Project Topics

Engineering Projects

Spam Zombie Detection and Blocking Mechanism

Published on Jan 19, 2017


The software is for automation of Health Centre monitoring. The software includes maintaining patient details, providing services, fixing appointments, providing and maintaining all kinds of tests for patients and billing report generator.


In today‟s computing world, almost in every aspect internet plays an important role in our daily lives. Internet not only influences the people to do positive works but also influences the people to trouble others by posing many attacks like doing spam sending viruses in email. These attacks are posed by the attackers indirectly or directly. There are two types of Attacks, one of them is manual attacks and the other type is automatic attacks. Most of the successful attacks are from the automated generated code injected by the attackers. These are very dangerous some of them are DDos, Dos, E-mail Worms, Worms, Viruses. These machines are called zombies, drones or compromised machines. The Report of 2012 march says that more than 75% of traffic is spam and in that 0.4% was malicious these spam zombies is tough job for the system administrators.

The Spam message is an unwanted message to the users because it occupy the network bandwidth, disk space, connection time, money and hide viruses inside spam message .The Spam filtering is a technique that classifies a message into two categories (good and spammed message). Effective spam filter aims to minimize the false positive percentage. There are many methods available to filter out the spam.


A. CI Anti spamtechniques:

I Anti-spam techniques is designed on basis of behavioral characteristics. This anti-spam techniques need to understand the behavioral characteristics of spammers that distinguishes it from senders of non spam messages. The effectiveness and feasibility of CI anti-spam techniques[3] affected by the behavioral characteristics of the spammers such as distributions of non-spam and spam messages by spam statistics, ratios of spam messages from different spammers etc.


1. Botsniffer has very promising detection accuracy with very low false positive rate.


1. Botnet CC traffic is difficult to detect.

B. DMTP filtering models

In Receiver-pull, it allow receiver to control over system and when they want any data from sender. In Sender-push, delivery of traffic is controlled by a sender and receivers just accept whatever sender had sent. DMTP (Differentiated mail transfer protocol) [5] which is a pull based model as a counterpart to the spam problem, which grants the control over the message delivery to the receiver. In the push-based white list and black list are defined along with receiver to determine whether to accept the message.


1. DMTP can easily deploy on internet.

2. Spam retrieval behaviors of receivers determine the delivery rates of spam.


1. New patterns are hard to detect.

C. Spam Zombie Blocking:

This system not only detects the spam zombies, but also has the facility of blocking the zombie system. This system mainly gives results on the basis of the false negative and false positive error rates. These false negative and false positive probabilities can be bounded by user defined threshold value. If the number of the messages send by the particular machine exceeds the threshold limit then that machine is considered to be compromised.

D. Parameter based filtering:

Parameter based filtering is used to classify a message as either non spam or spam by considering the parameters of the message. Parameters of a message include To, From, Received by, IP address, Subject etc. By using parameter based filtering most of the spam messages will be filtered or detected. Most of the messages will be filtered by checking the parameters of the message [6]. Some of the parameter based filtering techniques are Whitelists, Blacklists, Challenge/Response methods etc.Blacklists are the IP addresses/e-mail addresses/domain names of the real time spammers. In real world many black lists are available. These are called Realtime Black Lists (RBL).

E. Content Based Spam filtering Methods

Content based spam filtering works on the content of the message ie, the body of the message to decide it‟s a spam or not spam. These methods depend on the training of the previous messages.

F. K-Nearest Neighbors:

This method of classification is based on the distance measure among the messages. The distance can be measured based on the features between the messages by equations like Euclidean distance measurement. This method doesn‟t need training phase where the incoming messages will be directly measured with the available sample messages. So the time complexity of each message is of O (n)

Spam Zombie Detection

System Architecture

The Existing system has some basic steps like Account authentication, sending mails, SOPT detection, CT detection, and PT detection as shown in fig 1 to identify the message is spam or not.

Spam Zombie Detection

The SPOT System or existing system will be connected with more than two machines. So when machines send message it will be 1st pass to the SPOT system as shown in fig2.


The Proposed system use ID 3 algorithm logic [10] where it will get called instead of SPOT decision making logic for filtering purpose. Decision tree is one of widely used filtering or classification method in Data Mining field whose core problem is the choice of splitting attributes. In ID3 algorithm, information theory is applied to choose the attribute that has the biggest information gain as the splitting attribute in each step and a recursive way is used to generate decision tree until threshold condition is reached. The proposed improved ID3 based on weighted modified information gain called ωID3 is introduced in the proposed system.

Also proposed system will store newly introduced email spam behavior in the database which will also not only increase but also help to filter the input messages without any calculation and it will increase the existing system performance. In the existing system ID3 algorithm will try to improve the performance and accuracy of the system. As shown in fig 3 the ID3 algorithm will do the filtering process just before the message is sent to SMTP server. So the steps will be Account authentication (Gmail authentication), sending mails, ID3 algorithm, CT detection, and PT detection. Additional to this database also help to try to improve the performance of existing system.


The blocking functionalities of the system works as follows:

1: System is a Zombie.

2: Let n be the number of the important mails.

3: „GId‟ be Gmail ID and „PW‟ Gmail password,

4: declare total and threshold value

4: if (entered Gmail ID==GId) and (Pass word== PW) then

5: start sending the mail for compose mail screen.

6: for loop i=0 to i<5

7: choose the correct mail sender

8: If(choose correct sender) then

9: total++

10: endIf

11: endfor

12: if (total>=threshold) then

13: continue with account.

14: change your password.

15: else

16: block the machine temporarily.

17: senders „Physical address‟ block.

18: endelse

19: else

20: enter correct question and password.

21: endelse.

If the system is found as a Zombie system it has blocked temporarily and the user of that system when tries to login then he is informed that the system has been blocked. If the user wants to recover the system then it works as per the above algorithm.

Performance Evaluation

Finding out the accuracy of the spam detection system is to find out how accurate it detects the spam from test dataset. Proposed system used the down emails from different training data set. True Positive (TP), states the number of spam mails correctly classified as spam. True Negative (TN) states the number of non spam mails correctly classified as non spam. False Positive (FP) states the number spam mails classified as non spam. False Negative (FN) states the number of non spam mails classified as spam. Accuracy gives the performance of the system. The following tables show that increase in the dataset the accuracy of system is improved.

Spam Zombie Detection


proposed system is using the ID3 algorithm in order to detect the spam zombies. Depending upon the threshold limit system design to minimize the number of the required observation for detecting the spam zombies. The proposed system gives a complete spam detection system which can efficiently process the matching of spam emails. It also provide the temporarily blocking mechanism in which if the system is identified as the spam zombie then the system physical address gets blocked so that it cannot send the spam messages further.

The proposed system will try to help to recover the blocked or compromised system in case if the system was hacked by an attacker and was used as a spam zombie. The systems Performance was evaluated on basis its accuracy. The comparative study of existing and proposed system concludes that proposed system is more efficient and accurate.


[1] Zhenhai Duan, Peng Chen, Fernando Sanchez Florida State University, Yingfei Dong “Detecting Spam Zombies by Monitoring Outgoing Messages”University of Hawaii, Mary Stephenson, James Barker Florida State University,IEEE Transaction on dependable and secure Computing Vol.9 No2 March/April 2012

[2], Maria Namestnikova on June 14, 2012.

[3] Alex Brodsky University of Winnipeg, Winnipeg, MB, Canada, R3B 2E9, “Dmitry BrodskyMicrosoftCorporation,Redmond”,WA,USA,98033,USENIX Association Berkeley, CA, USA ©2007

Related Projects