JP6361837B2 - Training apparatus, training method, and training program - Google Patents

Description

  The present invention relates to a training apparatus that trains an operator who is engaged in a plant or factory so that an appropriate response can be taken when an accident or abnormal situation occurs.

  It is necessary to train operators who are engaged in industrial control system plants and factories so that they can take appropriate actions in the event of an accident or abnormal situation. Particularly in recent years, industrial control systems are often connected to the Internet, and accordingly, cyber attacks against industrial control systems are increasing. Therefore, it is necessary to train operators to respond appropriately when the industrial control system is subjected to a cyber attack. As a conventional technique related to such cyber attack training, for example, in Patent Document 1, a user (training instructor) is a simulation that adjusts the difficulty level of training by changing the number of terminals of the system to be trained and its IP address. An apparatus is disclosed.

WO2015-029464

  In order to realize the training device, it is necessary to construct a system simulating the production environment of a plant or factory on the training device, and to generate an accident or abnormal situation on that system. Since the training scenario depends on the system configuration of each plant or factory, if a new scenario occurs frequently, such as an accident or abnormal situation caused by a cyber attack, the training scenario for each individual system Need to be developed frequently. In particular, for systems with different system configurations, such as the functions and network configurations of the devices that make up the production environment, it is necessary to develop a scenario that conforms to this system from scratch, even for the same training scenario. There is a problem that the construction period and cost of the increase.

  The present invention has been made to solve the above problems, and an object of the present invention is to provide a general-purpose training apparatus that can automatically generate the same training scenario for a plurality of systems having different system configurations.

  In order to solve the problems described above, the training apparatus of the present invention includes an information storage unit that stores functional information indicating the function of the device and attack classification information indicating the classification of the execution step of the cyber attack, and a training target system. System configuration information indicating the information of the device to be configured is input, and the function information of the basic attack scenario including the execution step in which the function information and the attack classification information are associated is rewritten to the device of the system configuration information corresponding to this function information. A scenario generation unit that generates an individual attack scenario that is an individual attack scenario that matches the system configuration information of the training target system.

  According to the present invention, the system configuration information of the training target system is input, and the basic attack scenario expressed by the device function and a series of attack steps is rewritten so as to match the system configuration information. By generating an individual attack scenario that is an attack scenario, the same training scenario can be automatically generated for a plurality of systems having different system configurations, and the construction period and cost of training scenario development can be suppressed.

It is a figure which shows one structural example of the training apparatus 1 which concerns on Embodiment 1. FIG. 2 is a diagram illustrating an example of a hardware configuration of a training device 1 according to Embodiment 1. FIG. 3 is a flowchart showing a flow of operations of the training apparatus 1 according to the first embodiment. It is a figure which shows an example of the training object system used in Embodiment 1. FIG. It is a figure which shows an example of the system configuration information. It is a figure which shows an example of the function information. It is a figure which shows an example of the system configuration information. It is a figure which shows an example of the attack classification information 7. FIG. It is a figure which shows an example of basic attack scenario information. 4 is a flowchart (part 1) showing a flow of operations of a scenario generation unit 5; 6 is a flowchart (part 2) illustrating the flow of operations of the scenario generation unit 5; 6 is a flowchart (part 3) illustrating a flow of operations of the scenario generation unit 5; 10 is a flowchart (part 4) showing a flow of operations of the scenario generation unit 5; It is a figure which shows an example of the premise step determination information 9. FIG. It is a figure which shows an example of individual attack scenario information.

Embodiment 1 FIG.
FIG. 1 is a diagram illustrating a configuration example of the training apparatus 1 according to the first embodiment.
In FIG. 1, the training apparatus 1 includes a system configuration input unit 2, a scenario input unit 3, an information storage unit 4, and a scenario generation unit 5.

  The system configuration input unit 2 displays a list of functional information 6 stored in the information storage unit 4 to a user (teacher) or a training device developer (hereinafter collectively referred to as a user). This is an input interface for inputting system configuration information such as a device name and network name existing in the system in association with the displayed function information 6.

  The scenario input unit 3 displays the function information 6, the attack classification information 7, and the scenario template 8 stored in the information storage unit 4 to the user, and is a basic attack scenario that the user wants the trainer to train. This is an input interface for inputting attack scenario information.

  The information storage unit 4 is a memory that stores function information 6, attack classification information 7, scenario templates 8, and prerequisite step determination information 9.

  The scenario generation unit 5 uses the system configuration information input by the system configuration input unit 2, the basic attack scenario information input by the scenario input unit 3, and the premise step determination information 9 stored in the information storage unit 4. To generate individual attack scenarios for individual system configurations. The scenario generation unit 5 includes a determination unit 10, a complement unit 11, and an output unit 12.

Next, the hardware configuration of the training apparatus 1 will be described.
FIG. 2 is a diagram illustrating an example of a hardware configuration of the training apparatus 1 according to the first embodiment.

  The training device 1 is a computer, and each component of the training device 1 can be realized by a program. As a hardware configuration of the training apparatus 1, a processor 20, an external storage memory 21, a main storage memory 22, an input interface 23, and an output interface 24 are connected to the bus.

  The processor 20 is connected to other hardware via a bus and controls these other hardware. The processor 20 is an IC (Integrated Circuit) that performs processing. Specifically, the processor 20 is a CPU (Central Processing Unit) that executes a program.

  The external storage memory 21 is, for example, a ROM (Read Only Memory), a flash memory, a hard disk device, or the like.

  The main memory 22 is, for example, a RAM (Random Access Memory). The information storage unit 4 is realized by the external storage memory 21 or the main storage memory 22. Alternatively, the information storage unit 4 may be realized by both the external storage memory 21 or the main storage memory 22.

  The input interface 23 is a port connected to an input device such as a mouse, a keyboard, and a touch panel. Specifically, the input interface 23 is, for example, a USB (Universal Serial Bus) terminal. The input interface 23 may be a port connected to a LAN (Local Area Network).

  The output interface 24 is a port to which a cable of an output device such as a display is connected. The output interface is, for example, a USB terminal. The display is specifically an LCD (Liquid Crystal Display).

  The program is normally stored in the external storage memory 21, and is loaded into the main storage memory 22 and sequentially read into the processor 20 and executed. This program is a program that realizes the functions described as the system configuration input unit 2, the scenario input unit 3, the information storage unit 4, and the scenario generation unit 5 that constitute the training apparatus 1.

  Further, an operating system (OS) is also stored in the external storage memory 21, and at least a part of the OS is loaded into the main storage memory 22, and the processor 20 executes the above program while executing the OS.

  Further, information and data stored in the information storage unit 4 and information, data, signal values, and variable values indicating processing results of the system configuration input unit 2, the scenario input unit 3, and the scenario generation unit 5 are stored in the main memory 22. Stored as a file.

  2 is merely an example of the hardware configuration of the training apparatus 1, and the hardware configuration of the training apparatus 1 is not limited to the configuration illustrated in FIG. 2 and may be other configurations. Good.

Next, operation | movement of the training apparatus 1 which concerns on Embodiment 1 is demonstrated.
FIG. 3 is a flowchart showing a flow of operation of the training apparatus 1 according to the first embodiment.

First, in step S001, the user inputs system configuration information, which is information related to the configuration of a system to be trained, using the system configuration input unit 2.
FIG. 4 is a diagram illustrating an example of a training target system used in the first embodiment.
In FIG. 4, the training target system includes a network 1 and a network 2. The operating device 1 (30) and the operating device 2 (31) are connected to the network 1, and the operating device 2 (31), the controller 32, the control target device 1 (33), and the control target device 2 ( 34) is connected.

  The user inputs a device name existing in the training target system through the system configuration input unit 2. The training apparatus 1 displays a list of functions of the function information 6 stored in advance in the information storage unit 4 for each input device name, and associates the device name with the function to the user. Be able to do it. As a result of the user selecting the displayed function for the input device name, the training apparatus 1 outputs the system configuration information 1 in which the device name and the function are associated with each other.

FIG. 5 is a diagram illustrating an example of the system configuration information 1.
In FIG. 5, the system configuration information 1 includes a device name input by the user and a function selected by the user.

FIG. 6 is a diagram illustrating an example of the function information 6.
The function information 6 includes a function stored in advance in the information storage unit 4 and a description thereof. The explanation of the function is information for performing input support for the user, and therefore may not be necessary.

  Next, the user inputs a network name existing in the training target system through the system configuration input unit 2. The training device 1 displays the device name input by the user for the input network name, and allows the user to associate the network name with the device name connected to the network. As a result, the training apparatus 1 outputs system configuration information 2.

FIG. 7 is a diagram illustrating an example of the system configuration information 2.
In FIG. 7, the system configuration information 2 includes a network name (in the figure, “network” is indicated by “NW”), connected devices 1 to connected devices 4.

  Next, in step S002, the user uses the scenario input unit 3 to create a basic attack scenario, which is a content that the trainer wants to train, by specifying a series of attack steps.

  The training device 1 displays the attack classification information 7 obtained by dividing the attack scenario in units of attack classification, and the function information 6 serving as the attack source and the attack target, and can associate the attack classification with the device name to the user. Like that. The attack classification information 7 is information stored in advance in the information storage unit 4.

  Further, the user designates an attack step in which the characteristics of the basic attack scenario that the trainer wants to train are markedly displayed from the series of attack steps. The attack step specified here is an attack step that is always set when creating an individual attack scenario according to the system configuration of each training target system. Hereinafter, this attack step is called an essential step. As a result, the training device 1 outputs basic attack scenario information.

FIG. 8 is a diagram illustrating an example of the attack classification information 7.
In FIG. 8, the attack classification information 7 includes an attack classification.

FIG. 9 is a diagram illustrating an example of basic attack scenario information.
In FIG. 9, basic attack scenario information includes attack step numbers. , Attack classification, attack source (function), attack target (function), and information indicating whether it is essential.

  In the present embodiment, the basic attack scenario information is created from scratch, but the basic attack scenario information generated in the past is stored as a template in the scenario template 8 of the information storage unit 4, and this step (S002) It may be loaded at the time of execution. In this case, the user may be able to appropriately specify information indicating the attack classification, attack source, attack target, and necessity.

  Next, in step S003, the scenario generation unit 5 generates an individual attack scenario according to the system configuration input by the user. A detailed processing flow of step S003 will be described later.

  Next, in step S004, the scenario generation unit 5 outputs the generated individual attack scenario and ends the process.

Next, the detailed processing flow of step S003 will be described with reference to FIGS.
FIG. 10 is a flowchart (part 1) illustrating the flow of operations of the scenario generation unit 5.
FIG. 11 is a flowchart (part 2) of the operation flow of the scenario generation unit 5.
FIG. 12 is a flowchart (part 3) illustrating the flow of the operation of the scenario generation unit 5.
FIG. 13 is a flowchart (part 4) of the operation flow of the scenario generation unit 5.

  First, in step S101, the determination unit 10 of the scenario generation unit 5 determines whether or not the function required in the essential step exists in the system. The determination unit 10 refers to the basic attack scenario information, confirms the function required when executing the essential step, and determines whether a device corresponding to the function is registered in the system configuration information 1.

  For example, in the basic attack scenario information of FIG. An essential step is set in No. 4, which is an attack for performing unauthorized communication from the EWS to the PLC. Therefore, the functions required in the essential steps are EWS and PLC, and the determination unit 10 can determine from the system configuration information 1 in FIG. 5 that both the EWS and PLC functions exist in the system. it can. If it is determined in step S101 that it exists, the process proceeds to step S102. If it is determined in step S101 that it does not exist, the process proceeds to step S106, and an individual attack scenario for training cannot be generated, and the process ends.

  Next, in step S102, the determination unit 10 of the scenario generation unit 5 determines whether the function required when executing the essential step is singular or plural. For example, the attack step no. In 1, the required function is singular, and in other attack steps, there are multiple functions. If the required function is singular, the process proceeds to step S104. If there are a plurality of necessary functions, the process proceeds to step S103.

  Next, in step S103, the determination unit 10 of the scenario generation unit 5 determines whether or not a plurality of functions necessary for executing the essential steps are connected to each other. This process is determined by associating the functions of the system configuration information 1 with the names of devices connected to each other in the system configuration information 2. If it is connected, the process proceeds to step S104. When not connected, it progresses to step S106, and the process is complete | finished as the separate attack scenario for training cannot be produced | generated.

Next, in step S104, the determination unit 10 of the scenario generation unit 5 determines a step that is a prerequisite when executing the essential step. For this determination, premise step determination information 9 is used.
FIG. 14 is a diagram illustrating an example of the premise step determination information 9.
In FIG. 14, the prerequisite step determination information 9 is obtained by associating the attack classification of the essential step with the attack classification necessary for the prerequisite step. In the example of the first embodiment, since the essential step is unauthorized communication from EWS, it can be seen from FIG. 14 that infection to EWS is necessary as a premise. After determining the premise step in this way, the process proceeds to step S105.

  Next, in step S105, the determination unit 10 of the scenario generation unit 5 determines whether a step corresponding to the premise step determined in step S104 exists in the basic attack scenario information. In the example of the first embodiment, the attack step No. Since there is an infection from HMI to EWS in No. 3, it is determined in step S105 that it exists, and the process proceeds to step S107. Hereinafter, a step that is a premise of an essential step in the attack steps described in the basic attack scenario information is referred to as a premise step. A step that is neither an essential step nor a prerequisite step is called an optional step. If it is determined in step S105 that it does not exist, the process proceeds to step S106, and an individual attack scenario for training cannot be generated, and the process ends.

  Next, in step S107, the determination unit 10 of the scenario generation unit 5 determines whether a function necessary for executing the premise step exists in the system. Since the determination method is the same as step S101 in the case of the essential step, the description is omitted. If it is determined in step S107 that it exists, the process proceeds to step S108. If it is determined in step S107 that it does not exist, the process proceeds to step S112 or step S113.

  Next, in step S108, the determination unit 10 of the scenario generation unit 5 determines whether the function necessary for executing the premise step is singular or plural. If the required function is singular, the process proceeds to step S110. If there are a plurality of necessary functions, the process proceeds to step S109.

  Next, in step S109, the determination unit 10 of the scenario generation unit 5 determines whether or not functions necessary for executing the premise step are connected to each other. Since the determination method is the same as step S103, the description is omitted. If it is connected in step S109, the process proceeds to step S110. If not connected in step S109, the process proceeds to step S114 or step S115. In the example of the first embodiment, attack step No., which is a premise step, is performed. 3 is infection from HMI to EWS. Since the system configuration information 2 indicates that the operation terminal 2 and the controller are connected, the process proceeds to step S110.

  Next, in step S110, the determination unit 10 of the scenario generation unit 5 registers the function of the essential step and the prerequisite step in the individual attack scenario by changing the device name input by the user. In the process up to step S110, the essential steps and the prerequisite steps are in a state that can be executed in the training target system. Therefore, in step S110, the device name input by the user is assigned to the function required in the essential step and the prerequisite step, and is registered in the individual attack scenario, and the process proceeds to step S111. In the example of the first embodiment, attack step No. which is an essential step. 4 is the controller, the attack target is the control target device 1, and the attack step No. The operation terminal 2 is set as the attack source 3 and the controller is set as the attack target and registered.

  Next, in step S111, the determination unit 10 of the scenario generation unit 5 determines whether or not the determination in step S101 has been performed for all the essential steps set in the basic attack scenario information. When it determines, it progresses to step S116, and when not determining, it progresses to step S101.

  Next, in step S112 and step S113, the complementing unit 11 of the scenario generation unit 5 performs processing in the case where the function required in the premise step does not exist in the system in step S107. In the processing up to step S112 and step S113, since it is determined that there is a function required in the essential step, the function that becomes the attack source in the premise step exists in step S112 and step S113. This is the case. In creating an individual attack scenario, the premise step is required to make the attack as a result, that is, the attack target. Therefore, the content of the step is changed so that the same result as the result of the premise step is obtained from the system configuration information 1, the system configuration information 2, and the attack classification information 7.

  For example, it is assumed that the premise step is infection from HMI to EWS, and there is no HMI. In this case, the process proceeds from step S107 to step S112, and the content of the step is changed so that EWS is directly infected in step S112. There are several types of means for direct infection. For example, there are infection via a USB memory, infection by opening a mail attached file, infection by downloading a file containing malware from a website, and the like. A premise step is created by randomly selecting from these. When training is performed, the reality of the training is increased and the effectiveness of the training is improved by reproducing the attack scenario triggered by the trainer actually performing the infection method selected here.

  Step S113 is processing when the attack classification is internal investigation. The complementing unit 11 of the scenario generation unit 5 performs a step of performing an internal investigation from a function connected to the EWS, and further adds a step of infecting the function.

  As described above, the essential steps can be executed by changing the contents of the prerequisite steps. In the above, although the example of infection and internal investigation was given as attack classification, operation of training device 1 is not restricted to this, and other attacks may be performed.

  Next, in step S114 and step S115, the complementing unit 11 of the scenario generation unit 5 performs processing when the system has the functions necessary for the premise step in step S107 but they are not connected to each other. Do.

  In step S114, the complementing unit 11 of the scenario generation unit 5 sets the secondary infection via the function connected to the target function. If the function does not exist, it is directly infected. The complement unit 11 extracts the premise step from the basic attack scenario information. Next, the complementing unit 11 selects a device having the attack target function of the premise step from the system configuration information 1, is interconnected with this from the system configuration information 2, and also interacts with the attack source function of the premise step. When there is a connected device (referred to as an intermediate device), it is a step of performing secondary infection via the intermediate device. Specifically, it is assumed that the infection progresses in the order of the attack source of the premise step → the intermediate device → the attack target of the premise step. If there is no intermediate device, that is, if the path connecting the attack source of the premise step and the attack target of the premise step cannot be connected through any path, the step of directly infecting the attack target is performed as in step S112 And

  Step S115 is processing when the attack classification is internal investigation. As in step S114, once the intermediate device is once infected, an internal investigation is performed from the intermediate device.

  In step S114 and step S115, as in the case of step S112 and step S113, examples of infection and internal investigation are given as attack categories. However, the operation of the training apparatus 1 is not limited to this, and other attacks are performed. Also good.

  Next, in steps from step S116, processing for registering or deleting the optional step in the individual attack step is performed. The training device 1 refers to the basic attack scenario information and determines the attack step number. Judgment is made in the order of the option step with the smallest. When the contents of the premise step are changed in order to execute the essential step, it is necessary to match with the optional step. First, in step S116, the complement unit 11 of the scenario generation unit 5 selects the option that is the current determination target. It is determined whether the function required when executing the step includes the same function as that required when executing the prerequisite step. When the same thing is included, it progresses to step S117, and when it does not include the same thing, it progresses to step S126.

  Next, in step S117, the complementing unit 11 of the scenario generation unit 5 determines whether or not the content of the step has been changed when determining the relevant premise step. In the example of the first embodiment, the attack step No. of the original basic attack scenario information. In the case of infection from HMI 3 to EWS, HMI was registered as operation terminal 2 and EWS as a controller as it was. Therefore, the content of the premise step is not changed and the process proceeds to step S122. If the contents of the prerequisite step are changed, the process proceeds to step S118.

  Next, in step S118, the complementing unit 11 of the scenario generating unit 5 deletes the optional step in which the content of the prerequisite step is changed. The reason for this is that if the prerequisite step is changed by matching the required step and the prerequisite step, the scenario is completed as an attack scenario, so that an inconsistency occurs if the optional step is left.

  Next, in step S119, the complementing unit 11 of the scenario generation unit 5 determines whether all the optional steps set in the basic attack scenario information have been determined in step S116. When it determines, it progresses to step S120. If not, the process proceeds to step S121, and the next optional step is determined.

Next, in step S120, the output unit 12 of the scenario generation unit 5 outputs the registered individual attack scenario and ends the process. This step S120 corresponds to step S004.
FIG. 15 is a diagram illustrating an example of individual attack scenario information.
In FIG. 15, the individual attack scenario information includes an attack step No. , Attack classification, attack source (device name), attack target (device name).

  Next, step S122 is a step when it is determined in step S117 that the content of the premise step has not been changed. If there is no change, the basic attack scenario information input by the user has both a case where the basic attack scenario information can be executed without change and a case where the information cannot be changed. In step S122, it is determined whether a function that is necessary when executing the determination target option step and that is not the same as the function required in the premise step exists in the system. In the example of the first embodiment, OA-PC corresponds to the function and does not exist in the system. If not, the process proceeds to step S129 or step S130. When it exists, it progresses to step S123.

  Next, in step S123, the complementing unit 11 of the scenario generation unit 5 determines whether or not the function required in the premise step and the function required in the option step are connected to each other. If it is connected, the process proceeds to step S124, and the device name input by the user is assigned and registered in the individual attack scenario. If not connected, the process proceeds to step S125, and the option step is deleted.

  Next, step S126 is a step when it is determined that the function required when executing the optional step to be determined in step S116 does not include the same function as that required when executing the prerequisite step. It is. In such a case, the prerequisite step can be executed regardless of whether or not the corresponding optional step is executed. Therefore, whether or not the function required for executing the optional step exists in the system is determined. Decide whether to register for individual attack scenarios. When it exists, it progresses to step S127. If it does not exist, the process proceeds to step S125, and the step is deleted.

  Next, in step S127, the complementing unit 11 of the scenario generating unit 5 determines whether the function required when executing the optional step is singular or plural. Since the determination method is the same as step S102, the description is omitted. If it is singular, it is determined that it can be executed, and the process proceeds to step S124 where the device name input by the user is assigned and registered in the individual attack scenario. If there are more than one, the process proceeds to step S128.

  Next, in step S128, the complementing unit 11 of the scenario generation unit 5 determines whether or not a plurality of functions necessary for executing the optional step are connected to each other. Since the determination method is the same as step S103, the description is omitted. If it is connected, it is determined that execution is possible, the process proceeds to step S124, and the device name input by the user is assigned and registered in the individual attack scenario. When not connected, it progresses to step S125 and deletes the said step.

  Next, step S129 or step S130 is processing when it is determined in step S122 that the function required for the optional step does not exist in the system. Since the processing method is the same as step S112 and step S113, the description is omitted.

  As described above, in the invention of the first embodiment, the system configuration information of the training target system is used as an input, and the basic attack scenario expressed by the device function and a series of attack steps is rewritten so as to match the system configuration information. In addition, by generating individual attack scenarios that are individual attack scenarios for the training target system, the same training scenario can be automatically generated for multiple systems with different system configurations, and the construction period and cost of training scenario development are reduced. The effect that it can be obtained.

1 training apparatus, 2 system configuration input unit, 3 scenario input unit, 4 information storage unit, 5 scenario generation unit, 6 function information, 7 attack classification information, 8 scenario template, 9 prerequisite step determination information, 10 determination unit, 11 complement Unit, 12 output unit, 20 processor, 21 external storage memory, 22 main storage memory, 23 input interface, 24 output interface, 30 operation device 1, 31 operation device 2, 32 controller, 33 control target device 1, 34 control target device 2.

Claims (7)

An information storage unit for storing functional information indicating the function of the device and attack classification information indicating a classification of cyber attack execution steps;
System configuration information indicating information on the devices constituting the training target system is input, and the function information of the basic attack scenario including the execution step in which the function information and the attack classification information are associated with each other is used as the function information. A training apparatus comprising: a scenario generation unit that rewrites the corresponding system configuration information to the device and generates an individual attack scenario that is an individual attack scenario that matches the system configuration information of the training target system. A mandatory step that is essential for training is specified for the execution step included in the basic attack scenario,
The scenario generation unit
The training apparatus according to claim 1, further comprising a determination unit that determines whether or not the essential step is executable. The training device according to claim 2, wherein the scenario generation unit generates the individual attack scenario when the determination unit determines that the essential step is executable. The training apparatus according to claim 1, wherein the basic attack scenario includes an execution step in which the attack scenario is played back in response to a trainee's action. The information storage unit stores a scenario template that stores the basic attack scenario generated in the past,
The training apparatus according to claim 1, wherein the scenario generation unit generates the individual attack scenario based on the scenario template. A training method for a training apparatus comprising an information storage unit and a scenario generation unit for storing functional information indicating a function of a device and attack classification information indicating a classification of cyber attack execution steps,
The function information of the basic attack scenario including the execution step in which the scenario generation unit is input with system configuration information indicating information on the devices constituting the training target system and associates the function information with the attack classification information. Training method comprising a scenario generation step of generating an individual attack scenario that is an individual attack scenario that matches the system configuration information of the training target system by rewriting the system configuration information corresponding to the function information . Computer
Information storage means for storing functional information indicating the function of the device and attack classification information indicating a classification of cyber attack execution steps;
System configuration information indicating information on the devices constituting the training target system is input, and the function information of the basic attack scenario including the execution step in which the function information and the attack classification information are associated with each other is used as the function information. Scenario generation means for rewriting the corresponding system configuration information to the device to generate an individual attack scenario that is an individual attack scenario that matches the system configuration information of the training target system,
Training program to function as.

Images