JP5568651B2 - Mail check program, method, and apparatus - Google Patents

Description

  The present invention relates to a mail check program, method and apparatus.

  Furthermore, the invention related to the present invention relates to a mail erroneous transmission preventing apparatus, method and program. More specifically, the present invention relates to a technique for preventing erroneous mail transmission by presenting risk information regarding information leakage to a mail sender and alerting the user before sending the electronic mail.

  As e-mail systems are widely used, the impact of information leakage accidents caused by e-mail is becoming more serious. According to the statistics of the Japan Information Processing Development Corporation (for example, from 2005 to 2007), the number one cause of information leakage accidents is wrong transmission due to mail, fax, and e-mail address errors. Leakage accidents due to erroneous email transmission accounted for about 5.7% of the total number of leaked accidents.

  Information leakage accidents due to email transmission are mostly caused by human errors, such as inadvertently setting the destination (destination address) of outgoing emails. In general, inadvertent mistakes become apparent in response to indications from mistransmitted recipients without the parties being aware. As a result, the occurrence of one accident may have a significant adverse effect. Also, considering the fact that the use of e-mail systems is expanding, the accident rate is expected to increase. Therefore, it is increasingly important to prevent erroneous transmissions caused by human errors such as inadvertent mistakes.

  Conventionally, as a countermeasure for preventing erroneous email transmission, a mechanism has been provided that presents risk information to the sender prior to email transmission to prompt address confirmation and alert. As an example, there is known a system in which a security policy is registered in advance, and when a planned outgoing mail violates the security policy, risk information is presented to the sender to alert the sender.

  In addition, as an example of the conventional method, a white list is prepared in which reliable destinations are registered in advance, and when the destination of outgoing mail is registered in the white list, address check processing is suppressed, thereby reducing security There is a system that reduces the effort of the sender in the check process.

  In another example of the conventional method, an e-mail transmission log is accumulated, and the intimacy between the sender and the intended recipient is stored from the transmission log, and a threshold value for whether or not transmission is possible according to the intimacy. In addition, there is a system that analyzes the wording used in the body of the outgoing mail and judges the result of the analysis based on a threshold value corresponding to the familiarity of the recipient of the outgoing mail.

JP 2007-293635 A JP 2006-059297 A

  In measures to prevent erroneous email transmission, there is a trade-off between the strength of information security check processing and the convenience and operational comfort of the email transmission system. If the security strength of the security check process is high (that is, strictly), the convenience and operation comfort of the system are impaired, and the user feels inconvenience. As a result, the check measures will be lost, and against the intention, there is a risk that the information to be protected cannot be protected. Also, if the security strength is too low, the necessary checks will be insufficient and the security effect will be reduced. Therefore, it is important to balance convenience and information security strength so that the effect of security measures is maximized.

  In addition, the conventional methods of security measures cause the following problems that impair the effectiveness of the measures due to continued operation.

(1) Decrease in freshness of risk information When the same check process is repeated, the user becomes accustomed, and the alerting power of risk information decreases. That is, while the risk information provided to the user is new, the risk information has an alerting power. However, if similar information is repeatedly provided, the user will become accustomed to the information and perform a checked operation without sufficiently checking the contents, so the effectiveness of the countermeasure cannot be maintained.

(2) Inappropriate check strength In general, if the user is overly obliged to comply with the rules, for example, if the check frequency is high or there are many check items, the risk is buried in excessive information. This increases the risk of a simple mistake. At the same time, there is a possibility that the user's operation burden will increase and hinder the original work. Therefore, excessive rule compliance obligations make it difficult to maintain the motivation of users to cooperate in security check measures and increase the risk that the check process will become garbled and formalized.

(3) Whitelist management burden E-mail is often used as a means of communication in the business field because of the convenience and speed of e-mail. In addition, the mobility of human resources is increasing. For this reason, a destination (e-mail address) that the user can trust changes every day. In such a situation, the user or administrator must keep the white list for registering reliable destination addresses as e-mail destinations always up-to-date, and evaluate and manage each address in the white list. Is a heavy burden. In particular, when a transaction or the like is completed, it is normally unnecessary to send an email to such an address, and it is necessary to immediately delete it from the white list. Must be thorough. However, if this deletion work is left to the user, there is a possibility that it will be difficult to comply with the security policy as a whole because each user will have different deletion decisions and deletion times.

(4) The burden of inventory of information assets When risk information is provided using the accumulated transmission history information (transmission log), the risk information is changed along with the resetting of the security policy due to changes in transaction status, organizational changes, etc. It is necessary to delete the stored information that is the premise of provision, so-called inventory of information assets. If this inventory is entrusted to each user, the overall security policy may not be thoroughly enforced, as is the case with whitelist management. Further, although it is possible for the administrator to perform work or force the user to perform the work, there is a problem that the burden on the administrator increases.

  An object of the present invention is to provide an apparatus having an autonomous mail filtering function capable of providing risk information for a mail scheduled to be transmitted while maintaining a security policy and preventing users from getting used to preventing erroneous mail transmission. Is to provide.

The program disclosed in the present application is a program for causing a computer to execute processing for requesting confirmation of whether or not transmission is possible for one or more destinations when sending mail to one or more destinations. For each destination, calculate the storage rate by using an expression indicating a model in which the value is attenuated according to the retroactive transmission order from the latest at the time of processing , and 2) the value obtained by summing up the storage rate for each destination is the weight of the destination 3 ) For a destination where the user weight satisfies a predetermined criterion among the one or more destinations, transmission permission / inhibition simplified compared to a destination not satisfying the predetermined criterion The processing for requesting confirmation is executed by the computer.

  Furthermore, the outline of a typical apparatus disclosed in the present application will be briefly described as follows. That is, when the transmission log is updated by mail transmission, the disclosed apparatus calculates a storage rate indicating how much the mail transmitted in the past is stored for each mail in the transmission log. This calculation is performed using an equation indicating a model in which the value is gradually reduced according to the retroactive order of the mail transmission from the reference at the time of processing. Then, the above-mentioned device generates a user weight list that counts the storage rate of each mail address for each address and records the counted value as the weight of the address.

  Further, when receiving the user's scheduled mail, the above apparatus refers to the stored user weight list, identifies the destination weight of the planned mail, and compares the identified weight with a predetermined threshold. , Decide whether to suppress the provision of information related to destination confirmation.

  When the weight of the destination of the email to be sent exceeds the threshold, the provision of information related to destination confirmation is suppressed. This is because, in the past, mail transmission to the destination has been performed recently, or the number of transmissions to the destination is large, and it can be considered that the probability that the user stores the mail transmission to the destination is high.

  On the other hand, when the weight of the destination of the planned outgoing mail is less than or equal to the threshold value, information related to the destination confirmation is provided to the user, and the planned outgoing mail is transmitted after obtaining a confirmed response. This is because, in the past, it can be considered that a long time has passed since the mail transmission to the destination, or the number of transmissions to the destination is small, and the probability that the user stores the mail transmission to the destination is low.

  Furthermore, since the transmission log is updated when the scheduled transmission mail is transmitted after the destination is confirmed, the above device calculates the storage rate of the previously transmitted mail based on the updated transmission log, and further stores the storage rate. To calculate the destination weight and rewrite the user weight list.

  According to the disclosed apparatus, the probability that the user stores the mail is calculated based on the transmission log that changes from day to day, and the weight of each destination is obtained by counting the calculated storage rates. Therefore, by decreasing the storage rate of the outgoing mail, the weight of the same destination changes for each user weight list generation process, and the determination of whether or not to check the destination also changes. For this reason, since it is possible to confirm the destination of the planned outgoing mail in a state where it is difficult for the user to become accustomed, the effectiveness of the countermeasure for preventing erroneous mail transmission can be maintained in the operation of the information security check.

It is a figure which shows the structural example of the mail erroneous transmission prevention apparatus disclosed in this application. It is a figure which shows the example of a general forgetting curve. It is a figure for demonstrating the weight of a destination. It is a figure which shows the example of a transmission log. It is a figure which shows the example of process control information. It is a figure which shows the example of a user weight list | wrist. It is a figure which shows the example of the reciprocal model used by a user weight list production | generation process. It is a figure which shows the example of the exponent model used by a user weight list production | generation process. It is a figure which shows the example of the linear model used by a user weight list production | generation process. It is a figure which shows the example of a calculation algorithm performed with the user weight list | wrist production | generation part in one Example. It is a figure which shows the example of a destination confirmation screen. It is a figure which shows the relationship between a data adoption period and the transmission log used as a process target in the 1st transmission log acquisition example. It is a figure which shows the relationship between a data adoption period and the transmission log used as a process target in the 2nd transmission log acquisition example. It is a figure which shows the example of data acquired from an external server. It is a processing flowchart which shows the process outline | summary of the mail erroneous transmission prevention apparatus in one Example. It is a processing flow figure of the destination determination process in one Example. It is a processing flow figure of the threshold value setting process in one Example. It is a processing flow figure of user weight list generation processing in one example. It is a figure which shows the verification result by the measurement data about the user weight list | wrist by each model type | formula, and the ratio of the transmission destination address of "trusted".

  FIG. 1 is a diagram illustrating a configuration example of an erroneous mail transmission preventing apparatus according to an embodiment of the present invention.

  The erroneous mail transmission preventing apparatus 1 performs data transmission / reception between the mail transmission server 2 and the sender terminal 3 used by the user as a proxy of the mail transmission server (SMTP server) 2.

  First, the processing operation of the erroneous mail transmission preventing apparatus 1 will be outlined.

  The erroneous mail transmission preventing apparatus 1 stores a transmission log, which is mail transmission history information, every time a mail is transmitted. When the transmission log is updated, for each email sent within a predetermined period in the past, the storage rate indicating the percentage of remembered destinations of the email is determined based on the order of transmission (sending Calculate the storage rate of each mail destination in the transmission log using an equation that shows a model that decreases gradually according to the retroactive date, number of times of transmission, etc., and count the storage rate of each mail destination for each destination. A user weight list is generated as the weight of each destination (user weight list generation processing).

  When receiving the planned email from the sender terminal 3, the erroneous email transmission preventing apparatus 1 refers to the user weight list of the transmission source (user) of the planned email to identify the destination weight of the planned email. Then, the identified weight is compared with a predetermined threshold TH, and if the user weight of the destination of the planned mail to be sent exceeds the threshold TH, the destination is determined as “trusted” (destination determination processing).

  If one of the destinations of the planned email to be sent is not determined to be “trusted”, the erroneous email transmission preventing apparatus 1 performs a process of providing information related to the destination confirmation of the planned email to be sent as a destination confirmation. Thereafter, when a confirmed response is obtained from the user, the planned outgoing mail is transferred to the mail transmission server 2, and if the confirmed response is not obtained, the outgoing mail is sent as an error mail addressed to the sender. 2 (destination confirmation processing).

  In the user weight list generation process described above, the mail storage rate is calculated using an expression representing a model similar to a model generally known as a forgetting curve.

  FIG. 2 is a diagram illustrating an example of a forgetting curve. The forgetting curve is a curve proposed by the psychologist Hermann Ebbinghaus that represents a model of human forgetting of medium- and long-term memory. , It indicates that the frequency decays exponentially with the passage of time (days).

  The erroneous mail transmission preventing apparatus 1 uses a calculation formula representing a model similar to the forgetting curve in FIG. 2, that is, a model in which a value is gradually decreased (decreased) according to the retroactive transmission order from the latest. Calculate the storage rate of each mail destination. According to this calculation process, the storage rate of the most recently sent mail is high, and the storage rate of the mail sent at a time when a considerable amount of time has passed is low.

  Further, the erroneous mail transmission preventing apparatus 1 adds up the storage ratio of mails transmitted in the past for each destination and sets it as the weight of the destination. For example, when the mail transmissions are concentrated at a time close to recently, or when the number of transmissions is large throughout the period, the weight of the destination increases. On the other hand, when the time / number of times has elapsed since the mail transmission or when the number of transmissions is small throughout the period, the weight of the destination becomes small.

  Even in the case of e-mail transmission, if the transmission has been recent, the other party (destination) who sent the e-mail several times is well remembered. Assuming that there is no relationship, it is considered that an address with a higher memory ratio in a certain period has a higher remembered certainty, and that it is reliable to set that address as the destination. .

  FIG. 3 is a diagram for explaining the principle of destination weight. A black circle shown in FIG. 3 represents a mail transmission date addressed to the transmission destination A, and a white square represents a mail transmission date addressed to the transmission destination B. In addition, the vertical axis represents the memory ratio, and the horizontal axis represents the number of days (how many days ago) retroactive from a certain reference (for example, the current day) within a predetermined period.

The curve shown in FIG. 3 is a model representing the memory rate fg (d) that remembers mail d days ago,
fg (d)
dεMA (P) (where MA (P) is a set of mail transmission dates of the destination P)
Then, the weight for destination P is
Weight (P) = Σfg (d)
It becomes.

  In the model shown in FIG. 3, the storage rates of the two mails recently sent to the destination A are calculated as “0.8” and “0.5” and sent to the destination B long before. Assume that the storage rates of the four mails are calculated as “0.21”, “0.21”, “0.21”, and “0.21”.

The weights of destination A and destination B are respectively
Weight (A) = 1.3,
Weight (B) = 0.84
It becomes.

  Next, in the address determination process, the erroneous mail transmission preventing apparatus 1 identifies the weight of the transmission destination address of the planned mail to be transmitted from the user weight list, and compares the weight with a threshold value TH (= 1.0).

  Since the transmission destination “A” of the scheduled transmission mail has a weight (1.3) exceeding the threshold value TH, it is determined that the transmission destination A is “trusted”. Further, since the transmission destination “B” has a weight (0.84) equal to or less than the threshold value TH, it is determined that the transmission destination B is not “trusted”.

  In the destination confirmation process, if there is a destination address that is determined not to be “trusted”, a destination confirmation screen for the planned outgoing mail is displayed to prompt the user to confirm the destination.

  The weight of the transmission destination calculated by the erroneous mail transmission preventing apparatus 1 is obtained based on the storage rate that gradually decreases according to the transmission order (date / time) of the transmission logs accumulated every day. For this reason, even if the weight of the destination address exceeds the threshold value TH and is determined to be “trusted” at the present time, it may not be determined as “trusted” without exceeding the threshold value TH in future processing.

  As described above, even if the mail is sent to the same destination address, the determination result of whether or not to check the destination changes for each transmission operation. As a result, the user is less likely to become accustomed to address confirmation, and the effectiveness of the security check process can be maintained.

  Next, each processing means of the erroneous mail transmission preventing apparatus 1 will be described.

  As shown in FIG. 1, the erroneous mail transmission preventing apparatus 1 includes a mail transmission / reception unit 10, a transmission log storage unit 11, a user weight list storage unit 12, a process control information storage unit 13, a risk degree setting information storage unit 14, a transmission log. An acquisition unit 15, a user weight list generation unit 16, a destination determination unit 17, a destination confirmation unit 18, and a processing control information setting unit 19 are provided.

  The mail transmission / reception unit 10 acquires the planned transmission mail transmitted from the sender terminal 3, holds the acquired transmission planned mail in an internal mail storage unit (not shown), and accepts a transmission request from the destination confirmation unit 18. Then, the stored transmission schedule mail is transferred to the mail transmission server 2. The mail transmission / reception unit 10 includes a mail transmission processing unit and a mail reception processing unit.

  The transmission log storage unit 11 stores a transmission log including a transmission source, a transmission destination, and a transmission date and time of the mail transferred to the mail transmission server 2.

  FIG. 4 is a diagram illustrating an example of a transmission log. The transmission log includes information on a transmission source address indicating a mail sending source, a transmission destination address indicating a mail destination, and a transmission date and time indicating transmission processing date and time. The transmission destination address is also recorded in the transmission destination address.

  The process control information storage unit 13 stores process control information for controlling the processes of the user weight list generation unit 16 and the destination determination unit 17.

  FIG. 5 is a diagram illustrating an example of processing control information. The process control information includes parameters for the number of target logs (WIN), the minimum storage rate (FG), the threshold initial value (TH0), the start of data adoption period (From), and the end of data adoption period (To).

  The target log number (WIN) is the maximum number of transmission logs to be processed by the user weight list generation unit 16. For example, it indicates how many previous e-mails (transmission logs) from the reference time (now) are used as processing targets.

  The minimum storage rate (FG) is the minimum value of the storage rate (fg), and is the storage rate of the mail transmitted at the earliest (oldest) in the transmission log to be processed.

  The threshold initial value (TH0) is an initial value of the threshold TH of the determination process of the destination determination unit 17.

  The start (From) and end (To) of the data adoption period indicate the start and end of a period indicating the transmission date and time of the transmission log to be processed by the user weight list generation unit 16. By setting the data adoption period, transmission logs that are not suitable for the judgment target can be eliminated.

  The user weight list storage unit 12 stores a user weight list generated by the user weight list generation unit 16 for each sender (source address).

  FIG. 6 is a diagram illustrating an example of a user weight list. The user weight list includes the destination address to be processed, the weight calculated for the destination address, the number of past transmissions indicating the total number of past transmissions, the registration date and time indicating the registration date and time in the user weight list, the latest Including the last transmission date and time indicating the date and time of transmission.

  As shown in FIG. 6, when the transmission destination addresses in the user weight list are sorted in descending order of weight, when the destination determination unit 17 performs the determination process with the threshold TH = 3.6, the weight is set to the threshold TH. The upper four destination addresses that exceed are determined to be “trusted”, and the destination confirmation process for the planned outgoing mail is suppressed.

  The risk degree setting information storage unit 14 stores risk degree setting information for controlling a risk degree calculation process related to information leakage.

  The risk level setting information includes the domain type (domain name classification) of the destination address, the format and presence of the attached file, whether the email title contains a specific expression, or a specific expression in the email body. Is defined in advance and the degree of risk of mail that satisfies the condition. The risk degree is a numerical value of 1.0 or more, and the risk value is set to be higher as the value is larger. For example, as the condition setting based on the classification of the domain name, the risk level is set depending on whether it is a competitor (companies in the same industry), mobile phone mail, or mailing list.

  The address determination unit 17 calculates the risk level (RF) of the planned outgoing mail based on the risk level setting information.

  The transmission log acquisition unit 15 stores, in the transmission log storage unit 11, the transmission log of the planned transmission mail transferred to the mail transmission server 2 by the mail transmission / reception unit 10.

  The user weight list generation unit 16 obtains the process control information from the process control information storage unit 13 and the transmission log from the transmission log storage unit 11 when a predetermined period or opportunity, for example, the completion of the transmission log update process, occurs. , The storage rate of each mail address in the transmission log is calculated using a predetermined formula, and the sum of the storage rates of each source address is calculated as the weight of the transmission address. Then, a user weight list for recording the weight of each destination address is created and stored in the user weight list storage unit 12.

  FIG. 7 is a diagram illustrating an example of a model representing a calculation formula used in the process of the user weight list generation unit 16. The model in FIG. 7 is a reciprocal model, and is based on an equation for obtaining the storage rate at the elapsed time d, that is, the probability f (d) of remembering mail d days ago by the reciprocal.

  The model equation of FIG. 7 can be expressed as follows.

f (d) = (1-F) (W / d-1) / (W-1) + F
d: elapsed time,
W (WIN): upper limit of elapsed time,
F (FG): minimum memory rate,
The elapsed time d indicates how many days before the reference date the transmission date of the mail to be processed is, for example, in units of days. For example, d takes an integer value from 1 to W.

  The upper limit W of the elapsed time indicates the upper limit of the number of processes of the elapsed time d (mail transmission date) as a processing unit. That is, W is a parameter that indicates the upper limit of the number of transmission days retroactive from the reference date, which is the maximum value of elapsed time. For example, W = 50.

  The minimum memory rate F is a parameter indicating a probability (memory rate) of storing a mail transmitted on the transmission date that is the upper limit W of the elapsed time, that is, the most recent transmission date (for example, a transmission date that goes back 50 days). is there. For example, F = 0.2.

  In the present embodiment, the user weight list generation unit 16 executes a destination address weight calculation algorithm using the above model formula.

  The user weight list generation unit 16 is not limited to the model formula shown in FIG. 7, and may use another known model formula, for example.

  8 and 9 are diagrams illustrating examples of models representing calculation formulas used in the processing of the user weight list generation unit 16.

  The model in FIG. 8 is an exponential model proposed by the original forgetting curve, and the probability of storage by the elapsed time t, for example, the probability f (t) of remembering the mail t days ago when the number of days is the unit. , By an expression obtained by an exponential function.

  The model equation of FIG. 8 can be expressed as follows.

f (t) = e ^ (-At)
However, A = -In (F) / W
t: elapsed time (days),
W (WIN): upper limit of elapsed time,
F (FG): minimum memory rate,
The elapsed time t indicates, for example, how many days before the transmission date of the mail to be processed when the number of days is used as a unit. For example, t takes an integer value from 1 to W. The upper limit W of the elapsed time is the same as that shown in FIG. However, when the equation of the model in FIG. 8 is applied to the storage rate of the destination of the transmitted mail, the attenuation of the storage rate is somewhat moderate. For example, there is a risk that an inappropriate trusted determination may be performed such as processing including an erroneously transmitted destination, or reducing the weight of a destination address that has been transmitted several times in the past.

  The model in FIG. 9 is a linear model, which is based on an equation for obtaining a probability of storage by elapsed time d, for example, a probability f (d) of remembering mail d days ago by a linear function when the number of days is used as a unit. It is.

  The model equation of FIG. 9 can be expressed as follows.

f (d) = (F-1) / (W-1) d + (WF) / (W-1)
d: elapsed time,
W (WIN): upper limit of elapsed time,
F (FG): minimum memory rate,
The elapsed time d, the upper limit W of the elapsed time, and the minimum storage rate F are the same as those shown in FIG. However, even in the model equation of FIG. 9, the attenuation of the memory rate becomes more gradual, and there is a risk of making a wrong trusted determination.

  FIG. 10 is a diagram illustrating an example of a calculation algorithm executed by the user weight list generation unit 16 in one embodiment.

  In the processing according to the source code example of the calculation algorithm shown in FIG. 10, the calculation by the reciprocal model equation shown in FIG. 7 is performed using the following parameters.

WIN: number of processing targets,
FG: Minimum storage rate TH: Destination determination threshold Here, the number of mail transmissions i is used as the value of WIN, and i is a natural number up to WIN. The storage rate f (i) is a probability of storing a mail related to transmission at the transmission count i (transmission executed i times before a certain standard transmission process). The storage rate f (1) of emails sent with the number of transmissions i = 1 (most recent) is “1”, and the rate of storage of emails sent with the number of transmissions i = WIN (the number of transmissions that goes back most of the data used). f (WIN) is “0.2” which is the minimum storage rate fg.

  The weight (wt (addr)) of the transmission destination address (addr) is the sum of the storage rates f (i) of the mails transmitted to the transmission destination address (addr).

  When the transmission scheduled mail is stored in the mail storage unit of the mail transmission / reception unit 10, the destination determination unit 17 refers to the user weight list corresponding to the transmission source address of the transmission planned mail and sets the transmission destination address of the transmission planned mail. Identify weights. Then, the identified weight is compared with the threshold value TH. When the weight exceeds the threshold value TH, the destination address is determined as “trusted”. When the weight is equal to or less than the threshold value TH, the destination address is changed to “ It is determined that it is not “trusted”.

  The destination determination unit 17 calculates a risk level RF corresponding to a condition to which the scheduled transmission mail is applied when a condition for determining the risk level is stored in the risk level setting information storage unit 14, and based on the calculated risk level Then, the threshold TH0 may be increased or decreased, and the changed threshold TH may be used to compare with the destination weight. The risk level RF is a numerical value of 1.0 or more, and depending on the risk level, the risk level is high (= 2.0), medium risk level (= 1.5), and low risk level (= 1.0). To get.

Then, the destination determination unit 17
TH = TH0 * RF
As described above, the threshold value TH is changed.
For example, it is assumed that the following conditions are defined in the risk degree setting information storage unit 14.

“Risk of 1.0 when no file is attached to a destination mail other than foo.co.jp”,
“Risk of 1.5 when a file is attached to a destination mail other than foo.co.jp”
Further, it is assumed that the threshold value TH0 is “3.6” and the transmission destination address of the planned transmission mail is “ddd@ddd.ddd”.

  Since the degree of risk is low (1.0) when no file is attached to the planned mail to be sent, referring to the user weight list in FIG. 6, the weight of the destination address (3.9993) is the threshold TH = It becomes larger than 3.6 * 1.0 = 3.6, and it is determined as “trusted”.

  However, since the risk is moderate (1.5) when a file is attached to the email to be sent, the weight of the destination address is less than the threshold TH = 3.6 * 1.5 = 5.4. , “Trusted” is not determined.

  In this way, even if the transmission destination address is the same, the security level can be strengthened by changing the “trusted” determination condition according to the state of the planned mail to be transmitted.

  The destination confirmation unit 18 displays a destination confirmation screen for providing risk information including destination confirmation on the sender terminal 3 when the transmission destination address of the planned outgoing mail is not determined to be “trusted”. When the destination confirmation unit 18 receives confirmation from the sender terminal 3 regarding the displayed destination, the destination confirmation unit 18 requests the mail transmission / reception unit 10 to transfer the scheduled transmission mail.

  The generation and output processing of the destination confirmation screen of the destination confirmation unit 18 is realized using existing destination confirmation processing.

  The processing of the destination confirmation unit 18 is suppressed when it is determined that all the transmission destination addresses of the planned transmission mail are “trusted”.

  FIG. 11 is a diagram showing an example of a destination confirmation screen provided to the user.

  The address confirmation screen 100 in FIG. 11 is a screen that prompts confirmation of the address of the planned outgoing mail.

  In the destination confirmation screen 100, a destination address whose weight is greater than or equal to the threshold TH is displayed with a check mark 101 indicating that the destination has been confirmed in advance in the check box for inputting confirmed, and a destination address whose weight is smaller than the threshold TH. The check box field is displayed as a blank field 103. The destination confirmation screen 100 of FIG. 11 shows a state in which a check mark is input by a user operation in the blank box 103 of the check box of the destination address whose weight is smaller than the threshold value TH.

  In the address confirmation screen 100, based on the processing result of the risk degree calculation process, when it is determined whether the scheduled mail is a mail addressed to another company in the same industry, a mobile mail, or a mailing list, You may make it display the emphasis display and warning message of applicable transmission planned mail.

  The user confirms all destination addresses on the destination confirmation screen 100, inputs a check mark in the check box field 103, and selects the button 105 for instructing a transmission request, whereby the sender terminal 3 Is transmitted to the erroneous mail transmission preventing apparatus 1 from the terminal. When the button 107 for canceling the transmission on the destination confirmation screen 100 is selected, a transmission cancellation request is made to the erroneous mail transmission preventing apparatus 1, and the planned transmission mail held by the mail transmission / reception unit 10 becomes an error mail for the sender. Converted.

  It should be noted that if the button 105 is not selected even after 20 seconds have elapsed since the check mark is entered in all the check box fields 103 on the address confirmation screen 100 in FIG. 11, there is a user transmission request. As a result, a confirmation message is transmitted from the sender terminal 3 to the erroneous mail transmission preventing apparatus 1.

  When the destination confirmation unit 18 is notified of confirmation from the sender terminal 3, the destination confirmation unit 18 requests the mail transmission / reception unit 10 to transfer. The mail transmission / reception unit 10 transfers the planned mail to the mail transmission server 2. The transmission log acquisition unit 15 stores in the transmission log storage unit 11 the transmission source address, transmission destination address, and date and time of transmission processing of the forwarded transmission planned mail. When transmission cancellation is notified from the sender terminal 3, the mail transmission / reception unit 10 is requested to transfer an error mail. The mail transmission / reception unit 10 transfers the error mail whose destination is converted to the sender's address to the mail transmission server 2. In this case, the transmission log acquisition unit 15 does not accumulate transmission logs.

  The process control information setting unit 19 sets or updates the process control information stored in the process control information storage unit 13. The processing control information setting unit 19 cooperates with a management server (management server / HR management server) 4 such as a management server that manages an information security policy related to the mail transmission server 2 or a user's HR management system. Automatically acquires information related to information security policy changes and information related to organizational changes, and sets and updates model formula parameters for process control information.

  Further, the process control information setting unit 19 can set or update the process control information by periodically obtaining an input operation of the administrator or information of the management server 4.

  The data adoption period of the process control information is a condition for extracting the transmission log to be processed by the user weight list generation unit 16 according to the transmission date and time. Therefore, the transmission log to be processed varies depending on the data adoption period setting.

  The user weight list generation unit 16 acquires a transmission log to be processed by any of the following processes.

1) First, transmission logs up to the upper limit number given by WIN from the reference are extracted (i or d is a natural number up to WIN), and then the transmission date and time is within the data adoption period from the extracted transmission logs. Example of processing that uses only transmission log (first acquisition example),
2) First, extract a transmission log whose transmission date and time is within the data adoption period, and then use the extracted transmission log in order of transmission date and time up to the upper limit number given by WIN from the reference (second example) Acquisition example).

  FIG. 12 is a diagram for explaining the relationship between the data adoption period and the transmission log to be processed in the case of the first acquisition example.

  In FIG. 12, the arrow from left to right represents the passage of time. WIN (= 100) is the number of logs to be processed on a daily basis, and now indicates the reference time, and is the current date.

  In case 1 of FIG. 12, a data adoption period is set from the start date (From) to the end date (To) after the current date. This indicates that records within the future data adoption period are to be processed. First, since records in the transmission log are extracted with the transmission date and time corresponding to the transmission date from the current date to 100 days ago, the record to be processed in the processing of the current date is zero.

  In case 2 of FIG. 12, a data adoption period is set in which the start is before the current date and the end is after the current date. Records to be processed in processing on the current date are records from the current date to the first transmission date of the data adoption range.

  In case 3 of FIG. 12, a data adoption period in which the start and end are before the current date and time is set. Since the start and end are within the range of 100 days before the current date, the records to be processed in the processing on the current date are records of the transmission date and time from the start to the end of the data adoption period.

  In case 4 of FIG. 12, a period is set in which the start period is before 100 days ago. Accordingly, the record to be processed in the processing on the current date is a record of the transmission date and time from 100 days before to the end of the data adoption period.

  In case 5 of FIG. 12, a data adoption period is set such that the start and end are before 100 days ago. Records to be processed by processing on the current day are zero.

  In case 6 of FIG. 12, a data adoption period is set in which the start period is before 100 days ago and the end period is after the current day. Therefore, the record to be processed in the processing on the current date is a record of the transmission date and time from 100 days ago to the current date.

  FIG. 13 is a diagram for explaining the relationship between the data adoption period and the transmission log to be processed in the case of the second acquisition example.

  In cases 4 and 5 shown in FIG. 13, a data adoption period in which the start and end are before the current date is set. In this case, the records to be processed in the processing on the current day are the transmissions of the number of days going back to 100 days from the end of the records that were sent during the beginning and end of the data adoption period. It becomes a log.

  In the cases 1, 2, 3, and 6 of FIG. 13, data is extracted in the same manner as the cases 4 and 5 of FIG. 13, but the same results as the case to which the same numbers shown in FIG. Become.

  As shown in FIGS. 12 and 13, in the first acquisition example, the parameter WIN has a meaning of a value for setting the upper limit of the process. In the second acquisition example, the parameter WIN has a meaning of a value for setting the processing accuracy. In the present embodiment, the user weight list generation unit 16 adopts the first acquisition example described above, and performs processing according to each parameter of the processing control information set so as to have the relationship shown in case 6 of FIG. .

  Furthermore, the data adoption period of the processing control information is updated based on information transmitted from the management server 4.

  The process control information setting unit 19 acquires information such as a change of the user's affiliation and management information of the affiliation project from the management server 4 and sets a data adoption period of the process control information.

  When a user changes affiliation from the Y division to the X division due to personnel changes, the destination used in the business of the Y division to which he previously belonged is usually unnecessary in the moved X division Because it can be handled.

  FIG. 14 is a diagram for explaining the update of the data adoption period due to personnel changes.

  FIG. 14A shows an example of personnel change information that is automatically collected periodically from the management server 4 by the erroneous mail transmission preventing apparatus 1.

  Personnel change information includes a change date indicating a date on which the change is executed, a target person ID for identifying a target person to be transferred, and a change destination indicating a new affiliation. In this example, the header part (mail ID) of the user's mail address is used as the target person ID.

  The process control information setting unit 19 acquires personnel change information shown in FIG. 14A from the external management server 4 periodically (for example, every day at 9:00 am). Of the records including the user ID of the user himself / herself, the record with the latest date and time is acquired. Then, when the change date and time of the acquired record is earlier than the current date, the start time (From) of the learning data period of the process control information stored in the process control information storage unit 13 is updated to the change date and time.

  The process control information setting unit 19 does not update the process control information when there is no record with the latest date and time.

  For example, when the personnel change information in FIG. 14A is acquired, it is assumed that the process control information setting unit 19 obtains a record of the change date December 1, 2008 for the user aaa. The process control information setting unit 19 updates the start date (April 1, 2008) and end date (April 1, 2009) of the data adoption period before the update with the transfer date that acquired the start date. As a result, the data adoption period after the update is from December 1, 2008 to April 1, 2009.

  The sender terminal 3 includes a mail transmission / reception unit (mailer) 31 and a confirmation display input unit 32.

  The mailer 31 performs a known mail transmission / reception process, and transmits the mail that has been subjected to the transmission process to the erroneous mail transmission preventing apparatus 1. The confirmation display input unit 32 performs display processing of known XML (Extensible Markup Language) or HTML (Hyper Text Markup Language) data, and displays the destination confirmation screen 100 transmitted from the erroneous mail transmission prevention device 1 as the sender terminal 3. Displayed on a display unit (not shown), and a destination confirmation inputted by a user operation via an input device (not shown) is transmitted to the erroneous mail transmission preventing apparatus 1. FIGS. 15-18 is a figure which shows the processing flow of the mail erroneous transmission prevention apparatus 1 in one Example.

  FIG. 15 is a process flow diagram showing the flow of the entire process of the erroneous mail transmission preventing apparatus 1.

  When the erroneous mail transmission preventing apparatus 1 starts processing, the address determination unit 17 reads processing control information and a user weight list (step S1). The mail transmitting / receiving unit 10 receives the scheduled mail from the mailer 3 and holds it in the mail storage unit (step S2).

  Then, the destination determination unit 17 executes a destination determination process for the scheduled transmission mail held in the mail storage unit (step S3). Details of the destination determination process will be described later with reference to FIGS.

  In the destination determination processing in step S3, when the destination determination unit 17 determines that all the transmission destination addresses of the planned outgoing mail are “trusted” (Y in step S4), the destination confirmation unit 18 Destination confirmation processing for the scheduled mail is suppressed (step S5).

  On the other hand, when it is determined in the destination determination process in step S3 that even one transmission destination address of the planned outgoing email is not “trusted” (N in step S4), the destination confirmation unit 18 sends the destination of the planned outgoing email. As confirmation processing, a destination confirmation screen is output (step S6).

  After that, when the destination confirmation screen is displayed on the client side and the user confirms the destination address that is not “trusted” and the mailer 3 notifies “confirmed”, the scheduled email can be sent. (Y in step S7), the mail transmitting / receiving unit 10 transfers the scheduled mail to the mail transmission server 2 (step S8). The transmission log acquisition unit 15 acquires the transmission log of the forwarded transmission planned mail and stores it in the transmission log storage unit 11 (step S9).

  Thereafter, the user weight list generation unit 16 executes a user weight list generation process (step S10), and the erroneous mail transmission preventing apparatus 1 ends the process. Details of the user weight list generation processing will be described later with reference to FIG.

  Next, the destination determination process in step S3 will be described.

  FIG. 16 is a process flow diagram of the destination determination process.

  The destination determination unit 17 executes a threshold setting process (step S30). Details of the threshold setting process will be described later with reference to FIG.

  The destination determination unit 17 acquires the weights of all the transmission destination addresses of the planned transmission mail from the user weight list, and compares each of the acquired weights with the threshold value TH (step S31). If the weight of the transmission destination address exceeds the threshold value TH (Y in step S31), the destination determination unit 17 determines that the transmission destination address is “trusted” (step S32). When the weight of the transmission destination address is equal to or less than the threshold value TH (N in Step S31), it is determined that the transmission destination address is not “trusted” (Step S33). The determination results of steps S32 and S33 are passed to the destination confirmation unit 18.

  FIG. 17 is a process flowchart of the threshold setting process in step S30.

  The address determination unit 17 calculates the risk level (RF) of the planned outgoing mail based on the risk level setting information in the risk level setting information storage unit 14 (step S301), and sets the threshold TH0 based on the calculated risk level. The threshold value TH is obtained by changing (step S302).

  Next, the user weight list generation process in step S10 will be described.

  FIG. 18 is a process flow diagram of the user weight list generation process.

  The user weight list generation unit 16 initializes the weights of the user weight list stored in the user weight list storage unit 12. Further, based on the process control information in the process control information storage unit 13, an expression f () for calculating the storage rate is generated (step S101).

  Next, the user weight list generating unit 16 repeats the processing of the following steps S103 to S105 for each record r of the transmission log in the transmission log storage unit 11 (loop loop1: steps S102 to S106).

  In the loop loop 1, the user weight list generation unit 16 determines whether the record r is within the target log number (WIN) from the latest reference (now) (step S103), and if the record r is within WIN. If this is the case (Y in step S103), it is further determined whether the transmission date of the record r is within the data adoption period stored in the processing control information storage unit 13 (step S104). If the transmission date of the record r is within the data adoption period (Y in step S104), the record r is added to the target log that is a record group used in the processing (step S105).

  Next, the user weight list generation unit 16 sorts the records r of the target log in the order of transmission date and time (step S107), and processes the following steps S109 to S113 for each record s of the target log arranged in the new order. Is repeated (loop loop 2: steps S108 to S114).

  In loop loop 2, the user weight list generation unit 16 calculates the storage rate f (s) of the record s based on the storage rate calculation formula f () generated in the process of step S 101 (step S 109). A transmission destination (send) is taken out (step S110). And the process of step S112 is repeated about each transmission destination address (p) of the taken out transmission destination (loop loop3: steps S111 to S113).

  In the loop loop 2, the user weight list generation unit 16 adds the calculated storage rate f (s) to the weight of the corresponding transmission destination address (p) and updates the user weight list (step S112).

  Due to the difference in the models employed by the user weight list generation unit 16 (see FIGS. 7 to 9), the ratio of transmission destination addresses determined as “trusted” by the destination determination unit 17 changes.

  FIG. 19 is a diagram illustrating a result of verifying the relationship between the calculation formula indicating each model of the user weight list generation process and the ratio of the “trusted” transmission destination address in the destination determination process using actually measured data.

  In the verification result of FIG. 19, the ratio of the “trusted” destination address in the destination determination process by the user weight list generated by the expression of the reciprocal model (see FIG. 7) is 23.361%, the exponent model (FIG. 8) is 33.722%, the linear model (see FIG. 9) is 35.356%, and the non-attenuated (flat) model is 40.456%. there were.

  In the verification result, the proportion of destination addresses determined as “trusted” decreases in the order of the reciprocal model <exponential model <linear model <non-attenuation model. In other words, the steeper slope of the model formula used in the user weight list generation process, the smaller the evaluation of the destination weight and the greater the strength of information security. As shown in the figure, it should be noted that the gradual slope should not be excessively gradual. For example, if you use a model with a slow slope of memory rate attenuation, the memory rate of the wrongly sent destination is treated as an influential value, and the weight of the destination that has been sent several times in the past is difficult to reduce. It can be seen that there is a risk that the trusted judgment is likely to contain errors.

As shown in the above embodiment, the following effects can be obtained when the erroneous mail transmission preventing apparatus 1 is applied to the mail transmission system. That is,
(1) The erroneous mail transmission preventing apparatus 1 updates the transmission log by the user's mail transmission, and also updates the user weight list referred to in the destination determination process. As a result, while maintaining the security level, excessive user confirmation operations are reduced, and repeated confirmation of the same content is reduced to make it difficult for the user to become accustomed. Therefore, it is possible to suppress a decrease in the sustainability of the security check effect caused by user's familiarity.

  (2) The erroneous mail transmission preventing apparatus 1 automatically updates the user weight list using the transmission log. Thereby, it is possible to reduce the burden on the administrator and the user regarding the maintenance management of the list of reliable destination addresses.

  (3) The erroneous mail transmission preventing apparatus 1 controls the removal of data (transmission log) used for the user weight list generation process based on the setting of the process control information. This eliminates the need for work related to information inventory, such as reviewing the destination address of each user and checking work by the administrator, thereby reducing the management burden.

  In addition, when the security policy is changed by the processing control information setting unit 19 periodically acquiring information of the management server 4, the processing control information is automatically set in conjunction with the information change of the management server 4. The setting can be changed, and it is possible to respond in a unified manner without the trouble of the user or administrator. In particular, when there is a business change or organizational change, the business period before the status change is acquired from the management server, and the data adoption period setting is automatically changed so that the transmission log for that business period is not subject to processing. By doing so, it is possible to change (decrease) the weight of the destination address trusted in the business before the change. As a result, it is possible to reduce the error of erroneously sending new business information to a destination that was frequently sent in the previous business. As a result, thorough adherence to the security policy can be improved and the efficiency of operation management by the administrator can be improved.

  In the above description, the case where the invention mainly made by the present inventor is applied to the mail erroneous transmission prevention processing has been described. However, the present invention is not limited to this application example, and the scope of the gist of the description. Of course, various modifications are possible.

  A mail erroneous transmission preventing apparatus 1 disclosed as an embodiment of the present invention is a program read and executed by a computer, and a mail transmission / reception unit 10, a transmission log acquisition unit 15, and a user weight list generation included in the mail erroneous transmission preventing apparatus 1. Each processing unit of the unit 16, the destination determination unit 17, the destination confirmation unit 18, and the processing control information setting unit 19 can be constructed as a program module. This program can be stored in an appropriate recording medium such as a portable medium memory, a semiconductor memory, or a hard disk, which can be read by a computer, and is provided by being recorded on these recording media. Alternatively, this program is provided by transmission and reception using various communication networks via a communication interface.

  The erroneous mail transmission preventing apparatus 1 disclosed as the above embodiment has been described as an apparatus that is provided outside the sender terminal 3 and transmits / receives data to / from the sender terminal 3 via a network (for example, LAN). However, the erroneous mail transmission preventing apparatus 1 may be constructed in the sender terminal 3. The erroneous mail transmission preventing apparatus 1 constructed in the sender terminal 3 operates in the same manner as the processing operation in the above embodiment, but the destination confirmation screen 100 generated by the destination confirmation unit 18 is a binary serialized in the main memory. The data is transferred to the confirmation display input unit 32 as data.

DESCRIPTION OF SYMBOLS 1 Mail erroneous transmission prevention apparatus 10 Mail transmission / reception part 11 Transmission log memory | storage part 12 User weight list memory | storage part 13 Process control information memory | storage part 14 Risk degree setting information memory | storage part 15 Transmission log acquisition part 16 User weight list production | generation part 17 Destination determination part 18 Destination confirmation unit 19 Processing control information setting unit 2 Mail transmission server 3 Sender terminal 31 Mail transmission / reception unit (mailer)
32 Confirmation display input section 4 Management server / HR management server

Claims (4)

A program for causing a computer to execute processing for requesting confirmation of whether or not transmission is possible for one or more destinations when sending mail to one or more destinations,
For each of the one or more destinations, the storage rate is calculated using an equation indicating a model in which the value is attenuated according to the retroactive transmission order from the latest during processing ,
Creating a user weight list for recording the storage rate for each destination as a weight of the destination;
Of the one or more destinations, for a destination where the user weight satisfies a predetermined criterion, the computer is caused to execute a process for requesting confirmation of whether transmission is simpler than a destination that does not satisfy the predetermined criterion. A mail check program characterized by Retroactive to a period of transmission order from the last time the processing, the mail check program according to claim 1, characterized in that is determined based on personnel change history of the user is a target that requests the confirmation. A method in which when a mail is transmitted to one or more destinations, a computer executes a process of requesting confirmation of whether or not transmission is possible for the one or more destinations.
For each of the one or more destinations, the storage rate is calculated using an equation indicating a model in which the value is attenuated according to the retroactive transmission order from the latest during processing ,
Creating a user weight list for recording the storage rate for each destination as a weight of the destination;
Among the one or more destinations, the computer executes a process for requesting confirmation of whether transmission is simpler for destinations where the user weight satisfies a predetermined criterion than for destinations not satisfying the predetermined criterion. A mail check method characterized by An apparatus for executing a process for requesting confirmation of whether or not transmission is possible for one or more destinations when sending mail to one or more destinations,
For each of the one or more destinations, a storage rate is calculated using an expression indicating a model in which the value is attenuated according to the retroactive transmission order from the latest at the time of processing, and a value obtained by totaling the storage rate for each destination is calculated. A first processing unit for creating a user weight list to be recorded as the weight of
Among the one or more destinations, a second processing unit that requests a confirmation of transmission availability more simplified than a destination that does not satisfy the predetermined criterion for a destination whose user weight satisfies the predetermined criterion; A mail check device characterized by comprising.

Images