SlideShare a Scribd company logo
1 of 68
INSECS - INTELLIGENT NETWORK
SECURITY SYSTEM
Thesis by: Nadun N Rajasinghe
Supervisors: Dr. Jagath Samarabandu and Dr. XianbinWang
Department of Electrical and Computer Engineering,
The University ofWestern Ontario,
London.
INTRODUCTIO
N
Security issues in Modern distributed networks
■ Insecure web interface - Using default password and having insecure password recovery method
■ Insufficient or non-existent authentication/authorization
■ Vulnerable network services – communication ports have known vulnerabilities they will be exploited by hackers
■ Lack of transport encryption
■ Insecure software/firmware - software needs to be verified by signatures
■ Poor physical security
■ Data Breaches and Loss
■ Account or Service Hijacking
■ Malicious Insiders
■ HypervisorVulnerabilities – This will cause intruder to gain access to all virtual machines
■ Denial of Service
Different defensive
strategies
Firewall
Encryption
Anti-Virus
Authentication
Intrusion
Detection/Prevention
What is an Intrusion Detection System
■ An Intrusion Detection System is a device or software that monitors a system for malicious
activity.
How does an Intrusion Detection System work?
■ There are two popular detection strategies.
■ Anomaly Detection ( Behavior based) – Model normal behavior.Anything deviating
from that is an intrusion.
■ Misuse Detection ( Signature based) – Identify details about attacks and create attack
signatures. If traffic match these, it is considered an intrusion.
■ IDS can be further classified based on the Scope of operation.
■ Network-based (NIDS) – Intrusion Detection which relies on network traffic
information.
■ The usual place to have these is inside the network after the firewall.
■ Sometime these are also placed in selected places inside the network.
■ Host-based (HIDS) - Intrusion Detection based on system activity in a host.
■ These are placed at each host.
How does an Intrusion Detection System work?
■ There are two popular detection strategies.
■ Anomaly Detection – Model normal behavior.Anything deviating from that is an
intrusion.
■ Misuse Detection – Identify details about attacks and create attack signatures. If
traffic match these, it is considered an intrusion.
■ IDS can be further classified based on the Scope of operation.
■ Network-based (NIDS) – Intrusion Detection which relies on network traffic
information.
■ The usual place to have these is inside the network with the firewall.
■ Sometime these are also placed in selected places inside the network.
■ Host-based (HIDS) - Intrusion Detection based on system activity in a host.
■ These are placed at each host.
Areas of Research Interest
■ Looked at 2 aspects of intrusion Detection
■ Problems with Intrusion Datasets
■ Intrusion Detection System for Modern Distributed networks including IoT and Cloud
EXISTING CHALLENGES
AND
CONTRIBUTIONS OFTHIS
RESEARCH
Challenges
in Intrusion
Datasets
■ IDS rely heavily on Datasets for training or learning. If the Dataset is faulty,
so is the IDS.
■ A standard dataset is created for conditions available at that time
■ TheVersions of software and services
■ The nature of attacks and attack tools
■ Dataset has attributes, the dataset creators thought were good
■ Users of the dataset has no choice on customizability
■ Once dataset is created and released it’s fixed
■ No Choice on the Output format of Dataset
o PCAP in most cases, sometimes CSV
■ No choice over the attributes
o Ex. NSL-KDD has 41 attributes that cannot be changed
Dataset
Creation
■ INSecS-DCS - an on-demand intrusion dataset creation software that can run
on a network of choice and gives the user the ability to make fully customized
datasets.
1. A conference Paper – INSecS-DCS:A highly customizable Intrusion dataset
creation framework (CCECE, Quebec City 2018)
2. Available for Public use under an MIT license for the advancement of Intrusion
Detection research.
Does not
work!
Challenges
of Intrusion
Detection in
Modern
Networks
■ TheWidely Used Intrusion Detection Systems do not support Distributed
Intrusion Detection.
■ HighTraffic overhead leads to Lower Intrusion Detection Speeds
■ Detecting Multi stage attacks
Traffic
Volume
Detection
Speed
Challenges
of Intrusion
Detection in
Modern
Networks
■ Intrusion Detection systems need Adaptive learning to save detection effort
■ Limited Computing power in IoT devices.
■ Many IoT devices have low computing power
■ InternalAttacks
■ not all IDS architectures are able to detect internal attack.
IDS
Network
Data
Rule updates
Intrusion
Detection
■ INSecS-IDS – a flexible Intrusion Detection System fit for modern networks
due to it’s design and architecture, which is aimed at supporting IoT, Cloud and
Distributed systems.
■ A journal Paper – ACMAn IDS with Hierarchical Decision Making within a
Framework for Intelligent Network Security System
■ A survey of Intrusion detection as applied to IoT, Cloud and other distributed
networks.
■ A Flexible IDS design that allows third party Applications.
■ A comparison of the performance of INSecS-IDS compared to Snort.
■ Possible example setups for INSecS-IDS in IoT and Cloud applications.
PROPOSED
SOLUTION
Intelligent Network
Security Systems
Intrusion Detection
Framework
(INSecS-IDF)
INSecS-DCS
Dataset
Creation
Software
INSecS-IDS
Intrusion
Detection
System
 The Learning module extracts the rules out
of the dataset to be used in the IDS.
Learning
Module
 In this research the learning modules
functionality is simulated human expertise
INSECS-DCS
INSecS-DCS
Dataset
Creation
Software
RELATEDWORK AND
BACKGROUND
INSECS-
DCS
Dataset Creator
■ ID2T:A DIY dataset creation toolkit for Intrusion Detection Systems [56]
■ They have addressed some of the problems like
 Ability to create Datasets locally, anytime
 Insert attacks of your choice
 Attribute selection
■ Still not addressed,
 Get processed Dataset (not just RAW PCAP)
 Make new attributes withTime properties
 Change attributes
Processed Dataset
Packet
Extract
Packet
information
Packet
Extract
Overall
Information
Packet
Packet
Raw Dataset
Processed Dataset
■ Even in the datasets that provide these kind of attributes, you cannot make new ones or
customize them.
Destination
IP
Source IP
Last 100
 There are some datasets like Kyoto and NSL-KDD that offer these kind of processed attributes
but you cannot
1. change things like the number “100”.
2. Cannot add anyTemporal property to it
IMPLEMENTATION
OF INSECS-DCS
INSECS-
DCS
Architecture
DCS
Architecture
TShark is a network protocol analyzer
• It lets you capture packet data from a live network
• Read packets from a previously saved capture file
• TShark's native capture file format is pcap format, which is also the
format used by tcpdump and various other packet analyzer
Packet
Capturing
Packet Pre-
Processing
Collecting individual
packet information
Collecting overall packet
information for time
window
Dividing traffic into
time windows
Raw Dataset Processed Dataset
Packet Pre-
processing
Packet
Capturing
Packet Pre-
Processing
Collecting individual
packet information
Collecting overall packet
information for time
window
Dividing traffic into
time windows
Raw Dataset Processed Dataset
• Convert the Captured packets into a format suitable for the algorithm to
use easily. In this case, a dictionary of key value pairs.
Collecting
Individual
packet
information
Packet
Capturing
Packet Pre-
Processing
Collecting individual
packet information
Collecting overall packet
information for time
window
Dividing traffic into
time windows
Raw Dataset Processed Dataset
• Information from individual packets are collected by selecting key value pairs of
interest.These include,
o Protocols used -TCP, UDP, IP, FTP, SMTP, SSH, SSL,ARP, DHCP,HTTP
o Source and destination information - IP address, port numbers
Dividing
traffic flow
into time
windows
+
collecting
overall
features
Packet
Capturing
Packet Pre-
Processing
Collecting individual
packet information
Collecting overall packet
information for time
window
Dividing traffic into
time windows
Raw Dataset Processed Dataset
• Select a time window and analyzing the traffic flow during that time. (
customizable time window)
• As opposed to getting information from just the individual packets here we get
information on overall traffic behavior during the time window and identify
common trends in traffic
Dividing
traffic flow
into time
windows
+
collecting
overall
features  Before INSecS-DCS, you had to do this manually to the entire dataset if you
wanted more attributes.
 This allows users to create attributes like the following listed below
Attribute Description
connection pairs The number of different source and destination pairs
num ports number of different port numbers used
src bytes the total amount of source traffic
tcp frame length the total amount of frame bytes forTCP traffic
udp length the total amount of UDP data
num ssl total number of packets containing SSL traffic
PCAP file CSV or txt
Packet
Capturing
Packet Pre-
Processing
Collecting individual
packet information
Collecting overall packet
information for time
window
Dividing traffic into
time windows
Raw Dataset Processed Dataset
Processed
and Raw
Datasets
RESULTS
INSECS-
DCS
INSecS-DCS
■ To show the capability of the dataset creator, I used the PCAP files from the ISCX dataset and made
a processed dataset out of it.
■ The ISCX dataset is updated once a year.Therefore, it is a somewhat updated dataset.
■ This experiment was only to show the capability of INSecS-DCS. I still maintain the argument that
the best dataset for you is the dataset you create for your network.
■ INSecS-DCS is provided for public use under an MIT license
■ Hosted on GitHub - https://github.com/nrajasin/Network-intrusion-dataset-creator
COMPARISON
INSECS-
DCS
Comparison
of DCSWith
ID2T toolkit
Capability INSecS-DCS ID2T
Ability to Label dataset Yes Yes
Open Source Yes Yes
Raw PCAP dataset Yes Yes
Has a GUI No Yes
Allows attack injection within the software No Yes
Ability to divide traffic into time window and get overall
traffic attributes
Yes No
Ability to select input method ( packets captured on a
network of choice or get a raw PCAP dataset from another
source )
Yes No
Processed dataset that can fed intoWEKA and other ML
tools directly
Yes No
Attribute selection for processed dataset Yes No
INSECS-IDS
 The Dataset creator is capable of making
custom datasets that fit the network.
 The IDS is capable of providing Intrusion
Detection in a distributed manner, handling
large volumes of data.
 Signature based IDS on detection
strategy, that works with a Rule set.
 It is a Network based IDS on detection
scope but has nodes at each host in
network.
 The Learning module extracts the rules out
of the dataset.
RELATEDWORK AND
BACKGROUND
INSECS-IDS
■ Snort:A free, widely used Intrusion Detection/ Intrusion Prevention System
(IDS/IPS). It is a rule-based IDS/IPS which has a large rule base.
■ Bro:An open-source, passive, Unix based network intrusion detection
system that works well with large traffic loads.A key advantage Bro has is
its deep analysis involving a broad range of protocols.
■ Suricta: Considered to be a close competitor to Snort, with an advantage
over Snort in detection speed. According to literature, Suricta supports
Snort rules which makes it a strong performer as Snort is the most widely
supported IDS.
■ OSSEC:An Open Source, host-based Intrusion Detection System that
works for most Operating Systems. It performs log analysis, file integrity
checking,Windows registry monitoring, etc, within the host and
communicates using encrypted channels with the OSSEC server for
intrusion detection.
Widely Used Intrusion Detection Systems
Does not directly support
Distributed setup
Does not directly support
Distributed setup
Speed comes at the cost
of resources and also
does not support
distributed setups
Host Based architecture
does not offer network
wide protection
Existing solutions
■ Nezarat [36], has proposed using game theory to model the scenario involving a detection
agent, reporting from different nodes in the distributed system, and an intruder.
■ An IDS capable of detecting DoS attacks for distributed client-server systems using separate
client detectors and server detectors has been proposed by Kshirsagar et al. [13].The method
uses a rule-based approach to detecting attacks but details of rule generation are not explained.
■ By running different instances of Snort in a distributed manner, at different nodes across the
network, researchers have controlled the load on each of them by careful resources usage
observation [64]. However, due to processing all the instances with multiple Snort instances, the
researchers were not able to complete the experiment because of resource insufficiency.
■ Noorman et al. [38] has proposed an embedded security system with a separate microprocessor
that relies on knowing each node in the IoT system.The isolation capabilities ensure that even if
one node is compromised, intruders will not be able to hack into the other nodes using the
compromised node.
■ Lee et al. [24] proposed a distributed IDS solution based on analyzing the energy consumption at
each node, to detect anomalies.
INSecS – Intrusion Detection System
These are the main features of the proposed IDS
■ Designed for Distributed operation
■ Less detection delay with Large data volumes
■ Less resource usage
■ Adaptive rule updating
■ Ability to detect multi-step attacks faster
■ Detect internal attacks
IMPLEMENTATION
OF INSECS-IDS
INSECS-IDS
Logical Architecture – Hierarchical Decision Making
Primary
Decision Node
Complex
Decision
Engine
Primary
Decision Node
Primary
Decision Node
Only traffic that is suspected of being part of
attack is sent for more processing
This is different from the traditional approach
because there all traffic is checked for all
attacks.
IDS controller
The IDS Controller is responsible for
initiating the Packet handler and the initial
Primary Decision Nodes.
The initial PDNs can be detectors for
reconnaissance attacks or the most
common attacks.
More can be activated based on the
network behavior.
Packet Handler
As soon as the packet handler is initiated,
it runs Tshark, which captures the stream
of packets going in and out of the node
and generates an object with all the
packet attributes and values for each and
every packet.
This object is then sent to a pre-
processing unit in order to generate a
Python dictionary, which is broadcasted to
all the active buffers.
There is a buffer for each active PDN and
the corresponding PDN reads the stream
of dictionaries continuously.
Data Buffers
These Units function as Queues for the
streams of packets into the Primary
decision node engine. These prevent the
different decision units interfering with
each other’s operations.
Primary Decision
Node -1
This unit is responsible for determining
whether a packet is a possible threat or
not. For example, in the Portscan PDN, all
packets with SYN flag set to ON, are
marked as possible attack packets.
This reduces the traffic load on the
Intrusion detection process because
packets which have no association with
the attack are not subjected to further rule
checks.
There is a PDN at each network node.
Since there are perform lightweight tasks,
they can even be placed in low resource
devices like IoT.
Primary Decision
Node - 2
The traffic suspected of belonging to an
attack, is then sent to the Complex
Decision Engine using HTTP.
This choice was made because it is a very
common protocol and can be used easily
for the purpose.
Complex Decision
Engine
The complex decision engine performs the
complex rule and condition checks to make
final decision on whether there is an intrusion
or not.
This unit is placed in a device with a lot of
system resources as it handles all the data
streams from all the PDNs in the network.
The complex processes include,
• Dividing traffic streams into time
windows
• Counting instances of Different
incidents
• Condition and rule checks
• Determining if there is an intrusion
Multi Stage attack
If a reconnaissance attack is detected, the IDS
assumes that another attack would follow.
The CDE would then initiate new PDNs to
detect possible attacks if they are not already
switched on.
IDS adaptive behavior
Once an attack is detected, the IDS gets
updated with all the details of the attack.
This includes information like,
• The Source IP
• The ports used
• Geo location if available
If an attack were to originate from the same
source, it can be detected immediately
because of this updating process.
Experiment- 1
■ An experimental setup was used to validate the proposed setup.
■ 3 PDNs were used. Portscan, Ipsweep, Smurf.
■ The CDE was setup on another Server in the same network.
■ The DARPA 98 dataset was used as the traffic source.
■ The test cases were as follows:
Test case 1:Test the validity of using hierarchical decision making for faster Intrusion
detection
Two setups with identical features were used.The only difference was, one setup had hierarchical
decision making. I selected 3 dataset files containing the smurf attack. Only one detector was activated.
Test case 2: Ensure the results of test 1 are repeatable for multiple attacks
Two similar setups but with 3 PDNs and one dataset file containing 3 attacks (Smurf, Portscan, IPsweep)
were used.
Experiment - 2
■ Another feature of INSecS-IDS is the flexible design.The functionality of the CDE can be achieved
using a Complex Event Processor.
■ A Complex Event Processor is a software that can process streams of realtime data.
■ The advantage of this is, users who are used to SQL like queries can easily use the IDS.
■ The disadvantage would be features like adaptive rule updating would not work unless the CEP is an
adaptive one.
■ The CEP picked for this experiment was an adaptive CEP designed and implemented by a fellow lab
member.
■ The same 2 test cases mentioned before were used to validate this claim.
Experiment - 3
■ To test the effectiveness of INSecS-IDS compared to an industry grade IDS, another experiment
was devised.
■ Snort, (a popular IDS described earlier) was used to detect the same 3 attacks from the same
dataset files as Experiment 1.
RESULTS
INSECS-IDS
Detection Results for INSecS-IDS
Week 1,Wednesday 55.6 5.090
Week3,Wednesday 37.9 6.837
Week 5, Monday 150.65 124.266
Experiment 1 –Test Case 1
Experiment 1 –Test Case 2
Smurf 124.621 14.202
IP sweep 124.616 14.202
Port scan 124.613 14.202
■ The same accuracy was observed in all test cases. Meaning that all the attack
instances listed by the dataset creators were detected by the IDS.
Week 1,Wednesday 55.2 5.474
Week3,Wednesday 37.9 7.240
Week 5, Monday 150.65 133.58
Experiment 2 –Test Case 1
Experiment 2–Test Case 2
Smurf 124.621 14.9
IP sweep 124.616 14.9
Port scan 124.613 14.9
■ The same accuracy was observed in all test cases. Meaning that all the attack
instances listed by the dataset creators were detected by the IDS.
Detection Results for INSecS-IDS
COMPARISON
INSECS-IDS
Comparing Snort with INSecS-IDS
■ Experiment 3 showed that Snort also detected the same attack instances identified by
INSecS.
■ There was no method to compare the detection time because snort did packet reading and
Intrusion Detection on the same computer.There was no way to get the level of distributed
functionality that INSecS has.
■ Snort has too many rules, with multiple rule bases. Sometimes there are overlapping rules for
the same function. In comparison INSecS-IDS rule base is well structured to avoid duplication.
CONCLUSIONS
Conclusions for DCS and IDS
■ INSecS-DCS can be used for making customizable datasets for any network.
■ INSecS-IDS performs accurate Intrusion Detection in a distributed environment with low system
resources.
■ Hierarchical decision making provides a clear advantage in detection time and large traffic
volume handling.
Example IoT setup
IoT
Primary
Decision
Node
Mobile device
Primary
Decision
Node
Server
Primary
Decision
Node
IDS
controller
Complex
Decision
Engine
The CDE and IDS controller can be
in device with considerable
computing resources.
The IoT devices can even be in
different networks.
Each IoT device can have PDNs
activated.
Example Cloud setup
Server
Primary
Decision
Node
IDS
controller
Complex
Decision
Engine
Server
Primary
Decision
Node
Server
Primary
Decision
Node
The servers can be hosted in
different networks. Still the
proposed IDS can be deployed.
Each Server has PDNs activated in
them. Any suspicious traffic is
forwarded to the CDE.
CDE can be hosted on a device
with considerable computing
capabilities.
FUTUREWORK
Improvements for DCS and IDS
■ The communication between the distributed nodes is done using HTTP. I plan to use
encrypted communication to prevent the IDS itself being compromised.
■ The DatasetCreator and the IDS needs to be connected to provide a better IDS
solution. Right now, a human expert has to go through the Datasets and make the
rules. By implementing the learning module, I plan to make that an automated
process.
■ The learning module will generate rules for the specific network, based on the
Datasets created by the INSecS-DatasetCreation Software.
THANKYOU!

More Related Content

What's hot

Optimized Intrusion Detection System using Deep Learning Algorithm
Optimized Intrusion Detection System using Deep Learning AlgorithmOptimized Intrusion Detection System using Deep Learning Algorithm
Optimized Intrusion Detection System using Deep Learning Algorithmijtsrd
 
A Study of Intrusion Detection and Prevention System for Network Security
A Study of Intrusion Detection and Prevention System for Network SecurityA Study of Intrusion Detection and Prevention System for Network Security
A Study of Intrusion Detection and Prevention System for Network SecurityIRJET Journal
 
Network and network security
Network and network securityNetwork and network security
Network and network securityRuchi Gupta
 
Network forensics
Network forensicsNetwork forensics
Network forensicsArthyR3
 
INTERNET SECURITY SYSTEM
INTERNET SECURITY SYSTEMINTERNET SECURITY SYSTEM
INTERNET SECURITY SYSTEMBhushan Gajare
 
Data Mining and Intrusion Detection
Data Mining and Intrusion Detection Data Mining and Intrusion Detection
Data Mining and Intrusion Detection amiable_indian
 
A Study on Data Mining Based Intrusion Detection System
A Study on Data Mining Based Intrusion Detection SystemA Study on Data Mining Based Intrusion Detection System
A Study on Data Mining Based Intrusion Detection SystemAM Publications
 
Deep learning approach for network intrusion detection system
Deep learning approach for network intrusion detection systemDeep learning approach for network intrusion detection system
Deep learning approach for network intrusion detection systemAvinash Kumar
 
INTRUSION DETECTION SYSTEM CLASSIFICATION USING DIFFERENT MACHINE LEARNING AL...
INTRUSION DETECTION SYSTEM CLASSIFICATION USING DIFFERENT MACHINE LEARNING AL...INTRUSION DETECTION SYSTEM CLASSIFICATION USING DIFFERENT MACHINE LEARNING AL...
INTRUSION DETECTION SYSTEM CLASSIFICATION USING DIFFERENT MACHINE LEARNING AL...ijcsit
 
Network Intrusion Prevention by Configuring ACLs on the Routers, based on Sno...
Network Intrusion Prevention by Configuring ACLs on the Routers, based on Sno...Network Intrusion Prevention by Configuring ACLs on the Routers, based on Sno...
Network Intrusion Prevention by Configuring ACLs on the Routers, based on Sno...Disha Bedi
 
Intrusion Detection System using AI and Machine Learning Algorithm
Intrusion Detection System using AI and Machine Learning AlgorithmIntrusion Detection System using AI and Machine Learning Algorithm
Intrusion Detection System using AI and Machine Learning AlgorithmIRJET Journal
 
Intrusion detection using data mining
Intrusion detection using data miningIntrusion detection using data mining
Intrusion detection using data miningbalbeerrawat
 
IDS Evasion Techniques
IDS Evasion TechniquesIDS Evasion Techniques
IDS Evasion TechniquesTudor Damian
 
Introduction to IDS & IPS - Part 1
Introduction to IDS & IPS - Part 1Introduction to IDS & IPS - Part 1
Introduction to IDS & IPS - Part 1whitehat 'People'
 
Intrusion Detection Techniques for Mobile Wireless Networks
Intrusion Detection Techniques for Mobile Wireless NetworksIntrusion Detection Techniques for Mobile Wireless Networks
Intrusion Detection Techniques for Mobile Wireless Networksguest1b5f71
 
Network forensics and investigating logs
Network forensics and investigating logsNetwork forensics and investigating logs
Network forensics and investigating logsanilinvns
 

What's hot (20)

Optimized Intrusion Detection System using Deep Learning Algorithm
Optimized Intrusion Detection System using Deep Learning AlgorithmOptimized Intrusion Detection System using Deep Learning Algorithm
Optimized Intrusion Detection System using Deep Learning Algorithm
 
A Study of Intrusion Detection and Prevention System for Network Security
A Study of Intrusion Detection and Prevention System for Network SecurityA Study of Intrusion Detection and Prevention System for Network Security
A Study of Intrusion Detection and Prevention System for Network Security
 
Network and network security
Network and network securityNetwork and network security
Network and network security
 
Snort IPS
Snort IPSSnort IPS
Snort IPS
 
Network forensics
Network forensicsNetwork forensics
Network forensics
 
INTERNET SECURITY SYSTEM
INTERNET SECURITY SYSTEMINTERNET SECURITY SYSTEM
INTERNET SECURITY SYSTEM
 
Data Mining and Intrusion Detection
Data Mining and Intrusion Detection Data Mining and Intrusion Detection
Data Mining and Intrusion Detection
 
A Study on Data Mining Based Intrusion Detection System
A Study on Data Mining Based Intrusion Detection SystemA Study on Data Mining Based Intrusion Detection System
A Study on Data Mining Based Intrusion Detection System
 
Deep learning approach for network intrusion detection system
Deep learning approach for network intrusion detection systemDeep learning approach for network intrusion detection system
Deep learning approach for network intrusion detection system
 
INTRUSION DETECTION SYSTEM CLASSIFICATION USING DIFFERENT MACHINE LEARNING AL...
INTRUSION DETECTION SYSTEM CLASSIFICATION USING DIFFERENT MACHINE LEARNING AL...INTRUSION DETECTION SYSTEM CLASSIFICATION USING DIFFERENT MACHINE LEARNING AL...
INTRUSION DETECTION SYSTEM CLASSIFICATION USING DIFFERENT MACHINE LEARNING AL...
 
Network Intrusion Prevention by Configuring ACLs on the Routers, based on Sno...
Network Intrusion Prevention by Configuring ACLs on the Routers, based on Sno...Network Intrusion Prevention by Configuring ACLs on the Routers, based on Sno...
Network Intrusion Prevention by Configuring ACLs on the Routers, based on Sno...
 
Intrusion Detection System using AI and Machine Learning Algorithm
Intrusion Detection System using AI and Machine Learning AlgorithmIntrusion Detection System using AI and Machine Learning Algorithm
Intrusion Detection System using AI and Machine Learning Algorithm
 
Network forensic
Network forensicNetwork forensic
Network forensic
 
Intrusion detection using data mining
Intrusion detection using data miningIntrusion detection using data mining
Intrusion detection using data mining
 
Network traffic analysis course
Network traffic analysis courseNetwork traffic analysis course
Network traffic analysis course
 
IDS Evasion Techniques
IDS Evasion TechniquesIDS Evasion Techniques
IDS Evasion Techniques
 
Introduction to IDS & IPS - Part 1
Introduction to IDS & IPS - Part 1Introduction to IDS & IPS - Part 1
Introduction to IDS & IPS - Part 1
 
Network Forensic
Network ForensicNetwork Forensic
Network Forensic
 
Intrusion Detection Techniques for Mobile Wireless Networks
Intrusion Detection Techniques for Mobile Wireless NetworksIntrusion Detection Techniques for Mobile Wireless Networks
Intrusion Detection Techniques for Mobile Wireless Networks
 
Network forensics and investigating logs
Network forensics and investigating logsNetwork forensics and investigating logs
Network forensics and investigating logs
 

Similar to INSECS: Intelligent networks security system

Intrusion detection and prevention system
Intrusion detection and prevention systemIntrusion detection and prevention system
Intrusion detection and prevention systemNikhil Raj
 
MMIX Peering Forum and MMNOG 2020: Packet Analysis for Network Security
MMIX Peering Forum and MMNOG 2020: Packet Analysis for Network SecurityMMIX Peering Forum and MMNOG 2020: Packet Analysis for Network Security
MMIX Peering Forum and MMNOG 2020: Packet Analysis for Network SecurityAPNIC
 
FALLSEM2023-24_BCSE353E_ETH_VL2023240100871_2023-05-25_Reference-Material-I.ppt
FALLSEM2023-24_BCSE353E_ETH_VL2023240100871_2023-05-25_Reference-Material-I.pptFALLSEM2023-24_BCSE353E_ETH_VL2023240100871_2023-05-25_Reference-Material-I.ppt
FALLSEM2023-24_BCSE353E_ETH_VL2023240100871_2023-05-25_Reference-Material-I.pptuseonlyfortech140
 
Network Analysis Mini Project 2.pdf
Network Analysis Mini Project 2.pdfNetwork Analysis Mini Project 2.pdf
Network Analysis Mini Project 2.pdftalkaton
 
Network Analysis Mini Project 2.pptx
Network Analysis Mini Project 2.pptxNetwork Analysis Mini Project 2.pptx
Network Analysis Mini Project 2.pptxtalkaton
 
EFFICACY OF ATTACK DETECTION CAPABILITY OF IDPS BASED ON ITS DEPLOYMENT IN WI...
EFFICACY OF ATTACK DETECTION CAPABILITY OF IDPS BASED ON ITS DEPLOYMENT IN WI...EFFICACY OF ATTACK DETECTION CAPABILITY OF IDPS BASED ON ITS DEPLOYMENT IN WI...
EFFICACY OF ATTACK DETECTION CAPABILITY OF IDPS BASED ON ITS DEPLOYMENT IN WI...IJNSA Journal
 
Intrusion detection and prevention
Intrusion detection and preventionIntrusion detection and prevention
Intrusion detection and preventionNicholas Davis
 
Intrusion Detection And Prevention
Intrusion Detection And PreventionIntrusion Detection And Prevention
Intrusion Detection And PreventionNicholas Davis
 
Forensic Analysis - Empower Tech Days 2013
Forensic Analysis - Empower Tech Days 2013Forensic Analysis - Empower Tech Days 2013
Forensic Analysis - Empower Tech Days 2013Islam Azeddine Mennouchi
 
VTU 8TH SEM INFORMATION AND NETWORK SECURITY SOLVED PAPERS
VTU 8TH SEM INFORMATION AND NETWORK SECURITY SOLVED PAPERSVTU 8TH SEM INFORMATION AND NETWORK SECURITY SOLVED PAPERS
VTU 8TH SEM INFORMATION AND NETWORK SECURITY SOLVED PAPERSvtunotesbysree
 
All Hope is Not Lost Network Forensics Exposes Today's Advanced Security Thr...
All Hope is Not LostNetwork Forensics Exposes Today's Advanced Security Thr...All Hope is Not LostNetwork Forensics Exposes Today's Advanced Security Thr...
All Hope is Not Lost Network Forensics Exposes Today's Advanced Security Thr...Savvius, Inc
 
Intrusion_Detection_By_loay_elbasyouni
Intrusion_Detection_By_loay_elbasyouniIntrusion_Detection_By_loay_elbasyouni
Intrusion_Detection_By_loay_elbasyouniLoay Elbasyouni
 
network-management Web base.ppt
network-management Web base.pptnetwork-management Web base.ppt
network-management Web base.pptAssadLeo1
 
Cours_4_IDS_IPS.pptx
Cours_4_IDS_IPS.pptxCours_4_IDS_IPS.pptx
Cours_4_IDS_IPS.pptxssuserc517ee1
 
Intrusion detection
Intrusion detectionIntrusion detection
Intrusion detectionProgrammer
 
Rethinking Security: Corsa Red Armor Network Security Enforcement
Rethinking Security: Corsa Red Armor Network Security EnforcementRethinking Security: Corsa Red Armor Network Security Enforcement
Rethinking Security: Corsa Red Armor Network Security EnforcementCorsa Technology
 
Science DMZ security
Science DMZ securityScience DMZ security
Science DMZ securityJisc
 

Similar to INSECS: Intelligent networks security system (20)

012
012012
012
 
Intrusion detection and prevention system
Intrusion detection and prevention systemIntrusion detection and prevention system
Intrusion detection and prevention system
 
Day4
Day4Day4
Day4
 
MMIX Peering Forum and MMNOG 2020: Packet Analysis for Network Security
MMIX Peering Forum and MMNOG 2020: Packet Analysis for Network SecurityMMIX Peering Forum and MMNOG 2020: Packet Analysis for Network Security
MMIX Peering Forum and MMNOG 2020: Packet Analysis for Network Security
 
FALLSEM2023-24_BCSE353E_ETH_VL2023240100871_2023-05-25_Reference-Material-I.ppt
FALLSEM2023-24_BCSE353E_ETH_VL2023240100871_2023-05-25_Reference-Material-I.pptFALLSEM2023-24_BCSE353E_ETH_VL2023240100871_2023-05-25_Reference-Material-I.ppt
FALLSEM2023-24_BCSE353E_ETH_VL2023240100871_2023-05-25_Reference-Material-I.ppt
 
Network Analysis Mini Project 2.pdf
Network Analysis Mini Project 2.pdfNetwork Analysis Mini Project 2.pdf
Network Analysis Mini Project 2.pdf
 
Network Analysis Mini Project 2.pptx
Network Analysis Mini Project 2.pptxNetwork Analysis Mini Project 2.pptx
Network Analysis Mini Project 2.pptx
 
EFFICACY OF ATTACK DETECTION CAPABILITY OF IDPS BASED ON ITS DEPLOYMENT IN WI...
EFFICACY OF ATTACK DETECTION CAPABILITY OF IDPS BASED ON ITS DEPLOYMENT IN WI...EFFICACY OF ATTACK DETECTION CAPABILITY OF IDPS BASED ON ITS DEPLOYMENT IN WI...
EFFICACY OF ATTACK DETECTION CAPABILITY OF IDPS BASED ON ITS DEPLOYMENT IN WI...
 
Intrusion detection and prevention
Intrusion detection and preventionIntrusion detection and prevention
Intrusion detection and prevention
 
Intrusion Detection And Prevention
Intrusion Detection And PreventionIntrusion Detection And Prevention
Intrusion Detection And Prevention
 
Forensic Analysis - Empower Tech Days 2013
Forensic Analysis - Empower Tech Days 2013Forensic Analysis - Empower Tech Days 2013
Forensic Analysis - Empower Tech Days 2013
 
VTU 8TH SEM INFORMATION AND NETWORK SECURITY SOLVED PAPERS
VTU 8TH SEM INFORMATION AND NETWORK SECURITY SOLVED PAPERSVTU 8TH SEM INFORMATION AND NETWORK SECURITY SOLVED PAPERS
VTU 8TH SEM INFORMATION AND NETWORK SECURITY SOLVED PAPERS
 
All Hope is Not Lost Network Forensics Exposes Today's Advanced Security Thr...
All Hope is Not LostNetwork Forensics Exposes Today's Advanced Security Thr...All Hope is Not LostNetwork Forensics Exposes Today's Advanced Security Thr...
All Hope is Not Lost Network Forensics Exposes Today's Advanced Security Thr...
 
Intrusion_Detection_By_loay_elbasyouni
Intrusion_Detection_By_loay_elbasyouniIntrusion_Detection_By_loay_elbasyouni
Intrusion_Detection_By_loay_elbasyouni
 
network-management Web base.ppt
network-management Web base.pptnetwork-management Web base.ppt
network-management Web base.ppt
 
Cours_4_IDS_IPS.pptx
Cours_4_IDS_IPS.pptxCours_4_IDS_IPS.pptx
Cours_4_IDS_IPS.pptx
 
IS-Types of IDPSs.pptx
IS-Types of IDPSs.pptxIS-Types of IDPSs.pptx
IS-Types of IDPSs.pptx
 
Intrusion detection
Intrusion detectionIntrusion detection
Intrusion detection
 
Rethinking Security: Corsa Red Armor Network Security Enforcement
Rethinking Security: Corsa Red Armor Network Security EnforcementRethinking Security: Corsa Red Armor Network Security Enforcement
Rethinking Security: Corsa Red Armor Network Security Enforcement
 
Science DMZ security
Science DMZ securityScience DMZ security
Science DMZ security
 

Recently uploaded

MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLSMANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLSSIVASHANKAR N
 
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Dr.Costas Sachpazis
 
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur EscortsHigh Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escortsranjana rawat
 
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...ranjana rawat
 
Extrusion Processes and Their Limitations
Extrusion Processes and Their LimitationsExtrusion Processes and Their Limitations
Extrusion Processes and Their Limitations120cr0395
 
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130Suhani Kapoor
 
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...ranjana rawat
 
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx  fake news detection using machine learningchaitra-1.pptx  fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learningmisbanausheenparvam
 
(PRIYA) Rajgurunagar Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(PRIYA) Rajgurunagar Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(PRIYA) Rajgurunagar Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(PRIYA) Rajgurunagar Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...ranjana rawat
 
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...Dr.Costas Sachpazis
 
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...Christo Ananth
 
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptxthe ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptxhumanexperienceaaa
 
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)Suman Mia
 
GDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSCAESB
 
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINEMANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINESIVASHANKAR N
 
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxWhat are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxwendy cai
 
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...Soham Mondal
 
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )Tsuyoshi Horigome
 

Recently uploaded (20)

MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLSMANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
 
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
 
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur EscortsHigh Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
 
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
 
Extrusion Processes and Their Limitations
Extrusion Processes and Their LimitationsExtrusion Processes and Their Limitations
Extrusion Processes and Their Limitations
 
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
 
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
 
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx  fake news detection using machine learningchaitra-1.pptx  fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learning
 
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptxExploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
 
(PRIYA) Rajgurunagar Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(PRIYA) Rajgurunagar Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(PRIYA) Rajgurunagar Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(PRIYA) Rajgurunagar Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
 
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
 
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
 
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptxthe ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
 
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
 
DJARUM4D - SLOT GACOR ONLINE | SLOT DEMO ONLINE
DJARUM4D - SLOT GACOR ONLINE | SLOT DEMO ONLINEDJARUM4D - SLOT GACOR ONLINE | SLOT DEMO ONLINE
DJARUM4D - SLOT GACOR ONLINE | SLOT DEMO ONLINE
 
GDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentation
 
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINEMANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
 
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxWhat are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptx
 
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
 
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )
 

INSECS: Intelligent networks security system

  • 1. INSECS - INTELLIGENT NETWORK SECURITY SYSTEM Thesis by: Nadun N Rajasinghe Supervisors: Dr. Jagath Samarabandu and Dr. XianbinWang Department of Electrical and Computer Engineering, The University ofWestern Ontario, London.
  • 3. Security issues in Modern distributed networks ■ Insecure web interface - Using default password and having insecure password recovery method ■ Insufficient or non-existent authentication/authorization ■ Vulnerable network services – communication ports have known vulnerabilities they will be exploited by hackers ■ Lack of transport encryption ■ Insecure software/firmware - software needs to be verified by signatures ■ Poor physical security ■ Data Breaches and Loss ■ Account or Service Hijacking ■ Malicious Insiders ■ HypervisorVulnerabilities – This will cause intruder to gain access to all virtual machines ■ Denial of Service
  • 5. What is an Intrusion Detection System ■ An Intrusion Detection System is a device or software that monitors a system for malicious activity.
  • 6. How does an Intrusion Detection System work? ■ There are two popular detection strategies. ■ Anomaly Detection ( Behavior based) – Model normal behavior.Anything deviating from that is an intrusion. ■ Misuse Detection ( Signature based) – Identify details about attacks and create attack signatures. If traffic match these, it is considered an intrusion. ■ IDS can be further classified based on the Scope of operation. ■ Network-based (NIDS) – Intrusion Detection which relies on network traffic information. ■ The usual place to have these is inside the network after the firewall. ■ Sometime these are also placed in selected places inside the network. ■ Host-based (HIDS) - Intrusion Detection based on system activity in a host. ■ These are placed at each host.
  • 7. How does an Intrusion Detection System work? ■ There are two popular detection strategies. ■ Anomaly Detection – Model normal behavior.Anything deviating from that is an intrusion. ■ Misuse Detection – Identify details about attacks and create attack signatures. If traffic match these, it is considered an intrusion. ■ IDS can be further classified based on the Scope of operation. ■ Network-based (NIDS) – Intrusion Detection which relies on network traffic information. ■ The usual place to have these is inside the network with the firewall. ■ Sometime these are also placed in selected places inside the network. ■ Host-based (HIDS) - Intrusion Detection based on system activity in a host. ■ These are placed at each host.
  • 8. Areas of Research Interest ■ Looked at 2 aspects of intrusion Detection ■ Problems with Intrusion Datasets ■ Intrusion Detection System for Modern Distributed networks including IoT and Cloud
  • 10. Challenges in Intrusion Datasets ■ IDS rely heavily on Datasets for training or learning. If the Dataset is faulty, so is the IDS. ■ A standard dataset is created for conditions available at that time ■ TheVersions of software and services ■ The nature of attacks and attack tools ■ Dataset has attributes, the dataset creators thought were good ■ Users of the dataset has no choice on customizability ■ Once dataset is created and released it’s fixed ■ No Choice on the Output format of Dataset o PCAP in most cases, sometimes CSV ■ No choice over the attributes o Ex. NSL-KDD has 41 attributes that cannot be changed
  • 11. Dataset Creation ■ INSecS-DCS - an on-demand intrusion dataset creation software that can run on a network of choice and gives the user the ability to make fully customized datasets. 1. A conference Paper – INSecS-DCS:A highly customizable Intrusion dataset creation framework (CCECE, Quebec City 2018) 2. Available for Public use under an MIT license for the advancement of Intrusion Detection research. Does not work!
  • 12. Challenges of Intrusion Detection in Modern Networks ■ TheWidely Used Intrusion Detection Systems do not support Distributed Intrusion Detection. ■ HighTraffic overhead leads to Lower Intrusion Detection Speeds ■ Detecting Multi stage attacks Traffic Volume Detection Speed
  • 13. Challenges of Intrusion Detection in Modern Networks ■ Intrusion Detection systems need Adaptive learning to save detection effort ■ Limited Computing power in IoT devices. ■ Many IoT devices have low computing power ■ InternalAttacks ■ not all IDS architectures are able to detect internal attack. IDS Network Data Rule updates
  • 14. Intrusion Detection ■ INSecS-IDS – a flexible Intrusion Detection System fit for modern networks due to it’s design and architecture, which is aimed at supporting IoT, Cloud and Distributed systems. ■ A journal Paper – ACMAn IDS with Hierarchical Decision Making within a Framework for Intelligent Network Security System ■ A survey of Intrusion detection as applied to IoT, Cloud and other distributed networks. ■ A Flexible IDS design that allows third party Applications. ■ A comparison of the performance of INSecS-IDS compared to Snort. ■ Possible example setups for INSecS-IDS in IoT and Cloud applications.
  • 16. Intelligent Network Security Systems Intrusion Detection Framework (INSecS-IDF)
  • 19.  The Learning module extracts the rules out of the dataset to be used in the IDS. Learning Module  In this research the learning modules functionality is simulated human expertise
  • 23. Dataset Creator ■ ID2T:A DIY dataset creation toolkit for Intrusion Detection Systems [56] ■ They have addressed some of the problems like  Ability to create Datasets locally, anytime  Insert attacks of your choice  Attribute selection ■ Still not addressed,  Get processed Dataset (not just RAW PCAP)  Make new attributes withTime properties  Change attributes
  • 25. Processed Dataset ■ Even in the datasets that provide these kind of attributes, you cannot make new ones or customize them. Destination IP Source IP Last 100  There are some datasets like Kyoto and NSL-KDD that offer these kind of processed attributes but you cannot 1. change things like the number “100”. 2. Cannot add anyTemporal property to it
  • 28. DCS Architecture TShark is a network protocol analyzer • It lets you capture packet data from a live network • Read packets from a previously saved capture file • TShark's native capture file format is pcap format, which is also the format used by tcpdump and various other packet analyzer Packet Capturing Packet Pre- Processing Collecting individual packet information Collecting overall packet information for time window Dividing traffic into time windows Raw Dataset Processed Dataset
  • 29. Packet Pre- processing Packet Capturing Packet Pre- Processing Collecting individual packet information Collecting overall packet information for time window Dividing traffic into time windows Raw Dataset Processed Dataset • Convert the Captured packets into a format suitable for the algorithm to use easily. In this case, a dictionary of key value pairs.
  • 30. Collecting Individual packet information Packet Capturing Packet Pre- Processing Collecting individual packet information Collecting overall packet information for time window Dividing traffic into time windows Raw Dataset Processed Dataset • Information from individual packets are collected by selecting key value pairs of interest.These include, o Protocols used -TCP, UDP, IP, FTP, SMTP, SSH, SSL,ARP, DHCP,HTTP o Source and destination information - IP address, port numbers
  • 31. Dividing traffic flow into time windows + collecting overall features Packet Capturing Packet Pre- Processing Collecting individual packet information Collecting overall packet information for time window Dividing traffic into time windows Raw Dataset Processed Dataset • Select a time window and analyzing the traffic flow during that time. ( customizable time window) • As opposed to getting information from just the individual packets here we get information on overall traffic behavior during the time window and identify common trends in traffic
  • 32. Dividing traffic flow into time windows + collecting overall features  Before INSecS-DCS, you had to do this manually to the entire dataset if you wanted more attributes.  This allows users to create attributes like the following listed below Attribute Description connection pairs The number of different source and destination pairs num ports number of different port numbers used src bytes the total amount of source traffic tcp frame length the total amount of frame bytes forTCP traffic udp length the total amount of UDP data num ssl total number of packets containing SSL traffic
  • 33. PCAP file CSV or txt Packet Capturing Packet Pre- Processing Collecting individual packet information Collecting overall packet information for time window Dividing traffic into time windows Raw Dataset Processed Dataset Processed and Raw Datasets
  • 35. INSecS-DCS ■ To show the capability of the dataset creator, I used the PCAP files from the ISCX dataset and made a processed dataset out of it. ■ The ISCX dataset is updated once a year.Therefore, it is a somewhat updated dataset. ■ This experiment was only to show the capability of INSecS-DCS. I still maintain the argument that the best dataset for you is the dataset you create for your network. ■ INSecS-DCS is provided for public use under an MIT license ■ Hosted on GitHub - https://github.com/nrajasin/Network-intrusion-dataset-creator
  • 37. Comparison of DCSWith ID2T toolkit Capability INSecS-DCS ID2T Ability to Label dataset Yes Yes Open Source Yes Yes Raw PCAP dataset Yes Yes Has a GUI No Yes Allows attack injection within the software No Yes Ability to divide traffic into time window and get overall traffic attributes Yes No Ability to select input method ( packets captured on a network of choice or get a raw PCAP dataset from another source ) Yes No Processed dataset that can fed intoWEKA and other ML tools directly Yes No Attribute selection for processed dataset Yes No
  • 39.  The Dataset creator is capable of making custom datasets that fit the network.  The IDS is capable of providing Intrusion Detection in a distributed manner, handling large volumes of data.  Signature based IDS on detection strategy, that works with a Rule set.  It is a Network based IDS on detection scope but has nodes at each host in network.  The Learning module extracts the rules out of the dataset.
  • 41. ■ Snort:A free, widely used Intrusion Detection/ Intrusion Prevention System (IDS/IPS). It is a rule-based IDS/IPS which has a large rule base. ■ Bro:An open-source, passive, Unix based network intrusion detection system that works well with large traffic loads.A key advantage Bro has is its deep analysis involving a broad range of protocols. ■ Suricta: Considered to be a close competitor to Snort, with an advantage over Snort in detection speed. According to literature, Suricta supports Snort rules which makes it a strong performer as Snort is the most widely supported IDS. ■ OSSEC:An Open Source, host-based Intrusion Detection System that works for most Operating Systems. It performs log analysis, file integrity checking,Windows registry monitoring, etc, within the host and communicates using encrypted channels with the OSSEC server for intrusion detection. Widely Used Intrusion Detection Systems Does not directly support Distributed setup Does not directly support Distributed setup Speed comes at the cost of resources and also does not support distributed setups Host Based architecture does not offer network wide protection
  • 42. Existing solutions ■ Nezarat [36], has proposed using game theory to model the scenario involving a detection agent, reporting from different nodes in the distributed system, and an intruder. ■ An IDS capable of detecting DoS attacks for distributed client-server systems using separate client detectors and server detectors has been proposed by Kshirsagar et al. [13].The method uses a rule-based approach to detecting attacks but details of rule generation are not explained. ■ By running different instances of Snort in a distributed manner, at different nodes across the network, researchers have controlled the load on each of them by careful resources usage observation [64]. However, due to processing all the instances with multiple Snort instances, the researchers were not able to complete the experiment because of resource insufficiency. ■ Noorman et al. [38] has proposed an embedded security system with a separate microprocessor that relies on knowing each node in the IoT system.The isolation capabilities ensure that even if one node is compromised, intruders will not be able to hack into the other nodes using the compromised node. ■ Lee et al. [24] proposed a distributed IDS solution based on analyzing the energy consumption at each node, to detect anomalies.
  • 43. INSecS – Intrusion Detection System These are the main features of the proposed IDS ■ Designed for Distributed operation ■ Less detection delay with Large data volumes ■ Less resource usage ■ Adaptive rule updating ■ Ability to detect multi-step attacks faster ■ Detect internal attacks
  • 45. Logical Architecture – Hierarchical Decision Making Primary Decision Node Complex Decision Engine Primary Decision Node Primary Decision Node Only traffic that is suspected of being part of attack is sent for more processing This is different from the traditional approach because there all traffic is checked for all attacks.
  • 46. IDS controller The IDS Controller is responsible for initiating the Packet handler and the initial Primary Decision Nodes. The initial PDNs can be detectors for reconnaissance attacks or the most common attacks. More can be activated based on the network behavior.
  • 47. Packet Handler As soon as the packet handler is initiated, it runs Tshark, which captures the stream of packets going in and out of the node and generates an object with all the packet attributes and values for each and every packet. This object is then sent to a pre- processing unit in order to generate a Python dictionary, which is broadcasted to all the active buffers. There is a buffer for each active PDN and the corresponding PDN reads the stream of dictionaries continuously.
  • 48. Data Buffers These Units function as Queues for the streams of packets into the Primary decision node engine. These prevent the different decision units interfering with each other’s operations.
  • 49. Primary Decision Node -1 This unit is responsible for determining whether a packet is a possible threat or not. For example, in the Portscan PDN, all packets with SYN flag set to ON, are marked as possible attack packets. This reduces the traffic load on the Intrusion detection process because packets which have no association with the attack are not subjected to further rule checks. There is a PDN at each network node. Since there are perform lightweight tasks, they can even be placed in low resource devices like IoT.
  • 50. Primary Decision Node - 2 The traffic suspected of belonging to an attack, is then sent to the Complex Decision Engine using HTTP. This choice was made because it is a very common protocol and can be used easily for the purpose.
  • 51. Complex Decision Engine The complex decision engine performs the complex rule and condition checks to make final decision on whether there is an intrusion or not. This unit is placed in a device with a lot of system resources as it handles all the data streams from all the PDNs in the network. The complex processes include, • Dividing traffic streams into time windows • Counting instances of Different incidents • Condition and rule checks • Determining if there is an intrusion
  • 52. Multi Stage attack If a reconnaissance attack is detected, the IDS assumes that another attack would follow. The CDE would then initiate new PDNs to detect possible attacks if they are not already switched on.
  • 53. IDS adaptive behavior Once an attack is detected, the IDS gets updated with all the details of the attack. This includes information like, • The Source IP • The ports used • Geo location if available If an attack were to originate from the same source, it can be detected immediately because of this updating process.
  • 54. Experiment- 1 ■ An experimental setup was used to validate the proposed setup. ■ 3 PDNs were used. Portscan, Ipsweep, Smurf. ■ The CDE was setup on another Server in the same network. ■ The DARPA 98 dataset was used as the traffic source. ■ The test cases were as follows: Test case 1:Test the validity of using hierarchical decision making for faster Intrusion detection Two setups with identical features were used.The only difference was, one setup had hierarchical decision making. I selected 3 dataset files containing the smurf attack. Only one detector was activated. Test case 2: Ensure the results of test 1 are repeatable for multiple attacks Two similar setups but with 3 PDNs and one dataset file containing 3 attacks (Smurf, Portscan, IPsweep) were used.
  • 55. Experiment - 2 ■ Another feature of INSecS-IDS is the flexible design.The functionality of the CDE can be achieved using a Complex Event Processor. ■ A Complex Event Processor is a software that can process streams of realtime data. ■ The advantage of this is, users who are used to SQL like queries can easily use the IDS. ■ The disadvantage would be features like adaptive rule updating would not work unless the CEP is an adaptive one. ■ The CEP picked for this experiment was an adaptive CEP designed and implemented by a fellow lab member. ■ The same 2 test cases mentioned before were used to validate this claim.
  • 56. Experiment - 3 ■ To test the effectiveness of INSecS-IDS compared to an industry grade IDS, another experiment was devised. ■ Snort, (a popular IDS described earlier) was used to detect the same 3 attacks from the same dataset files as Experiment 1.
  • 58. Detection Results for INSecS-IDS Week 1,Wednesday 55.6 5.090 Week3,Wednesday 37.9 6.837 Week 5, Monday 150.65 124.266 Experiment 1 –Test Case 1 Experiment 1 –Test Case 2 Smurf 124.621 14.202 IP sweep 124.616 14.202 Port scan 124.613 14.202 ■ The same accuracy was observed in all test cases. Meaning that all the attack instances listed by the dataset creators were detected by the IDS.
  • 59. Week 1,Wednesday 55.2 5.474 Week3,Wednesday 37.9 7.240 Week 5, Monday 150.65 133.58 Experiment 2 –Test Case 1 Experiment 2–Test Case 2 Smurf 124.621 14.9 IP sweep 124.616 14.9 Port scan 124.613 14.9 ■ The same accuracy was observed in all test cases. Meaning that all the attack instances listed by the dataset creators were detected by the IDS. Detection Results for INSecS-IDS
  • 61. Comparing Snort with INSecS-IDS ■ Experiment 3 showed that Snort also detected the same attack instances identified by INSecS. ■ There was no method to compare the detection time because snort did packet reading and Intrusion Detection on the same computer.There was no way to get the level of distributed functionality that INSecS has. ■ Snort has too many rules, with multiple rule bases. Sometimes there are overlapping rules for the same function. In comparison INSecS-IDS rule base is well structured to avoid duplication.
  • 63. Conclusions for DCS and IDS ■ INSecS-DCS can be used for making customizable datasets for any network. ■ INSecS-IDS performs accurate Intrusion Detection in a distributed environment with low system resources. ■ Hierarchical decision making provides a clear advantage in detection time and large traffic volume handling.
  • 64. Example IoT setup IoT Primary Decision Node Mobile device Primary Decision Node Server Primary Decision Node IDS controller Complex Decision Engine The CDE and IDS controller can be in device with considerable computing resources. The IoT devices can even be in different networks. Each IoT device can have PDNs activated.
  • 65. Example Cloud setup Server Primary Decision Node IDS controller Complex Decision Engine Server Primary Decision Node Server Primary Decision Node The servers can be hosted in different networks. Still the proposed IDS can be deployed. Each Server has PDNs activated in them. Any suspicious traffic is forwarded to the CDE. CDE can be hosted on a device with considerable computing capabilities.
  • 67. Improvements for DCS and IDS ■ The communication between the distributed nodes is done using HTTP. I plan to use encrypted communication to prevent the IDS itself being compromised. ■ The DatasetCreator and the IDS needs to be connected to provide a better IDS solution. Right now, a human expert has to go through the Datasets and make the rules. By implementing the learning module, I plan to make that an automated process. ■ The learning module will generate rules for the specific network, based on the Datasets created by the INSecS-DatasetCreation Software.

Editor's Notes

  1. Good morning all. The topic of my presentation today is Intelligent Network Security System (INSecS) based on the thesis of the same name. I am Nadun Rajasinghe and my Supervisors are Dr. Jagath Samarabandu and Dr. Xianbin wang.
  2. These are some of the Security Issues in Modern Distributed networks. Insecure passwords Insufficient authentication Use of vulnerable network services Data breaches Denial of Service and the list goes on……….
  3. There are different defences for different type of issues. Firewalls Anti virus software Intrusion Detection Systems Encryption techniques Proper Authentication are some of the popular ones.
  4. Out of the different defence methods, I selected Intrusion Detection as the Topic of my project. An Intrusion Detection System is a device or software that monitors a system for malicious activity.
  5. How does an Intrusion Detection System work? There are two popular detection strategies. Anomaly detection- Misuse detection- These can be further classified
  6. These different types of Intrusion Detectors have their own advantages and disadvantages.
  7. These different types of Intrusion Detectors have their own advantages and disadvantages.
  8. Next I will present the existing challenges in the research area selected and My contributions.
  9. So what is the solutions? The one size fits all approach does not work. I am proposing INSecS-Dataset creation software. This was published in CCECE 2018. The software is available for public use on Github.
  10. Now lets look at the challenges in Intrusion Detection in Modern networks. Reconnaissance attacks lead to more devastating attacks. So preparing for that is important.
  11. Taking these problems into account I propose, INSecS Intrusion Detection System.
  12. The Dataset Creation Software will take the packet stream and make datasets
  13. The Intrusion Detection System is placed in the network in a distributed manner.
  14. This is module has not been implemented yet.
  15. The Dataset Creation Software will take the packet stream and make datasets
  16. First lets look at the related work and background of Intrusion Dataset Creation
  17. This is the difference between Raw and Processed Datasets
  18. Consider the number of packets with the same source and destination IPs in the last 100 packets. This is a custom attribute that you can only get from processing raw capture files. There are some datasets like KDD and Kyoto, which provide these kinds of attributes. But the issue is you cannot customize these. For example you cannot change the number 100. or you cannot add any Temporal property to it.
  19. Let us look at work that has already been done to address some of these challenges
  20. The architecture of the Intrusion Dataset creator is as follows. In the next few slides I will go through the functionality of each of the blocks.
  21. Let us look at work that has already been done to address some of these challenges
  22. Let us look at work that has already been done to address some of these challenges
  23. In IDS classification I presented at the beginning, this falls under the following.
  24. Let us look at work that has already been done to address some of these challenges
  25. There are different Commercial grade Intrusion Detection Systems available. Snort is the most popular one.
  26. There are research based solutions for Intrusion Detection Systems suitable for distributed security. I have listed some of them here. But most of them different weaknesses Hence the proposed solutions! A main weakness of this is, they sample at calculated optimum intervals, which means attackers have a chance to slip through. A rule based IDS capable of DOS detection in a distributed system, was proposed by Kshirsagar. Details about rule generation are not explained. Snort has been suggested for distributed security by running instances of it on multiple servers but the researchers encountered resource issues. Using dedicated microprocessors with IDS capabilities, on suitable IoT devices has been proposed as a solution for IoT devices. Lee and others proposed a distributed IDS based on the energy consumption at each node to detect anomalies.
  27. Let us look at work that has already been done to address some of these challenges
  28. The logical architecture in the IDS is a hierarchical architecture. The key essence of it is Primary decision nodes at each network node makes simple decisions Determine if the traffic needs to be further analyzed for possible intrusion detection. Only traffic that has a chance of being related to an attack is forwarded onto a complex decision maker. This is expected to save valuable processing time and resources.
  29. The physical architecture is given in this diagram. I will go through each block to explain how it works.
  30. Let us look at work that has already been done to address some of these challenges
  31. Let us look at work that has already been done to address some of these challenges