Technique for detecting suspicious electronic messages

• US 10,572,664 B2
• Filed: 09/18/2017
• Issued: 02/25/2020
• Est. Priority Date: 09/19/2016
• Status: Active Grant

• Current Assignee: Retarus GmbH
• Original Assignee: Retarus GmbH
• Inventors: Hager, Martin, Grauvogl, Michael
• Primary Examiner(s): Pham, Luu T
• Assistant Examiner(s): Jackson, Jenise E
• Application Number:
  US15/706,796
• Publication Number:
  US 20180082062A1
• Time in Patent Office: 890 Days
• Field of Search: 726 24
• CPC Class Codes: G06F 21/564 by virus signature recognition
• CPC Class Codes: H04L 2463/121 Timestamp
• CPC Class Codes: H04L 2463/144 Detection or countermeasures against botnets
• CPC Class Codes: H04L 51/212 using filtering or selective blocking
• CPC Class Codes: H04L 63/123 received data contents, e.g. message integrity
• CPC Class Codes: H04L 63/1408 by monitoring network traffic monitoring network traffic per se H04L43/00
• CPC Class Codes: H04L 63/1416 Event detection, e.g. attack signature detection
• CPC Class Codes: H04L 63/145 the attack involving the propagation of malware through the network, e.g. viruses, trojans or worms
• CPC Class Codes: H04L 63/164 at the network layer
• CPC Class Codes: H04L 63/308 retaining data, e.g. retaining successful, unsuccessful communication attempts, internet access, or e-mail, internet telephony, intercept related information or call content

Abstract

The disclosure relates to a method of detecting suspicious electronic messages. The method is performed in a messaging server which is in communication with a plurality of message senders and a plurality of message receivers, and comprises the steps of: receiving electronic messages sent from the plurality of message senders to at least one message receiver; extracting from each received message at least one message sender feature and at least one message content feature; recording the extracted message sender features and message content features in a database; determining, on the basis of the message content features recorded in the database, whether a specific content feature that can be associated with a current message has already been recorded in the past; if the specific content feature has already been recorded in the past, determining, on the basis of the message sender features recorded in the database, a number of message senders that can be associated with the specific content feature; and classifying the current message as suspicious if the determined number of message senders that can be associated with the specific content feature exceeds a predetermined threshold value. Also disclosed is a messaging server implementing the above described method.

Claims

1. A method of detecting suspicious electronic messages, wherein the method is performed in a messaging server which is in communication with a plurality of message senders and a plurality of message receivers, wherein the method comprises the steps of:

receiving electronic messages sent from the plurality of message senders to at least one message receiver;

extracting from each received message at least one message sender address feature (AF) and at least one message content feature (CF);

recording the extracted at least one message sender address features (AF) and at least one message content features (CF) in a database;

determining, on the basis of the message content features (CFs) recorded in the database, whether a specific content feature that can be associated with a current message has already been recorded in the past;

if the specific content feature has already been recorded in the past, determining, on the basis of the message sender address features (AFs) recorded in the database, a number (N) of message senders that can be associated with the specific content feature; and

classifying the current message as suspicious if the determined number (N) of message senders that can be associated with the specific content feature exceeds a predetermined threshold value (N1), wherein the predetermined threshold value (N1) is dynamically adjusted,

wherein time-stamped message sender address features (AFs) and message content features (CFs) are recorded in two separate index data structures, wherein a first index data structure (IDX1) comprises a data set (ts, CF) of time-stamped message content features (CFs) and a second index data structure (IDX2) comprises a data set (ts, CF, AF) of time-stamped message content features (CFs) and message sender address features (AFs), and

wherein if the current message has been classified as suspicious, the method further comprising at least one of the following steps;

blocking the current message; and

subjecting the current message to an anti-virus (AV) analysis.

2. The method according to claim 1, further comprising generating timestamps (ts) and recording the timestamps (ts) along with the extracted message sender address features (AFs) and message content features (CFs) in the database.

3. The method according to claim 2, wherein the recording step further comprises:

organizing the time-stamped message sender address features (AFs) and message content features (CFs) into at least one index data structure (IDX1, IDX2).

4. The method according to claim 1, wherein the step of determining whether specific content features has already been recorded in the database comprises:

performing an identity or similarity check between the message content feature (CF) associated with the current message and the recorded message content features (CFs) in the database.

5. The method according to claim 4, wherein a database look-up is performed in order to determine whether a content feature record identical or similar to the specific content feature already exists in the database for a predetermined time window in the past.

6. The method according to claim 4, wherein if a message content feature record identical or similar to the specific content feature already exists in the database for a predetermined time window, determining how many message sender address features (AFs) can be related to the existing content feature record for the predetermined time window.

7. The method according to claim 5, wherein if a message content feature record identical or similar to the specific content feature already exists in the database for a predetermined time window, determining how many message sender address features (AFs) can be related to the existing content feature record for the predetermined time window.

8. The method according to claim 1, wherein the classifying step further comprises at least one of the following processes:

tagging the current message as suspicious message; and

registering the content of the current message as spam or malicious content in a blacklist.

9. The method according to claim 1, further comprising:

routing the current message to the intended message receiver if the anti-virus (AV) analysis reveals that the message is not malicious.

10. The method according to claim 1, wherein the at least one extracted message sender address feature (AF) is indicative of a sender address or sender address portion.

11. The method according to claim 1, wherein the at least one extracted message content feature (CF) is indicative of an attachment of the message, subject line content of the message, uniform resource locator (URL) comprised in the message and/or portions thereof.

12. A computer program product being stored on a non-transitory computer readable recording medium comprising program code portions for carrying out a method comprising the steps of:

receiving electronic messages sent from the plurality of message senders to at least one message receiver;

extracting from each received message at least one message sender address feature (AF) and at least one message content feature (CF);

recording the extracted at least one message sender address features (AF) and at least one message content features (CF) in a database;

determining, on the basis of the message content features (CFs) recorded in the database, whether a specific content feature that can be associated with a current message has already been recorded in the past;

if the specific content feature has already been recorded in the past, determining, on the basis of the message sender address features (AFs) recorded in the database, a number (N) of message senders that can be associated with the specific content feature; and

classifying the current message as suspicious if the determined number (N) of message senders that can be associated with the specific content feature exceeds a predetermined threshold value (N1), wherein the predetermined threshold value (N1) is dynamically adjusted,

wherein time-stamped message sender address features (AFs) and message content features (CFs) are recorded in two separate index data structures, wherein a first index data structure (IDX1) comprises a data set (ts, CF) of time-stamped message content features (CFs) and a second index data structure (IDX2) comprises a data set (ts, CF, AF) of time-stamped message content features (CFs) and message sender address features (AFs), and

wherein if the current message has been classified as suspicious, the method further comprising at least one of the following steps;

blocking the current message; and

subjecting the current message to an anti-virus (AV) analysis.

13. A messaging server for detecting suspicious electronic messages, wherein the messaging server is in communication with a plurality of message senders and a plurality of message receivers, the messaging server being configured to receive electronic messages sent from the plurality of message senders to at least one message receiver, the server comprising:

an analyser configured to extract at least one message sender address feature (AF) and at least one message content feature (CF) from each received message;

a recorder configured to record the extracted at least one message sender address features (AF) and at least one message content features (CF) in a database;

a determiner configured to determine, on the basis of the message content features (CFs) recorded in the database, whether a specific content feature that can be associated with a current message has already been recorded in the past, and if the specific content feature has already been recorded in the past, to further determine, on the basis of the message sender address features (AF) recorded in the database, a number (N) of message senders that can be associated with the specific content feature; and

a classifier configured to classify the current message as suspicious if the determined number (N) of message senders that can be associated with the specific content feature exceeds a predetermined threshold value (N1), wherein the predetermined threshold value (N1) is dynamically adjusted,

wherein time-stamped message sender address features (AFs) and message content features (CFs) are recorded in two separate index data structures, wherein a first index data structure (IDX1) comprises a data set (ts, CF) of time-stamped message content features (CFs) and a second index data structure (IDX2) comprises a data set (ts, CF, AF) of time-stamped message content features (CFs) and message sender address features (AFs), and

wherein if the current message has been classified as suspicious, the classifier further configured to;

block the current message; and

subject the current message to an anti-virus (AV) analysis.

14. The messaging server according to claim 13, further comprising a time stamper configured to provide a timestamp (ts) for each extracted message sender address feature (AF) and message content feature (CF).