KR101273112B1 - E-mail transmission control system using transmission rule reasoning result - Google Patents

Abstract

PURPOSE: An e-mail sending control system is provided to improve management convenience by avoiding sending restriction based on spam mail classification. CONSTITUTION: A list distribution unit divides a sending list in each domain. A domain control unit confirms a mail sending result based on a sending control rule by managing sending list information in each domain. The domain control unit is generated in each domain. A sending result is provided to the domain control unit and an estimation engine. A domain sending control management unit controls the order of the corresponding domain control unit. [Reference numerals] (AA) Start; (BB) Confirm a sending state; (CC,GG,II,KK,NN,QQ) Yes; (DD) Are spam mails classified?; (EE,HH,,OO,RR) No; (FF) Receiver list > 0 ?; (JJ) Current connected number >= Approved connecting number ?; (LL) Current sent number >= Approved sending number per minute ?; (MM) Session termination time = null ?; (PP) Store a session termination time; (SS) Current time - termination time > 60s; (TT) End

Description

E-MAIL TRANSMISSION CONTROL SYSTEM USING TRANSMISSION RULE REASONING RESULT}

The present invention relates to an e-mail sending control system using a sending control rule inference result, and in particular, determines a sending control rule for each receiving domain based on the inferred sending control rule, and manages sending according to the sending control rule. The present invention relates to an e-mail sending control system using the result of sending control rule inference.

Marketing using email is rapidly increasing due to the daily use of various communication means, and various research and policy restrictions for filtering such spam mails are increasing as the damage caused by advertising mails, ie, spam mails, is transmitted unilaterally regardless of the recipient's intention. This is being done.

Various e-mail service providers have adopted these spam filtering techniques and policies to block mails that are judged to be spam, thereby increasing subscriber satisfaction and reducing unnecessary loads. It's hard to tell if it's spam or if it's legitimate.

Therefore, even if a normal mail is sent with the recipient's consent, the mail may be blocked by the spam mail blocking policy. For example, regular e-mails or coupon e-mails sent to members who have agreed to receive e-mails may be blocked and the recipients may not receive them, and work related newsletters or academic newsletters may be blocked. It can also happen if you do not receive.

However, large portals and professional hosting companies that provide email services have a variety of anti-spam rules and policies applied to them, and not only to block them, but to block them as well. Due to the varying rules and policies, more and more e-mails fail to deliver in the event of normal bulk e-mail delivery. This failure to send a normal e-mail is becoming increasingly problematic in that it limits the right of the recipient.

Korean Patent Laid-Open Publication No. 10-2002-0029526, 'Methods and Systems for High-Speed Bulk Email Delivery,' distinguishes and indexes broadcast mails and individual mails for mass e-mail sending, and then sends a plurality of e-mail sending agents. A technique for sending a large amount of e-mail at a high speed by activating in parallel to perform an e-mail sending operation is disclosed.

In addition, Korean Patent Application Publication No. 10-2003-0017702, 'Mail Processing System' also discloses a distributed processing method for sending a large amount of e-mail, while increasing the load and processing speed of authenticating a user for sending such a large amount of mail. In order to solve the problem, a technique of simplifying user authentication and retrieval by applying a hash algorithm is disclosed to reduce the load of mail by reducing the authentication load.

Until the spam policy was strengthened, more attention was paid to the speed of sending than the success rate. Therefore, distributed mail processing to increase the number of hourly dispatches was focused. However, this method is no longer realistic due to the low success rate of the mail. There is no way.

Recently, research has been started to improve the success rate of mail delivery rather than high speed mail delivery or rapid mail delivery. However, no suitable solution has been proposed. For example, you can send a large amount of email by setting a fixed sending rule that takes into account spam blocking rules, and collect and resend only the failed mail. However, if the mail sending fails due to the spam blocking policy, the success rate is lowered. Even if the message is sent again, the mail sending will fail again due to the spam mail blocking policy.

As a result, in order to increase the success rate of sending mail, it is necessary to apply a very low level of sending rule to prevent spam blocking, or to request unblocking if it is blocked by spam mail. There is a problem that the delivery speed is low or the management cost is high.

Therefore, there is a need for a new mass mailing system and method capable of sending a mail to a recipient while automatically avoiding a spam mail blocking policy at an optimal level for a normal mass mail.

Korean Patent Publication No. 10-2002-0029526 Korean Patent Publication No. 10-2003-0017702

An object of the embodiments of the present invention for improving the above-mentioned problems is to determine a shipment control rule for each receiving domain based on the inferred shipment control rule, and to manage the shipment according to the shipment control rule to increase the success rate of the shipment. It is to provide an email sending control system using the result of sending control rule inference.

Another object of the embodiments of the present invention is to manage domain controllers that are dynamically generated according to a list for each domain, so that each domain controller distributes and manages whether or not it is sent based on dispatch control rule inference information for the corresponding domain. It is to provide an e-mail sending control system using the result of sending control rule inference to improve load balancing and management convenience by distributed management.

Another object of the embodiments of the present invention is to send the mail according to the classification of the spam mail so that the domain controller can perform the delay management based on the allowable value according to the dispatch control rule inference information even if the domain controller checks the current sending status and is not classified as the spam mail. It is to provide an e-mail sending control system using the result of sending control rule inference to avoid the restriction.

In order to achieve the above object, an email sending control system using a sending control rule reasoning result according to an embodiment of the present invention includes a sending control rule reasoning including an optimal value and a sending delay time for preset mail blocking elements. This is an e-mail sending control system using a sending control rule inference result that receives a result from an inference engine and applies a sending control rule for each domain on the basis of the result, and generates a sending list by receiving mail sending information and classifying by domain. A shipment list generator; A distribution list for distributing distribution lists provided by the distribution list generation unit for each domain, and a distribution suitability for managing the distribution list information for each domain provided by the list distribution unit and based on the distribution control rules for each domain received from the inference engine. A domain controller for verifying through a mail sending result, and a domain controller for each domain is created and managed as a list. A sending control unit provided to the domain control unit and the inference engine, and removing the domain control unit from the list if the sending is completed, and adjusting the order of the corresponding domain control unit in the list if waiting to be sent in the spam mail classification state. It includes.

The domain control rule for each domain that the domain controller receives from the inference engine is based on an optimal value for mail blocking factors including a maximum number of mails, a maximum number of email errors, and a maximum number of connections, and a delay time for delaying spam mail determination. Include information about

The domain controller compares the queue storing the MX domain information, the recipient information, and the content in the shipment list, and the current shipment status with the shipment rule configured based on the domain-specific dispatch control rule information received from the inference engine, and provides the result. And a domain sending control unit for providing the mail sending information of the level corresponding to the sending rule if the sending state is possible.

When the domain dispatch control unit receives the inquiry on whether to send a message from the domain dispatch control management unit, it checks whether the current status is a spam mail classification status, notifies the dispatch delay status during the dispatch delay time of the dispatch rule, and when the spam mail classification status is not. If there are no recipients, it checks the completion of mail sending and informs the completion status. If there are any remaining recipients, it checks whether the current number of transmissions is less than the allowable number of senders and whether the current number of connections is less than the number of allowable connections. If exceeded, the dispatcher can be notified of the delay status, and if it is less than the allowable level, the dispatcher can provide information related to the dispatch.

The domain controller may be configured to configure a current shipment state through a shipment result provided by the domain dispatch control manager and to provide the domain dispatch controller upon request.

The domain controller increases the number of shipments and the number of accesses when a sending result is received, and if the result is success, increases the success value, and in case of failure, classifies the cause according to spam classification, email error, and server response timeout. Set the delay mode when classifying spam mails, and increase the failure count in other cases.

In this case, if the cause of the failure is an email error, it is determined whether the number of allowable email errors is exceeded by accumulating the email errors, and when the excess is exceeded, the possibility of being classified as spam is reduced by setting the delay mode before being classified as spam mail. In addition, if the cause of the failure is exceeding the server response time, it is checked whether the cumulatively increased response waiting time exceeds the allowable response time, and if it exceeds, the delay mode is set before being classified as spam, thereby reducing the possibility of being classified as spam.

The e-mail sending control system using the shipment control rule inference result according to the embodiment of the present invention determines the sending control rule for each receiving domain based on the inferred sending control rule, and manages the sending according to the sending control rule to improve the sending success rate. It can increase the effect.

The e-mail sending control system using the sending control rule inference result according to an embodiment of the present invention manages a domain controller that is dynamically generated according to a list for each domain, and each domain controller is based on the sending control rule inference information for the corresponding domain. It is possible to distribute and manage whether the data is sent to the server, and load balancing and management convenience are improved by the distributed management for each domain.

The email sending control system using the sending control rule inference result according to the embodiment of the present invention may perform delay management based on the allowable value according to the sending control rule inference information even though the domain controller checks the current sending status and is not classified as spam mail. By doing so, it is possible to circumvent the sending restriction due to the classification of spam mails.

1 is a configuration example of a conventional email sending system.
2 is a block diagram of a mass email sending system according to an embodiment of the present invention.
3 is an example of an email sending result log according to an embodiment of the present invention.
4 is a block diagram of an inference engine according to an embodiment of the present invention.
5 is a flowchart illustrating an example of an operation method of a log analyzer according to an exemplary embodiment of the present invention.
6 is a flow chart for explaining an example of the operation of the inference unit according to an embodiment of the present invention.
7 is an example of an MX domain for describing a delay time estimation according to an embodiment of the present invention.
8 is a block diagram of a mail sending system according to an embodiment of the present invention.
9 is a flowchart illustrating an example of a method of applying a dispatch control rule of a domain dispatch controller according to an exemplary embodiment of the present invention.
10 is a flowchart illustrating an example of a configuration of a dispatch environment of a domain controller according to an exemplary embodiment of the present invention.
11 is a graph comparing the performance of a mass email sending system and a conventional email sending system according to an embodiment of the present invention.

The present invention as described above will be described in detail with reference to the accompanying drawings and embodiments.

1 is a configuration example of a conventional email sending system, when a user writes and receives an email as shown, a message to be delivered through a mail user agent (MUA) 11, which is a client program installed in the client terminal 10. And send it to the sending Mail Transfer Agent (MTA) 12 using Simple Mail Transfer Protocol (SMTP), the sending MTA 12 delivers the email message to the receiving MTA 22, and the receiving MTA 22. Stores it in the mail database 23.

The Post Office Protocol (POP) / IMAP (Internet Message Access Protocol) server 24 connected to the MUA 21 of the server receiver client terminal 20 receives the e-mail stored in the mail database 23 to receive the MUA 21. The receiver checks the mail received through it.

Since such e-mails provide information or data using SMTP, which is built on trust in computer networks, e-mail recipients are supposed to receive messages sent by the sender without any doubt. Therefore, in the case of the existing mass email sending system, in order to speed up the sending, the MTA was distributed and sent in large quantities to send the mail. However, due to the anti-spam policy for filtering the spam mail, this method no longer ensures the success of the sending. It became impossible.

Currently, various spam mail blocking technologies are applied differently to each portal or email service provider. These spam mail blocking technologies are largely analyzed by the contents of the mail and the content filtering technology that determines the spam mail by checking the predetermined keyword and the origin of the mail. It is divided into technologies that control the process of delivery to the destination.

Here, in the case of normal mail transmission, spam filtering is often performed by a delivery process control technology rather than content filtering. As a policy and technology for filtering the spam mail by controlling the delivery process, RBL (Realtime Blackhole list) and SPF ( There are technologies such as Sender Policy Framework (policy), open relay blocking, sending limit, graylisting, tarpiting, challenge response, and port 25 blocking of dynamic IP users. .

Among these, the RBL or the SPF policy can be avoided when a normal sender sends an e-mail with the recipient's consent. However, various spam filtering techniques, such as blocking open relay, limiting volume, graylisting, tarpiting, challenge response, and blocking port 25 of dynamic IP users Even when sending, it can still be a problem for mass emailing. However, some of these spam filtering techniques are difficult to adopt by large portals or large email service providers because of the delay of transmission or the inconvenience of recipients.

Therefore, after analyzing and reviewing spam mail blocking policies and technologies based on representative portals, in most cases, SPF and RBL policies are applied, and the main spams include sending volume limit, access limit, sending delay, and account error limit. It is confirmed that it is used as a mail blocking element.

Of course, the policies applied to each portal were somewhat different, the criteria were also different, and there was a possibility of dynamic variable of the criteria, so it was difficult to measure fixed values and apply them uniformly to the values of the corresponding domains.

Therefore, in the embodiment of the present invention, the inference engine can be inferred by applying an inference engine that can infer an optimal value of the mail blocking elements according to different criteria for each domain of such recipients, and based on the inference information, We propose a mass email sending system that increases the sending success rate by domain by controlling sending by domain, and actively infers and responds to optimal values even when the mail blocking factors are variable. .

2 is a block diagram of a mass email sending system according to an embodiment of the present invention.

As shown, the mass email sending system includes an inference engine unit 200 that analyzes the log of the shipment result and provides a shipment control rule reasoning result including an optimal value and a delay time for preset mail blocking elements, and The mail sending system receiving the sending control rule inference result from the inference engine unit 200 is configured to send an e-mail by applying the sending control rule for each domain based on this, and provide the result log to the inference engine unit 200. do.

The inference engine unit 200 obtains a log of the sending result of the email sending list for each domain from the log management unit 140 to reconstruct the start and end of the mail sending performed for one domain into a virtual session and identify a cause of failure. The log analysis unit 220 cumulatively managing the analysis database 230 as session information, and the value of the mail blocking element of the normal session for a predetermined domain based on the session information of the log analysis unit 220 is a spam mail session. If it is larger, the mail blocking factor is excluded from the spam classification to obtain the value of the allowed mail blocking factor, and the sending delay time according to the spam classification is determined through subsequent successful session check, so that the information on the mail blocking factors and the sending delay are And a reasoning unit 210 for providing the mail control system inference result including time to the mail sending system 100.

The mail sending system 100 receives the sender information and the content and divides them by domains to generate a shipment list 111 and a shipment list provided by the shipment list generator 110. And a sending control unit 120 to control the sending for each domain according to the sending control rule inference result received from the inference engine unit 200 and provide the log of the sending result.

The dispatch control unit 120 may include a list distribution unit 121 for dividing the shipment list 111 provided by the shipment list generation unit 110 for each domain, and the domains provided by the list distribution unit 121. A domain controller 122 which manages the shipment list information as a queue and checks the suitability of shipment based on the shipment control rule for each domain received from the inference engine through a mail sending result, and the domain controller for each domain 122 Manage the list and ask the domain controller 122 whether it can be sent or not, if the shipment is possible, receive the shipment-related information and send it through the MTA 130, and then provide the result to the domain controller 122 and the reasoning. It provides the log management unit 140 as a log of the shipment result so that the engine unit 200 can use it. Delete (122) and, if sent to the spam classification idle it includes the sending domain control manager (123) to adjust the order in the list.

Through the above configuration, after sending mail for each domain, the shipping result log is analyzed to infer the sending rule, and the sending rate is controlled within the allowable range by controlling the sending per domain based on the inferred result, thereby increasing the success rate. Even if it fails, this information can be used to infer sending rules, so that the more accumulated the information is, the more optimized the mail sending rule can be. As a result, when a mail is unsuccessful, unlike a general method in which subsequent mails have a low success rate, even if a mail fails, subsequent success rates are high, and a successful subsequent mail time can be optimized.

In order to easily understand the configuration of FIG. 2 according to the embodiment of the present invention, which has been described above, the basic operation method will be described in more detail with reference to the inference engine 200, and then the dispatch control rule provided by the inference engine unit 200. It will also be described in detail with respect to the mail sending system 100 that controls the sending by domain using the inference result.

The illustrated inference engine unit 200 obtains a shipment log collected from the number of mails sent from the log database 141 of the log manager 140, and analyzes the log by the log analyzer 220.

The log manager 140 is shown separately from the mail sending system 100 and the inference engine 200 for convenience, but may be included in one of the mail sending system 100 and the reasoning engine 200. More specifically, it may be included as a part of the configuration of the domain dispatch control manager 123.

In the embodiment of the present invention, the mail blocking factor obtained as a result of the analysis of the spam blocking policy of the major large email service provider (portal, hosting company, etc.) as described above is determined as the main sending rule reasoning factor, and optimized for the reasoning. The reasoning engine 200 and the mail sending system 100 optimized for the mail sending control through the reasoning values for the mail blocking elements are provided.

The characteristics of the anti-spam policy obtained in this way are that if a large amount of mail is sent in a short period of time, excessive access to the mail receiving server is caused, and mail sending to an nonexistent address (account) is excessive, mail reception is blocked. In some cases, such blocking may be released after a specific time depending on the service provider after the reception of the mail is blocked. According to the characteristics of the spam mail policy, the inference engine 200 analyzes the provided log to obtain the number of mails sent, the number of mail server accesses, and the number of email errors for each domain. And provide these mail blocking elements and sending delay time to the mail sending system 100.

Therefore, such a distribution log includes individual unit information for easily confirming the information on the mail blocking element.

3 is an example of an email sending result log according to an embodiment of the present invention. As shown in FIG. 3, a log is generated to include the information of Table 1.

Component Explanation MEM_SEQ Receiver sequence number EMAIL Recipient email address MX RECORD MX domain SENDER_IP Email sending server IP ERR1 Whether an error occurred before joining the domain ERR2 An error occurred during the email sending step ERR3 Error code received from incoming mail server when it fails SEND_DATE Email sent date ERRMSG Error message received from receiving mail server when it fails

Here, the MX (Mail eXchange) domain refers to a plurality of subordinate mail server domains that are actually set for a single receiving domain, and they are collectively managed as one domain.

As shown in Table 1, the dispatch log basically shows information on sending a single message and various error information generated each time.

The inference engine 200 analyzes the shipment log of the distribution list for each domain to obtain the number of mails sent, the number of mail server accesses, and the number of email errors for each domain. .

4 is a block diagram of the inference engine according to an embodiment of the present invention, as shown in the log analysis unit 220 aggregates the log information for each domain to convert the analysis into a single virtual session information for analysis The virtual session conversion analyzer 221 is included.

The virtual session conversion analysis unit 221 aggregates the log for each domain and terminates when the mail sending action is stopped from the first mail sending start time and mail is not resent for a predetermined time (for example, 1 minute). If the reason for the failure of the mail is spam mail classification, it is determined that the session is terminated, and a virtual session having a corresponding life cycle is configured.

The virtual session may include session data as shown in Table 2 below.

Component Explanation Domain Mx Domain Start Time Session start time End time Session end time Total Connect count Number of connections to mail server during session persistence Total Send Count Total number of dispatches sent during session persistence Session Time Session persistence time Connection count Connections per minute Send count Shipments Per Minute Total Sucess count Mail success Email Error Count The number of cases where the mail failed cause was an email error. Session End Cause of session termination (spam classification, mail suspension)

Here, the number of logs is used as the number of logs during the life cycle of the session, and the number of connections to the email server is obtained by dividing the number of mails sent to the mail server from the number of transmissions to obtain the number of connections per minute and the number of transmissions per minute. The number of e-mail errors is calculated by summing the number of cases of an e-mail address error (an address that does not exist) among the causes of the mail failure. The criteria for judging an email address (account) error is a response code of 5xy (permanent negative completion response) during the RCPT TO (command defined in SMTP) during the email sending phase, and the spam classification keyword may not exist in the response message. If it is. The criteria for determining the classification of spam mail, which is one of the causes of session termination, are shown in Table 3 below.

ERR2 Shipment Step Spam classification criteria One connect Any code other than success code (220) 2 HELO Any code other than success code (250) 3 MAIL FROM Any code other than success code (250) 4 RCPT TO 4yz code that is a temporary negative completion response
Spam classification keywords are present in the response message 5 DATA 4yz code that is a temporary negative completion response 6 DOT Any code other than success code (250) 7 QUIT Success regardless of response

A summary of the method for generating a virtual session of the log analyzer 220 as described above may be illustrated in the flowchart of FIG. 5.

5 is a flowchart illustrating an example of an operation method of a log analyzer according to an exemplary embodiment of the present invention. As shown in FIG. 5, the virtual session conversion analyzer 221 of the log analyzer 220 may complete a predetermined domain. By collecting the log of the log, the life cycle of the session is analyzed to calculate the number of connections and the number of shipments per minute.

If the success of the session is classified and classified, the session data is generated by counting the number of successful emails. If the session termination is due to an email error, the session data is generated by counting an email error. According to the spam classification as shown in Table 3, the session data including the spam classification information is generated and provided to the analysis database 230 and the inference unit 210.

The inference unit 210 is a block factor estimator 212 for obtaining an actual block element and an optimal value among the mail block elements, and a send delay time estimator 211 for obtaining an optimal delay until the next shipment when classifying spam mails. ).

6 is a flowchart illustrating an operation of the blocking element estimator 212 according to an embodiment of the present invention. As shown in FIG. It collects the maximum number of mail blocking elements such as the number of connections, and collects the maximum number of mail blocking elements such as the maximum number of mails, the maximum number of email errors, and the maximum number of connections in sessions where the session termination is normal.

If there is only a normal session without a spam session, the normal session value is returned as the session value. If there is only a spam session without the normal session, then the spam session value is returned as the session value, and if there is information about the normal session, the spam session is present. If there is a spam session value, the spam session value is excluded from the mail blocking element by comparing the spam session with the normal session for the maximum number of sending, maximum number of email errors, and maximum number of connections. Returns. In other words, an item with a normal session value greater than the spam session's value means that the item's current value is acceptable and is not considered spam by this. Therefore, the largest value of the item's normal session value is the current item's current value. It can be judged by an optimal value. Therefore, the information is included as allowable maximum information in the shipment control rule information to be provided to the mail sending system.

Since the determination process may use shipment logs obtained through multiple IP distributed shipment information for multiple MX domains, a plurality of virtual sessions for one domain may be substantially obtained for each MX domain and for each IP. The same analysis can be made easier.

On the other hand, as shown in Figure 4, the inference unit 210 includes a sending delay time estimating unit 211 for estimating the sending delay time until the next normal sending when the spam mail is classified, the sending delay time estimating unit 211 ) Determines the sending delay time until the start of a subsequent success session for the same domain based on the spam session when a spam session occurs for the same domain.

Here, the sending delay time estimator 211 determines the sending delay time through a subsequent success session check, but uses information obtained from another IP or the corresponding domain through the log or log analyzer accumulating information that is subsequently stored. By referring to all of the information on the multiple MX domains of the can determine the delay time for the entire domain or the MX domain of the domain.

FIG. 7 illustrates an example in which a plurality of MX domains exist for a single domain. In the case of hamail.net, 10 MX domains are configured so that at least 10 different virtual session information can be parallelized for each MX domain of the same domain. Since it can be obtained by analyzing the optimal sending delay time can be analyzed, if necessary, the inference unit 210 is to send a different delay time for each of the multiple MX domains of the same domain mail sending system 100 The shortest sending delay time among the sending delay times obtained from the successful MX domain may be determined as the sending delay time for the corresponding domain.

Of course, in the case of distributed processing through a plurality of IPs, information about more virtual sessions can be obtained in parallel for each IP, so that the blocking element estimator 212 and the sending delay time estimator 211 provide a plurality of MXs. Multiple IPs for domain and distributed processing provide information that mixes normal and spam classifications, successes, and failures, which can be used to optimize mail blocking factors and sending delays. As a result of repeated optimization, it is possible to estimate the optimal successful delivery rule for a domain that is frequently repeated.

An optimized sending control method of the mail sending system 100 that has received the sending control rule inference result obtained by the inference engine unit 200 will be described with reference to FIG. 8.

8 is a block diagram of a mass email sending system showing the configuration of the mail sending system 100 according to an exemplary embodiment of the present invention in more detail.

As shown, the mail sending system 100 is divided into a sending list generating unit 110 and a sending control unit 120.

The shipment list generating unit 110 checks the recipient file and checks the recipient module file, and generates a shipment list 113 for each domain by dividing the recipient file into domains. )

As described above, the dispatch controller 120 includes a list distributor 121, a domain controller 122, and a domain dispatch control manager 123, and the domain controller 122 is configured for each domain therein. The domain control unit 310 of the domain, and the domain control unit 320 is configured therein and manages the shipment list as a queue.

If there is no domain controller 310 of a specific domain corresponding to the dispatch list of the list distributor 121, the domain dispatch control manager 123 generates a corresponding domain controller 310 and manages the list together with other domain controllers. In addition, when receiving a dispatchable flag by inquiring whether or not the domain dispatch controller 320 in the domain controller 310 sends the message, the MTA 130 may further receive shipment related information (recipient list, mail content, MX domain address information, etc.). After sending through, the result is provided to the corresponding domain controller 310. If the flag indicating that the sending is completed is received, the corresponding domain controller 310 is deleted from the list. When receiving a flag indicating the status of spam mail when the domain sending control unit 320 in the domain control unit 310 managed by the list is sent or not, the domain control unit 310 adjusts the order in the list to adjust the sending inquiry order. Later, it is usually postponed to the last list if you have a spam category.

Here, the domain controller 310 generated to manage the shipment for a specific domain receives a shipment control rule inference value of the corresponding domain from the inference engine unit 200 and sets and updates the current shipment status based on this. The received sending control rule inference value includes an optimal value for mail blocking elements including a maximum number of mails, a maximum number of email errors, and a maximum number of accesses, and information about a delay time when determining spam mails.

That is, the domain controller 310 receives the shipment control rule inference value from the inference engine unit 200, manages the current shipment state, and analyzes the actual shipment control rule according to the request of the domain dispatch control manager 123. As a result, the domain sending control unit 320 configured therein manages the list of recipients to be provided and sent, the contents of mail, and the MX domain information.

The domain dispatch control unit 320 includes a queue for storing MX domain information, recipient information, and contents in a shipment list, and a shipment rule configured based on current shipment control rule information received from the inference engine unit based on a current shipment status. The results of the comparison are provided as follows, and an example of a method of confirming such a shipment status is shown in the flowchart shown in FIG.

9 is a flowchart illustrating an example of a method for applying a shipment control rule of the domain dispatch control unit 320 according to an embodiment of the present invention. As shown in FIG. If the status of spam mail is classified, whether or not the recipient is left, whether the current number of connections and the number of sendings are exceeded, and whether the sending is completed through the session completion time, the session delay flag (Session_delay return) and the session maximum flag (Session_MAX return) are determined. In response to the query from the domain dispatch control management unit 123, one of the flag (Session_empty return), the completion flag ((Session_End return)), and the sendable flag (Session_send return) indicating that the session is temporarily empty.

According to the sequence shown, first, it checks whether the current status is classified as spam mail. When receiving a message from the receiving mail server that the message has been classified as spam mail, it fails to send even if it tries to send mail for a certain time. When a request for confirmation is made that the mail can be sent in the spam classification mode for a time, the response is sent with a delay flag for the corresponding sending delay time.

If it is not currently classified as spam, check if there is a list of mail recipients. If there are no recipients in the list, check if there are no remaining recipients. If there are no remaining recipients even at the preset time (60 seconds, shown) Is determined to be completed and responds with a completion flag to delete the corresponding domain controller 310 from the list, and if the time is not yet set, an additional list may be further received. Answer.

If there are any remaining recipients in the recipient list, check whether the current number of connections exceeds the allowable number of connections specified in the shipment control rule, and whether the current number of shipments exceeds the allowable number of minutes per minute specified in the shipment control rule. In case of exceeding, the response is delayed by sending the session maximum flag, and in both cases, the sending process is sent by sending the sendable flag if it is within the allowable number. Various choices)

As a result of confirming the current sending status in the same manner as shown in FIG. Is provided to the corresponding domain controller 310, through which the domain controller 310 configures or reconfigures (updates) the current dispatch status.

FIG. 10 is a flowchart illustrating an example of a method for configuring a sending environment of the domain controller 310 according to an exemplary embodiment of the present invention. As shown in FIG. Check if the result is a success, if successful, increase the number of successes.

If it is failed to send, classify the cause according to spam classification, email error, and server response timeout, and set the current status to delay mode so that sending is delayed during the sending delay time according to spam classification. The case increases the failure number by readjusting (accumulating) the value of the failure.

If not shown, the reason for the failure, not the classification of spam mail, is to contact the receiving mail server and wait for a response. If the response does not occur during the preset response waiting time (Soket time out), the response waiting time is further increased by the preset level. Increases the value, if the response wait time is equal to or longer than the allowable response wait time, also set the current state to delay mode.

Also, if the failure, not the spam classification, is caused by an email error (no email address), accumulate the email error number and increase the failure number, if the cumulative email error is greater than or equal to the number of email errors allowed. It also sets the current state to delay mode.

This may be because the failure is due to an email error and the cumulative error number is equal to or greater than (or exceeds) the allowed email error, or when the response wait time that is cumulatively increased while exceeding the server response time is equal to or greater than (or exceeds) the response time. ) By setting the current status to delay mode (which can be set differently from the spam classification) before being classified as spam mail, the delay time can be reduced because the possibility of classification as spam mail is reduced.

In conclusion, as described above, the inference engine block infers the spam blocking rule and actively uses the inferred mail sending rule information to be used in the mail sending system. E-mail can be sent in a state, which can increase the success rate of the e-mail.

FIG. 11 is a graph comparing the performance of a mass email sending system and a conventional mass email sending system according to an embodiment of the present invention. In the email sending system according to an embodiment of the present invention, as information on a specific domain is accumulated, the optimal Since the success of sending control rule can be inferred, the success rate is higher when repeated sending, but in case of the existing system, the sending success rate is always the same.

In the above described and illustrated with respect to preferred embodiments according to the present invention. However, the present invention is not limited to the above-described embodiments, and various changes and modifications may be made by those skilled in the art without departing from the scope of the present invention. .

100: mail sending system 110: shipping list generation unit
120: sending control unit 121: list distribution unit
122: domain control unit 123: domain dispatch control management unit
130: MTA 140: log management unit
200: inference engine unit 210: inference unit
220: log analysis unit 230: analysis database

Claims (8)

delete delete delete Based on the result of the shipment control rule inference that receives the result of the inference control rule including the optimal value and the delay time for the preset mail blocking elements from the inference engine, and applies the control control for each domain based on the inference, As an email sending control system,
A shipment list generation unit configured to receive the mail shipment information and generate a shipment list by domain;
A distribution list for distributing distribution lists provided by the distribution list generation unit for each domain, and a distribution suitability for managing the distribution list information for each domain provided by the list distribution unit and based on the distribution control rules for each domain received from the inference engine. A domain controller for verifying through a mail sending result, and a domain controller for each domain is created and managed as a list. A sending control unit including a domain sending control management unit provided to the domain control unit and the inference engine, and deleting the domain control unit from the list if the sending is completed, and adjusting the order of the corresponding domain control unit in the list if waiting to be sent in the spam mail classification state. Including ,
The domain controller compares the queue with the MX domain information, the recipient information, and the content in the shipment list, and compares the current shipment status with a shipment rule configured based on the domain-specific dispatch control rule information received from the inference engine, and provides the result. And a domain sending control unit that provides the domain sending control management unit with mail sending information at a level that meets a sending rule if the mailing state is possible.
When the domain dispatch control unit receives the inquiry on whether to send a message from the domain dispatch control management unit, it checks whether the current status is a spam mail classification status, notifies the dispatch delay status during the dispatch delay time of the dispatch rule, and when the spam mail classification status is not. If there are no recipients, it checks the completion of mail sending and informs the completion status.If there are any remaining recipients, it checks whether the current number of transmissions is less than the allowable number of senders and whether the current number of connections is less than the number of allowable connections. If the number is exceeded, the sending delay status, and if the number is less than the allowable all, the e-mail sending control system using the results of the shipment control rule inference, characterized in that the information provided by the shipment control rules.
delete Based on the result of the shipment control rule inference that receives the result of the inference control rule including the optimal value and the delay time for the preset mail blocking elements from the inference engine, and applies the control control for each domain based on the inference, As an email sending control system,
A shipment list generation unit configured to receive the mail shipment information and generate a shipment list by domain;
A distribution list for distributing distribution lists provided by the distribution list generation unit for each domain, and a distribution suitability for managing the distribution list information for each domain provided by the list distribution unit and based on the distribution control rules for each domain received from the inference engine. A domain controller for verifying through a mail sending result, and a domain controller for each domain is created and managed as a list. A sending control unit including a domain sending control management unit provided to the domain control unit and the inference engine, and deleting the domain control unit from the list if the sending is completed, and adjusting the order of the corresponding domain control unit in the list if waiting to be sent in the spam mail classification state. Including ,
The domain controller compares the queue with the MX domain information, the recipient information, and the content in the shipment list, and compares the current shipment status with a shipment rule configured based on the domain-specific dispatch control rule information received from the inference engine, and provides the result. And a domain sending control unit that provides the domain sending control management unit with mail sending information at a level that meets a sending rule if the mailing state is possible.
The domain controller configures a current shipment state through a shipment result provided by the domain dispatch control manager, and provides the domain dispatch controller upon request.
The domain controller increases the number of shipments and the number of accesses when a sending result is received, and if the result is success, increases the success value, and in case of failure, classifies the cause according to spam classification, email error, and server response timeout. An e-mail sending control system using the result of the shipment control rule inference, characterized in that it is set to the delay mode when classifying spam mails and the failure number is increased in other cases.
The method of claim 6, wherein when the cause of the failure is an email error, the error is accumulated by checking whether the number of email errors has been accumulated and exceeded. E-mail sending control system using the result of sending control rule inference, characterized in that.
The method of claim 6, wherein if the cause of failure is exceeding the server response time, the cumulatively increased response waiting time is determined to exceed the allowable response time. E-mail sending control system using the result of sending control rule inference, characterized in that to lower.

Images