC.

  Furthermore, in the NIC simulator of the present embodiment, duplicate frames addressed to the OS are not input to the OS. Therefore, it is not necessary to modify the OS in order to construct a simulation environment. Therefore, there is an effect that it is possible to suppress an increase in management cost due to a change in the transmission / reception interface of the OS kernel or application.

  As can be understood from the description of the present embodiment, the OS 200 functions as a distribution unit that distributes the transfer destination of the input frame. That is, the OS 200 classifies the frame into an AP-addressed frame or a non-AP-addressed frame. Then, the OS 200 transfers the AP-addressed frame to the processing unit 130 via the frame receiving unit 120. In addition, the OS 200 transfers the non-AP addressed frame to the OS main function 230 and performs a predetermined process. The possessed functions other than the distribution function of the OS 200 are not particularly limited. Therefore, the OS 200 can be generalized as a distribution unit.

  The frame receiving unit 120 and the frame transmitting unit 150 are added as independent functional blocks in order to clearly indicate the function of receiving and transmitting a frame. Therefore, the frame receiving unit 120 and the frame transmitting unit 150 are not essential components of the present invention.

  From the above, the essential components for the information processing apparatus of the present embodiment are the filter management unit, the distribution unit, the processing unit, and the filter unit. FIG. 20 is a block diagram showing the configuration of the information processing apparatus 20 having only the components essential for the present invention.

  The filter management unit 21 sets frame distribution conditions by the distribution unit 22 and frame filter conditions by the filter unit 24.

  The distribution unit 22 outputs an input frame input from the outside to the processing unit 23 or the outside in accordance with the frame distribution condition set in the filter management unit 21.

  The processing unit 23 performs predetermined processing on the input frame from the distribution unit 22 and outputs the result to the filter unit 24 as an output frame.

  The filter unit 24 outputs or discards the output frame from the processing unit to the outside according to the filter condition of the frame set in the filter management unit 21.

  By providing the above blocks, the information processing apparatus 20 can perform processing only on the input frames distributed to the processing unit 23 by the distribution unit 22. Only output frames that satisfy a predetermined condition are output to the outside.

  In this way, the information processing apparatus 20 performs a desired process only for a frame that satisfies a certain condition, and the same frame is not output twice.

  The information processing apparatus 20 according to the present embodiment performs user-defined processing using an FPGA-NIC or the like, and then uses it for verifying the operation of the information processing system in which additional processing is performed on the result by the information processing apparatus. can do. That is, whether or not the user-defined process performed on the AP-addressed frame is a desired action can be confirmed by comparing the output frame of the information processing apparatus 20 with the expected value. If the output frame matches the expected value, it can be seen that the user-defined process processed by the information processing apparatus 20 is the desired process.

Therefore, if the user-defined process whose operation has been confirmed is designated as a process in the NIC that can be designated by the user, the information processing system is guaranteed to operate normally.
(Second Embodiment)
Next, a second embodiment of the present invention will be described in detail with reference to the drawings. In the first embodiment, the server executes the application using the FPGA-NIC that performs the user-defined process after confirming the operation of the user-defined process by simulation. On the other hand, in the second embodiment, the user-defined process is simulated on a server different from the server that executes the application using the FPGA-NIC that performs the user-defined process. That is, a server (hereinafter referred to as “simulation server”) that performs simulation of user-defined processing and a server that executes an application using an FPGA-NIC that actually performs user-defined processing (hereinafter “actual operation server”). Is different.)

  FIG. 21 is a block diagram illustrating a configuration of an information processing system according to the second embodiment. The same reference numerals are used for the same components as in FIG. The information processing system according to the second embodiment includes a simulation server 1 that executes the NIC simulator 500 and an actual operation server 2 that executes an application. Hereinafter, detailed configurations of the simulation server 1 and the actual operation server 2 will be described.

  In the actual operation server 2, the application 300 and the sub filter management unit 310 are executed under the management of the OS 200.

  The OS 200 includes an OS main function 230 and an OS filter unit 220. Since the OS filter unit 220 performs the same function as the OS filter unit 220 of the first embodiment, the description thereof is omitted.

  The sub filter management unit 310 sets the OS filter unit 220. The setting contents are based on a request from the filter management unit 110 of the simulation server 1. The secondary filter management unit 310 may be a service that enables remote login from the server 2.

  In the simulation server 1, the NIC simulator 500 is executed under the management of the OS 600. The NIC of the simulation server 1 is set to a promiscuous mode, that is, a mode for receiving all incoming frames. Although the NIC is not shown in FIG. 21, the NIC exists between the network and the server 1.

  The OS 600 includes a frame copy unit 610. The frame copy unit 610 copies only the frame input to the NIC of the simulation server 1 from the traffic flowing on the network to which the simulation server 1 and the actual operation server 2 are connected. To pass. That is, the frame whose destination is the server 2 and the frame copy whose destination is the broadcast address are acquired.

  The NIC simulator 500 includes a frame reception unit 120, a processing unit 130, a user filter unit 140, a frame rewriting unit 560, a frame transmission unit 150, and a filter management unit 110. The frame reception unit 120, the processing unit 130, the user filter unit 140, and the frame transmission unit 150 perform the same functions as those in the first embodiment, and thus description thereof is omitted.

  The frame rewriting unit 560 rewrites the transmission source address with the address of the simulation server 1 for the AP-addressed frame that has passed through the user filter unit 140.

  The filter management unit 510 performs setting so that the OS filter unit 220 of the actual operation server 2 discards the AP-addressed frame, and the user filter unit 140 in the NIC simulator 500 passes. The internal configuration of the filter management unit 510 is shown in FIG. The filter management unit 510 includes a setting information storage unit 111, a user filter setting unit 112, and a remote OS filter setting unit 513.

  The setting information storage unit 111 and the user filter setting unit 112 execute the same functions as those in the first embodiment. The setting information storage unit 111 holds a list of application communication ports. The setting information storage unit 111 is updated when the application is started / finished.

  The user filter setting unit 112 sets the user filter unit 140 so that only frames addressed to the AP described in the setting information storage unit 111 are passed.

  The remote OS filter setting unit 513 performs settings opposite to the user filter setting unit 112 with respect to the passing and discarding of frames with respect to the OS filter unit 220 of the actual operation server 2. That is, the remote OS filter setting unit 513 sends a command to the sub-filter management unit 310 of the actual operation server 2 so that the OS filter unit 220 discards the AP-addressed frame described in the setting information storage unit 111. Let

  Next, the overall operation of this embodiment will be described in detail with reference to FIGS.

First, the operation at the time of setting a filter will be described with reference to FIG. The operation when setting the filter is as follows.
(1) The filter management unit 510 updates the list in the setting information storage unit 111 at the time of starting and ending the application. Then, the filter management unit 510 sets the user filter unit 140 to pass only the AP-addressed frame.
(2) The filter management unit 510 performs a setting opposite to the user filter unit, that is, a setting for discarding the AP-addressed frame, in the OS filter unit 220 in the sub filter management unit 310 on the actual operation server 2. Send a request.
(3) The sub filter management unit 310 sets the OS filter unit 220 to discard the AP-addressed frame in accordance with a request from the filter management unit 110.

Next, an overall operation for processing a frame addressed to an AP input from the network will be described with reference to FIG. The operation when processing a frame addressed to AP is as follows.
(1) The OS filter unit 220 discards the AP-addressed input frame.
(2) The frame copy unit 610 generates a copy of a frame whose destination is the actual operation server 2 and a frame whose destination is a broadcast address among the input frames.
(3) The frame receiving unit 120 acquires the copied input frame and passes it to the processing unit 130. Here, the frame receiving unit 120 may buffer the input frame.
(4) The processing unit 130 performs software simulation of user-defined processing using the input frame received from the frame receiving unit 120 as an input.
(5) Based on the setting of the filter management unit 510, the user filter unit 140 passes the AP-addressed frame that is the output of the processing unit 130.
(6) The frame rewriting unit 560 rewrites the transmission source address of the frame that has passed through the user filter unit 140 with the address of the simulation server 1 and passes it to the frame transmission unit 150.
(7) The frame transmission unit 150 transmits the frame received from the frame rewriting unit 560 to the application of the actual operation server 2.
(8) The frame transmitted from the NIC simulator 500 of the simulation server 1 passes through the OS filter unit 220 of the actual operation server 2.
(9) The OS 200 analyzes the contents of the frame and transfers the frame to the application 300.
(10) The application 300 receives a frame.

Furthermore, an overall operation for processing a non-AP addressed frame input from the network will be described with reference to FIG. The operation when processing a non-AP addressed frame is as follows.
(1) The OS filter unit 220 passes the non-AP addressed frame.
(2) The OS main function 230 performs a process corresponding to the content of the frame on the non-AP addressed frame that has passed through the OS filter unit 220.
(3) The frame copy unit 610 generates a copy of the frame whose destination is the actual operation server 2 and the frame whose destination is the broadcast address among the input frames.
(4) The frame receiving unit 120 acquires the copied input frame and passes it to the processing unit 130.
(5) The processing unit 130 performs software simulation of user-defined processing using the input frame received from the frame receiving unit 120 as an input.
(6) The user filter unit 140 discards the non-AP addressed frame that is the output of the processing unit 130 based on the setting of the filter management unit 510.

  The effect of the second embodiment will be described. In the second embodiment, the NIC simulation and the application are executed on different servers. Therefore, the load on the application execution server (actual operation server 2) is reduced. Therefore, in the second embodiment, in addition to the effect of the first embodiment, there is an effect that an application can be executed at a higher speed during simulation.

  The present invention can be applied to a simulation for a purpose in which a user-defined process is performed by an attached processing apparatus, and then the processed frame is input to the main processing apparatus and processed. Accordingly, for example, when an information processing apparatus that performs processing such as extraction / analysis on stock price information input from a network is realized by a computer, it can be applied to a purpose of performing preprocessing using a NIC or the like.

DESCRIPTION OF SYMBOLS 1 1st communication apparatus 100, 500 NIC simulator 110 Filter management part 111 Setting information storage part 112 User filter setting part 113 OS filter setting part 120 Frame reception part 130 Processing part 140 User filter part 150 Frame transmission part 200, 600 OS
210 Frame Copy Unit 220 OS Filter Unit 230 OS Main Function 240 Packet Copy Unit 300 Application 310 Sub Filter Management Unit 400 User Level Protocol Stack 510 Filter Management Unit 513 Remote OS Filter Setting Unit 560 Frame Rewrite Unit 610 Frame Copy Unit

Claims (6)

An information processing system having a simulator and an operating system,
A first filter condition for a frame to be output from the simulator to the operating system is set , and the input frame output to the operating system according to a distribution condition for the input frame input to the information processing system includes the first filter condition . A filter management unit for sending a command to set the distribution condition so that the frame output to the operating system is not included according to the filter condition ;
A sub-filter management unit that sets the distribution condition according to the command from the filter management unit;
A copy unit that copies the input frame input to the information processing system and outputs the copy to the simulator, and a distribution unit that outputs the input frame to the operating system according to the distribution condition;
A processing unit that performs a first process on the input frame output from the distribution unit to the simulator and outputs the output frame to the filter unit;
A filter unit that outputs or discards the output frame to the operating system according to the first filter condition;
An information processing system comprising: The distribution condition defines a distribution destination according to whether or not the input frame is a target of the first process,
The distribution unit distributes the input frame to the processing unit when the input frame is a target of the first processing according to the distribution condition,
The processing unit performs the first process on the input frame when the input frame is a target of the first process,
The first filter condition defines a filter operation depending on whether the output frame is a target of the first process,
The filter unit outputs the output frame to the operating system when the output frame is the target of the first process according to the first filter condition, and the output frame is the target of the first process. The information processing system according to claim 1, wherein the output frame is discarded when not. The distribution condition defines a distribution destination depending on whether the input frame is a target of the second process and the first process in the operating system ,
The distribution unit distributes the input frame to the processing unit when the input frame is a target of the first process and the second process according to the distribution condition, and the input frame is the second process. Output the input frame to the operating system ,
The first filter condition defines a filter operation depending on whether the output frame is a target of the second process,
The filter unit outputs the output frame to the operating system when the input frame is the target of the second process according to the first filter condition, and the input frame is the target of the second process. 3. The information processing system according to claim 2, wherein the output frame is discarded when not. The distribution condition includes a second filter condition,
The distribution unit, the copy of the input frame using the frame copy function provided operating system output to the processing unit, the second using a filter function of the operating system comprises controlling the distribution unit 4. The information processing system according to claim 1, wherein the input frame is passed according to a filter condition and output to the operating system . By an information processing system having a simulator and an operating system,
A first filter condition for a frame to be output from the simulator to the operating system is set , and the input frame output to the operating system according to a distribution condition for the input frame input to the information processing system includes the first filter condition . Sending a command to set the sorting condition so that the frame output to the operating system is not included by the filter condition of
According to the command, set the distribution condition,
The input frame input to the information processing system is copied and output to the simulator, and the input frame is output to the operating system according to the distribution condition.
Performing a first process on the input frame output to the simulator, outputting as an output frame;
According to the first filter condition, the output frame is output to the operating system or discarded. Determining whether or not the first output frame output using the information processing method according to claim 5 and an expected value frame expected as a result of the first processing match;
When the first output frame and the expected value frame match, the second processing information processor row physician second output frames specified for the second input frame input from the network information processing how, characterized in that to specify that the processing unit of the network interface card, performs the first process of outputting to.

Images