Honeypot in Network Security
Honeypot in Network Security
Honeypot in Network Security
A
SEMINAR REPORT
ON
HONEYPOT IN NETWORK SECURITY
Abstract
Day by day, more and more people are using internet all over the world. It is
becoming a part of everyone’s life. People are checking their e-mails, surfing
over internet, purchasing goods, playing online games, paying bills on the
internet etc. However, while performing all these things, how many people know
about security?
Do they know the risk of being attacked, infecting by malicious software?
Even some of the malicious software are spreading over network to create more
threats by users.
How many users are aware of that their computer may be used as zombie
computers to target other victim systems?
As technology is growing rapidly, newer attacks are appearing.
Security is a key point to get over all these problems. In this thesis, we will make
a real life scenario, using honeypots.
Honeypot is a well design system that attracts hackers into it. By luring the
hacker into the system, it is possible to monitor the processes that are started and
running on the system by hacker.
In other words, honeypot is a trap machine which looks like a real system in
order to attract the attacker.
The aim of the honeypot is analyzing, understanding, watching and tracking
hacker’s behaviour in order to create more secure systems.
Honeypot is great way to improve network security administrators’ knowledge
and learn how to get information from a victim system using forensic tools.
Honeypot is also very useful for future threats to keep track of new technology
attacks.
Keywords
Honeypot, hacking, security, forensic analysis of honeypots, network.
HONEYPOT IN NETWORK SECURITY
Acknowledgments
We would like to thank Ola Flygt to guide us and support our work, and supply
the necessary equipment for our thesis process.
Table of content
1. INTRODUCTION .......................................................................................................... 1
1.1 Problem Description ..................................................................................................... 1
1.2 Motivation .................................................................................................................... 1
1.3 Goals ............................................................................................................................. 1
1.4 Restrictions ................................................................................................................... 2
1.5 Report Structure ............................................................................................................ 2
2. HONEYPOTS AND THEIR AIMS ............................................................................... 3
2.1 What is a honeypot ? .................................................................................................... 3
2.2 Research honeypots ...................................................................................................... 3
2.3 Production honeypots ................................................................................................... 3
2.3.1 Prevention .............................................................................................................. 3
2.3.2 Detection ................................................................................................................ 4
2.3.3 Response ................................................................................................................ 4
2.4 History of Honeypots ................................................................................................... 4
2.5 Advantages of honeypots ............................................................................................. 5
2.6 Disadvantages of honeypots ......................................................................................... 5
2.7 Levels of interactions in honeypots .............................................................................. 5
2.8 Wireless Honeypots ...................................................................................................... 6
2.8.1 Why Wi-Fi Honeypots ? ........................................................................................ 6
2.8.2 Wireless Honeypot History ................................................................................... 6
2.8.3 Honeyspot .............................................................................................................. 7
3. LOOKING AT THE SECURITY PROBLEMS CLOSER ............................................ 9
3.1 Legal issues with honeypots ......................................................................................... 9
3.1.1 Using honeypots are illegal or not? ....................................................................... 9
3.1.2 Different laws in different countries .................................................................... 10
3.2 Security risks .............................................................................................................. 11
3.3 Things to consider while configuring a honeypot ...................................................... 12
4. PRACTICAL IMPLEMENTATION ........................................................................... 13
4.1 Starting to honeypots .................................................................................................. 13
4.2 Starting with low level interaction honeypots :Hone ............................................ 13
4.3 Continuing with medium level of interaction honeypots : Nepenthes ....................... 16
4.4 High level of interaction honeypots :Honeywall ....................................................... 17
5. FORENSICS POINT OF VIEW ..................................................................................... 26
HONEYPOT IN NETWORK SECURITY
6. RESULTS ........................................................................................................................ 30
7. CONCLUSION ............................................................................................................... 32
8. REFERENCES ................................................................................................................. 34
1. INTRODUCTION
We will have two perspectives which are a forensic examiner and a hacker. We
will use variety of hacking tools and forensic examiner tools to have very
accurate results.
1.2 Motivation
First of all, we are very interested in this subject field of study.
So, our motivation for this thesis is to understand how security systems are
working and how an organization can be protected and being aware of the risks of
security flaws in the system.
We will learn how a system is working and how it can be developed. Once we
have the results, we will examine the output with forensic science tools.
While trying all these, we will come across some problems and we will try to
solve it.
At the same time we will have experience on creating and managing this kind of
systems for the future. If we see similar problems in a network, we will be able to
handle the system and recover the loss.
1.3 Goals
We will find answers to all the questions that we stated in problem description
part. Are the honeypots secure?
Does the hacker know that it is a trap system? If the hacker realizes that it is a
trap system does he continue attacking to it?
What does he gain from attacking it?
Is it possible for the hacker to reach other systems and compromise them? Our
perspective is to solve the problems related to security, how a honeypot can be
deployed, and the amount of information that we can get.
We will look into the restrictions honeypot implementationmainly in EU and
USA including which laws exist, how far a network security administrator can go
to obtain information and track the hacker.
We will explain and come up with some discussions regarding what should be
done and what should not be done with respect to the laws.
We will have some opinions and suggestions based on our work. While we will
be looking for answers for security problems, we will also evaluate and think
about the limits of the experiment.
1.4 Restrictions
We will state the legal issues and uses of honeypots as well as their risks.
Forth chapter will be practical implementation which includes the work we
accomplished.
Chapter five will cover other honeypot products which exist in the market.
Forensic part of the honeypots will be explained at chapter six.
We will write the interview about honeypots that we did with lawyer PehrJern at
chapter seven.
We will show the results at chapter eight and finish with conclusion at chapter
nine.
The objective is to learn how to protect a system better, they do not bring any
direct value to the security of an organization.
2.3.1 Prevention
2.3.2 Detection
This technology will help users to know if the network is compromised, but it
will not prevent hackers from attacking the system. For companies, such
detection systems are expensive.
At this point, honeypots are valuable to monitor the activity.
2.3.3 Response
In this part, we will give the history of honeypots so far according to Lance
Spitzner
(2002):
1990-1991: It is the first time that honeypot studies released by Clifford Stoll
(The Cuckoo’s Egg) and Bill Cheswick (An Evening WithBerferd).
1997: Deception Toolkit version 0.1 was introducted by Fred Cohen. After
Clifford Stoll (The Cuckoo’s Egg) and Bill Cheswick (An Evening With Berferd)
, Deception Toolkit gave an idea of first honeypot structure.
1998: First commercial honeypot was released which is known as CyberCop Sting.
1998: BackOfficer Friendly honeypot was introduced. It was free and easy to
configure. It is working under Windows operating system. Most of the people
tried this software and the concept of honeypot became more and more known
among people.
1999: After BackOfficer Friendly, people were more into this new technology.
Honeynet project started at this year. Also, Know Your Enemy papers were also
released. Thanks to these releases, people understood the aim of the honeypots
more.
2000-2001: Honeypots started to be used for capturing malicious software from
internet and being aware of new threats. Companies began to use honeypots in
their systems to improve security and see the malicious traffic.
2002: Honeypot concept became popular and honeypots improved their
functionalities, so they became more useful and interesting for both researchers
and companies.
There are many security solutions available in the market. Anyone can browse
the variety of choices through internet and find the most suitable solution for their
HONEYPOT IN NETWORK SECURITY
needs. Here are the reasons why we should choose honeypots according to
Mokube I. and Adams M. (2007):
Honeypots can capture attacks and give information about the attack type and
if needed, thanks to the logs, it is possible to see additional information about the
attack.
New attacks can be seen and new security solutions can be created by looking
at them. More examinations can be obtained by looking at the type of the
malicious behaviors. It helps to understand more attacks that may happen.
Honeypots are not bulky in terms of capturing data. They are only dealing with
the incoming malicious traffic. Therefore, the information that has been caught is
not as much as the whole traffic. Focusing only on the malicious traffic makes the
investigation far easier. Therefore, this makes honeypots very useful.
For the only malicious traffic, there is no need for huge data storage. There is
no need for new technology to maintain. Any computer can be used as a
honeypot system. Thus, it does not cost additional budget to create such a system.
They are simple to understand, to configure and to install. They do not have
complex algorithms. There is no need for updating or changing some things.
As there are several important advantages of using honeypots, there are also some
disadvantages of them as well. We are continuing with Mokube I. & Adams M.
(2007)’s studies:
We can only capture data when the hacker is attacking the system actively. If
he does not attack the system, it is not possible to catch information. If there is an
attack occuring in another system, our honeypot will not be able to identify it. So,
attacks not towards our honeypot system may damage other systems and cause
big problems.
There is fingerprinting disadvantage of honeypots. It is easy for an
experienced hacker to understand if he is attacking a honeypot system or a real
system. Fingerprinting allows us to distinguish between these two. It is a not a
wanted result of our experiment.
The honeypot may be used as a zombie to reach other systems and compromise
them.
This can be very dangerous.
Let us have a look at each of them and compare them one by one according to
Mokube, I. & Adams M.(2007:p.322-323).
With low interaction honeypots , one can get the least amount of data
compared to other honeypot systems. They are limited, so the risk that was taken
from intruder is not big either proportionally.
First of all, there is no operating system to deal with. They can be used to
identify new worms or viruses and analyzing the traffic that is going on through
network. Low level of interaction honeypots are easy to configure and
understand.
In our thesis, firstly we will understand the logic of this category and test how
efficient they are. Therefore, we will start our experiment using the most common
low level of interaction honeypot, which is Honeyd. Its last version (1.5c) has
been released on 2007 and we will use it.
Details of Honeyd 1.5c will be explained under practical implementation
chapter.
Medium interaction honeypots are more advanced than low interaction
honeypots. Still, operating system does not exist.
But this time, more information and more complicated attacks from the
hacker can be obtained. As it is more advanced, it has more security holes so that
hacker can access the system. Mwcollect, honeytrap and Nepenthes are some of
the medium interaction honeypots that are used today.
High interaction honeypots are the most advanced honeypots. Unlike low
interaction and medium interaction honeypots, there is an operating system. As a
consequence, the hacker can perform anything. Proportionally, more data can be
captured from the hacker’s activities. However, it is the most risky one when it
comes to security as it provides such an access to the hacker that he does not have
any restrictions. These kind of honeypots are very time consuming and difficult
to maintain.
Honeywall is a good example of a high interaction honeypot. We will come
back to these security issues covering all these kind of honeypots and discuss and
state the exact security problems and come up with some ideas to improve
security thanks to our laboratory work.
In this part, we looked into a different kind of honeypot systems which are
wireless honeypots.
The goal of deploying wireless honeypots is to capture behaviors of our system in
a wireless area and obtain some information and statistics. IEEE 802.11
technology is covered, and also other technologies are possible such as bluetooth.
We used Maggi F. &Zanero S., (2008) and Siles R., (2007) ideas on this part.
This Wi-Fi structure can be obtained with some access points, wired network and
some open-to-attack computers.
Wi-Fi honeypots are used to capture unauthorized traffic, and tries to answer
questions if it is possible to catch wardriving and hackers which are trying to
compromise wireless networks.
HONEYPOT IN NETWORK SECURITY
First idea of wireless honeypots was released by Kevin Poulsen in 2002. During
his experiments, he realized that networks are not secure and protected.
Intruders are trying to monitor your system, eavesdropping, hacking your system
through your wireless network.
Therefore, The Wireless Information Security Experiment started the work in
2002 in Washington USA.
After that, the leader of this foundation Rob Lee continued the experiments and
tried to answer questions related to wireless hacking, and understand the hackers’
ideas and tools, especially the logic behind it.
Late 2002, Tenebris organization in Canada did the monitoring for malicious
activities, and understood that there was a huge malicious traffic going on
through the network. They did the experiment using wireless honeypot.
WMON is wireless monitor module. This module captures the traffic in order
to have the network traffic information. It helps to understand the attacks, so this
module is quite significant at this point.
While deploying and start using a honeypot, there are some legal issues that a
person should know about. Every country has different laws regarding to
honeypot usage and information capturing.
These regulations are related to data security, collection of data and finally how
to use honeypots.
All these different laws are based on the quality of the data that a honeypot can
capture and a person who is deploying it.
In here, the type of the data and its contents are significant. It is not easy to say
that if using honeypots are illegal or not.
As we stated before, it depends on the intention and the usage of the information
that has been collected. Therefore, there are several steps to think about before
doing this job.
There are also several questions and approaches that you should ask yourself
during the experiment.
If it is for a company that you are deploying a honeypot rather than a
homemade honeypot to use it at home, then as a network administrator you have
other responsibilities as well.
HONEYPOT IN NETWORK SECURITY
First thing is to think about the country laws regarding to these, and then
company laws. Maybe the country is allowing you to experiment some things but
what happens if there is a restriction on it in your company?
Before taking any serious action, you should ask those questions to the
responsible people to make sure that you are doing something without violating
the laws of your country.
We will look into three main legal issues now, which are privacy, entrapment
and civil liability.
3.1.1.1 Privacy
Let us start with privacy issue. As the type of data we are gathering is important,
privacy and data leads us to confidentiality term in network security.
Our example is being a network administrator in a company. Does he have a
right to collect information from other employees in the company?
Accordingly, it is the same logic with the hacker.
Does the hacker have a right to do so? If we combine both of these situations,
then we come up with these: Does honeypot have a right to collect information
from the hacker and his/her friends? Privacy is relative here.
As there are several levels of interactions honeypots, the information that is
gained is also relative.
3.1.1.2 Entrapment
3.1.1.3 Civilliability
It should be noted that there is nothing to do with federal or law in this issue.
When that kind of problem occurs, you should consult state which means you
should talk about this problem with legal counsel.
According to Lance Spitzner (2002) there are four different laws in United States.
They are U.S Constitution (Fourth Amendment), The Federal Wiretap Act (can
be referred as Title III), The Pen Register / Trap and Trace Statue (Pen/Trap), The
Electronic Communication Privacy Act (ECPA).
The last three ones are taking care of privacy of communication.
First of all, fourth amendment is meaningful with the following situations
according to Lance Spitzner (2002).
“Where the honeypot is run b y government, the search and seizure provisions
of the Fourth Amendment may apply.”
And the second situation is “where aprivate actor is assisting a government
actor, that private actor may be deemed an extension, or agent of the government
actor for Fourth Amendment purposes.”
Wiretap Act and Pen/Trap are working together to investigate the traffic that is
happening at that time.
They are obtaining the data from the source such as the traffic information, IP
address and packets that has been sent through the network, port numbers and so
on.
3.1.2.2 EU regulations
In EU regulations, it is not very easy to find laws that exist like US. LorfSteyn is
explaining with these words how it is working in United Kingdom according to
the case of R versus Latif.
HONEYPOT IN NETWORK SECURITY
“The court has the discretion: it has to perform a balancing exercise.....the judge
must weigh in the balance the public interest in insuring that those that are
charged of grave crimes should be tried and the competing public interest in not
conveying the impression that court will adopt the approach that the end justifies
the means.
”Lakhani A.D states that in European Union and United Kingdom, problems with
privacy are handled with two major laws.
They are Directive 97/66/EC and Regulation of Investigatory powers act, 2000.
However second law only exists in United Kingdom.
Directive 97/66/EC is a law which handles the confidentiality of communication
including the network and other telecommunications for public systems.
Regulation of Investigatory powers act, 2000 is the version of USA’s federal
wiretap statue. They both handle privacy of communication interceptions.
4. PRACTICAL IMPLEMENTATION
In this chapter we are presenting our practical work.
We are starting with low interaction honeypot and then continue on a middle
level of interaction to finally conclude with a high level of interaction.
Figure 4.1 Honeyd virtual honeypots from Virtual Honeypots: From Botnet Tracking
toIntrusion Detection, ProvosN., Holz T. (2007)
HONEYPOT IN NETWORK SECURITY
As it is seen processes are designed to create some returns which are created by
personality engine to make it look good and logical according to our template for
Honeyd.Figure 4.1 Honeyd structure from Virtual Honeypots: From Botnet Tracking to
IntrusionDetection, ProvosN., Holz T. (2007)
Figure 4.2 Honeyd virtual honeypots from Virtual Honeypots: From Botnet Tracking
toIntrusion Detection, ProvosN., Holz T. (2007)
The difference between low level and medium level of interaction honeypots is
medium level of interaction honeypots are not working on network stack and do
the management on it.
As we stated it above, Nepenthes is developed with Mwcollectd.
According to the researches of Baecher P., Koetter M., Holz T., Dornseif M.,
Freiling F., (2006), Nepenthes is giving us a great way of expressiveness.
We can also create some vulnerability modules in it. That is what makes
Nepenthes more interesting to work on. Thanks to this unique module, we can
define so many vulnerabilities. This key point is also not possible in high level of
interaction honeypots.
Thus, in our experiment, it was very good to see and discover different sides of
different level of interaction honeypots. Nepenthes has scalability and flexibility
features.
We can compare it with several honeypots. For example, we deployed
Honeyd as a low interaction honeypot before. And, Honeyd cannot virtualize
complicated protocols.
Even high interaction honeypot Gen III is not very good at scalability. There
are limitations of creating several honeypots.
However, we learned that Nepenthes is covering the blanks that several
honeypot do not have. We can create many honeypots in our system and capture
data easily.
scanner offering a lot of options (type of scan, level of detail about the target,
etc…). The result of our scan is the figure 4.5 below:
Many precious information have been gathered, thanks to the scan we could
identify the operating system running on the target (Windows XP sp2 or sp3) and
which port are open (135, 139 and 445).
We also obtained some information about the network card used, we used the
default value for the MAC address of the machine, so Nmap detected it as a
Vmware network card.
We could have changed this address in order to avoid the virtual machine
detection.
On the report page, a link is provided to install the right patch and protect the
system.
On the hacker point of view, it just highlights which exploits he can use.
After the scan, we obtained the following result shown in the figure 4.6:
Nessus detected two critical problems, by opening the pane to take a closer
look at it, we saw that the machine had the vulnerability ms08-067. Figure 4.7 is
showing this vulnerability.
First of all, honeypot forensics is used to study and understand a hacker strategy and
his tools but not to prosecute him.
This science is very time consuming and according to honeynet project members,
one hour of hacker activity can lead to more than 40h of forensic work.
The suggested approach is to work on a copy of the original victim, that way the
analysis process can be repeated from the beginning without losing any important
data.
Forensic in computer science require a perfect knowledge of hacker techniques as
well as how different software works in general.
Forensic science is to find evidences to make researches on it and trying to find
some details and answers from it.
The forensic science branch that we are interested in our thesis is computer
forensics which is the same definition of forensic science but this time electronic
devices are involved with our researches. The necessary data is obtained from the
devices, and forensic investigators make deeper examination on them.
There are several roles and responsibilities for forensic investigation.
Forensic investigation is done with first responders, investigators, technicians,
evidence custodians, forensic examiners and forensic analysts. (Kipper G., (2007)).
Figure 6.1 Sebek output example from Raynal F. &Berthier Y. &Biondi P. &Kaminsky
D., (2004).
We should try to understand what this output means, and evaluate the results.
First of all, any intruder looks for a target to hack and for this purpose he tries to
obtain necessary information to do it.
There are some commands to have information about the target such as ifconfig,
netstat and so on. After he knows the system, he starts to do this job.
According to his aim, he installs some software on the target. For example, if he
wishes to connect to the system several times, he installs a backdoor to achieve his goal.
Once he finishes those initial preparations for the target, he jumps to the next step.
Next step is to completely use the system the way he wants it. To be able to keep up the
things that hacker is doing to the system,
HONEYPOT IN NETWORK SECURITY
6. RESULTS
We studied all level of interaction honeypots and configured them. As first level of
interaction honeypot, we deployed Honeyd.
We explained the logic behind it and installed it correctly.
Our findings about Honeyd are; Honeyd is the most popular low interaction
honeypot but its problem is its age.
The project is opensource but part of it is outdated and nobody seems to upgrade it.
On the other hand hacker tools are evolving, so identifying this honeypot is not
hard.
Honeyd is using an old version on Nmap fingerprint to create fake virtual operating
systems so by using a newer version of Nmap, the fake operating systems will not
be recognized and Nmap will detect that there is a problem.
Another limitation of Honeyd is the scripts bound to the different ports.
With a basic scan it is possible to find which ports are open but as soon as the
attacker tries to actually connect on a port, he will realize the service is fake.
For example the script used for a Web server, by connecting it using telnet,
thew server should send back replies but nothing is happening. Another problem
is one cannot understand if there is an incoming attack to the system or not.
Because there is no such alarm system that can make you understand that there is
an attack. Information gathering is not very smart either. As a result the hacker
can understand quickly that there is something wrong with the target and will
abort his attack. Even unprofessional intruders can compromise the honeypot
without spending too much time on it. Because it is very popular and easy to use
well known techniques such as Nmap. There is no additional approach needed for
it.
The last step was the implementation of a high interaction honeypot: Honeywall.
We explained his structure and how to set it up. Its architecture is solid, and an
intruder will not understand at the first glance that he is entering a honeypot system.
Honeywall is making a bridge between the attacker and the honeypots interface
making the link invisible. All the traffic going through that link can be saved and
analyzed.
HONEYPOT IN NETWORK SECURITY
7. CONCLUSION
In our thesis, we explained honeypot systems in detail, and implemented low interaction,
middle interaction and high interaction honeypots at laboratory.
Our goal was to understand their strategy and how they are working in order to lure
intruders towards the system.
We discovered their security flaws in order to help researchers and organizations.
Several companies are using honeypot systems to protect the whole organization’s network
security, and researchers are making academic experiments on them at schools.
As we all know network security is very significant for all computer systems because any
unprotected machine in a network can be compromised in any minute.
One may lose all the secret and important data of a company, which can be a great loss,
and it is also very dangerous that someone else knows your important personal information.
Thus, we tried to find answers for honeypots’ security using all interaction honeypots
possible.
Our main goal for our thesis was to see if honeypots are easy to hack and check if
they are really isolated from other networks like a organization’s network.
When a honeypot is compromised, is it possible to reach other systems and
compromise them too ? After the system is compromised, is it possible to track the
hacker by using necessary forensic science tools ? How efficient are they ? As we stated
in results and analysis part,we easily hacked all the honeypots that we used for our thesis.
Especially, low interaction honeypot Honeyd can be hacked easily without too much
effort. As we stated before, any amateur hacker can seize the system and also can see that
it is a trap system.
Therefore, Honeyd is not a good honeypot as its features are not efficient to fool the
hacker. As Honeyd is a deamon, it is just simulating a operating system’s services.
So, it is not possible to a hacker to seize other systems using Honeyd. For the
intruder, it will not take time to see that the system is not real, so he will not continue
compromising it.
He will leave the system. For forensic part, Honeyd’s log was sufficient to see the
actions of the hacker.
Next part was to try Nepenthes as medium interaction honeypots.
The result was quite similar. Thus,we came up with this conclusion: Low interaction
honeypots and medium interaction honeypots are just simulating the services of a real
system, because of that it is not possible to capture significant data from intruders.
They are slightly different from each other but the main idea is the same. As they are
not real operating systems , it is not risky to build them.
There is no need to mention about further attacks.
So, we moved on to the last level. After working low interaction and medium
interaction honeypots, we decided to deploy high interaction honeypots.
We studied on Honeywall. Even though it is time consuming and difficult, we
managed to create a structure and worked on it. Our result were more interesting than
before.High interaction honeypots are not virtualizing the system.
They are real systems.So, it is very risky but the captured information is important.
After deploying the implementation correctly, we successfully hacked the honeynet, but
not Honeywall itself. It was the result we were looking for.
HONEYPOT IN NETWORK SECURITY
REFERENCES
Alata E. &Nicomette V. &Kaâniche M. &Dacier M. &Herrb M.,2006.Lessons learned
from the deployment of a high-interaction honeypot, EDCC'06.
Crandall J.R. &Wu S.F & Chong F.T., Experiences Using Minos as a Tool for Capturing
and Analyzing Novel Worms for Unknown Vulnerabilities University of California,
Davis
University of California, Santa Barbara [Online] Available
at :
http://www.cs.unm.edu/~crandall/DIMVAMinos.ppt [Accessed 3 May 2010]