CBS
CBS
CBS
What is the PRIVACY AND ML: PROBLEM #1 ?
• ML needs (a lot of) data to learn!
• Machine learning is based on data to
• Find features
• Train the model
• Data is highly unique → allows many inferences
• Anonymizing may not work…
• Aggregation affects utility and requires careful evaluation
• Hide data
• Noise → differential privacy
• Encryption → homomorphic encryption, secure multiparty computation, …
What is the PRIVACY AND ML: PROBLEM #2 ?
• To obtain value, you must give data!
• To use the ML model, you need to provide a sample…
• If the model is remote (e.g., MLaaS), you are giving this sample out
• We can do privacy-preserving model evaluation:
• Predict on noisy data: affects utility
• Use advanced cryptography: affects performance
What is the PRIVACY AND ML: PROBLEM #3 ?
• The output reveals information!
• Membership inference
• Given the answer of the classifier, infer whether the queried example was used in training
• Attribute inference
• Given the answer of the classifier, infer whether a training sample had a particular attribute
Give a definition of the term adversarial example
• Coined by Szegedy et al. in 2014 [1] with respect to computer vision but
goes back further to work on spam and malware evasion by Srndic &
Laskov [2], Biggio et al. [3], and Dalvi et al. [4] in 2013
What is a Working definition for adversarial example
Inputs to a model that an attacker has designed to cause the model to make a mistake
• Set of techniques that aim to fool ML model
What are the 2 ADVERSARIAL ML ATTACKS ?
• Evasion attacks
• Most prevalent type of attacks
• Perturbation of test data, no influence over training data
• E.g., spammers trying to evade detection by obfuscating the content of spam e-mails
• Samples are modified to be classified as legitimate
• E.g., image-based spam: spam content embedded within an attached image to evade textual analysis by anti-spam filters
• Poisoning attacks
• Adversarial contamination of training data
• ML systems can be re-trained using data collected during operations
• E.g., intrusion detection systems (IDSs) often re-trained using such data
• An attacker may poison this data by injecting malicious samples during operation that subsequently disrupt re-training
Give some example
elf-driving cars:
Stop sign → speed-limit sign with scotch
What is the Objective of MACHINE LEARNING 101 ?
Objective: find model parameters that minimize empirical loss
In MACHINE LEARNING 101 How do we find those parameters/weights?
image
What is the Adversary’s objective ?
Adversary’s objective: find a perturbation that maximizes the loss
HOW TO SOLVE THE OPTIMIZATION PROBLEM?
One approach: Fast Gradient Sign Method (FGSM
HOW TO DEFINE SIMILARITY?
The similarity relation is often represented as adversarial cost constraint.
If the goal is to be imperceptible, the common cost is a norm of the perturbation
OTHER SIMILARITY RELATIONS?
• Constraint on perturbation norm assumes that similar images are close in, e.g., Euclidean distance…
• But this is not always true! [1]
• For example, similarity could be defined as small affine transformations [2] :
Why Defending in general is hard AGAINST ADVERSARIAL EXAMPLES ?
We can only defend against a particular threat model (e.g., perturbations up to epsilon norm), and normally no guarantees
What is the Standard approachfor DEFENDING AGAINST ADVERSARIAL EXAMPLES
Training on simulated adversarial examples, i.e., include adversarial examples in the training set
Is Adversarial samples restricted to computer vision ? Give 3 examples
Adversarial samples are not restricted to computer vision
• Twitter bot detection [1]
• Detection tools can easily be fooled by tweaking the number of replies or retweets
• Text classification [2]
• Audio [3]
• Misleading an automatic speech recognition neural network with some noise
• E.g., transcribing "without the dataset the article is useless" into
"okay google, browse to evil.com"
Give exabple of CYBERSECURITY INCIDENTS
• Data or infrastructure as targets
• Attacks against confidentiality, integrity or availability
• Examples of incident description:
• Man-in-the-Middle attack when data was being transferred through the
public Internet
• Phishing
• People running crypto mining on our infrastructure through our APIs
• Server was compromised
• Botnet communication
Give example of PRIVACY INCIDENTS
• Relation to AI vague (beyond data related)
• Attacks against confidentiality, targeting data, not ML model
• Examples of incident description:
• Patient data should not be shared
• Data privacy
• Incorrect use of temporary storage on AWS lambda led to a potential leak of
raw inference [...] potentially containing PII [...]
Give example of: POISONING INCIDENTS
• Attacks against integrity
• Incident descriptions:
• ML systems being retrained to provide false outputs
• Client/partner employees [...] feel threatened by automation, and either stall
or sabotage the labeling effort, harming the models
Give examples of EVASION INCIDENTS
Evasion → 4
• Attacks against integrity
• Incident descriptions:
• Intentional circumventions
• Users spam to optimize their strategy for job search
• A speed-sign as advertisement put onto a driving-school car was interpreted
by the camera as a valid street sign
• Autonomous vehicle image recognition errors leading to dangerous path
planning
WHAT IS COMPUTER SECURITY?
“Measures and controls that ensure confidentiality, integrity, and availability of information system assets including hardware, software, firmware, and information being processed, stored, and communicated.”
What are the 3 KEY SECURITY OBJECTIVES and explain them ?
Confidentiality
Integrity
Availability
What is Data confidentiality: ?
Data confidentiality: private or confidential information not made available or
disclosed to unauthorized individuals
What is Privacy?
Privacy: individuals control or influence what information related to them may be
collected and stored and by whom and to whom that information may be disclosed
What is Data integrity ?
Data integrity: information and programs are changed only in a specified and
authorized manner
What is system integrity ?
• System integrity: a system performs its intended function in an unimpaired manner,
free from deliberate or inadvertent unauthorized manipulation of the system
What is availability ?
Systems work promptly and service is not denied to authorized users
In terms of requirements and loss security (NIST standard) what i s Confidentiality ?
. Preserving authorized restrictions on information access and disclosure, incl. means for protecting privacy and proprietary information
In terms of requirements and loss security (NIST standard) what is Integrity?
Preventing improper information modification or destruction, incl. ensuring information non-repudiation and authenticity
In terms of requirements and loss security (NIST standard) what is Availability?
Availability
• Ensuring timely and reliable access to and use of information
What is Loss of confidentiality ?
unauthorized disclosure of information
What is Loss of integrity ?
unauthorized modification or destruction of information
What is Loss of availability?
disruption of access to or use of information (denial of service)
What are the 3 OTHER SECURITY OBJECTIVES and explain them
• Authenticity
• Verifying that users are who they say they are
• Accountability (non-repudiation)
• Entity cannot deny having done an action
• Anonymity
• Not being recognizable among a group of people
Give the 8 CHALLENGES IN SECURITY
1. Computer security is not as simple as it might first appear to the novice
2. In developing a particular security mechanism or algorithm, one must always consider potential attacks on those security features
3. Security mechanisms typically involve more than one particular protocol and also require that participants be in possession of some secret information which raises questions about the creation, distribution, and protection of that secret information
4. Attackers only need to find a single weakness, while the designer must find and eliminate all weaknesses to achieve perfect security
5. Security is still too often an afterthought to be incorporated into a system after the design is complete, rather than being an integral part of the design process
6. Security requires regular and constant monitoring
7. There is a natural tendency on the part of users and system managers to perceive little benefit from security investment until a security failure occurs
8. Many users and even security administrators view strong security as an impediment to efficient and user-friendly operation of an information system
What is an Asset ?
• Asset
• Hardware, software, data, communication networks, …
• E.g., a major application, high impact program, physical plant, mission critical system, …
What is an Adversary (or attacker, threat agent)
Individual, group, organization, or government that conducts detrimental activities
What is an Attack
Any kind of malicious activity that attempts to collect, disrupt, deny, degrade, or destroy information system resources or the information itself
What is a Threat ?
Any event with the potential to adversely impact organizational operations, organizational assets, individuals, other organizations, or the Nation through an information system via unauthorized access, destruction, disclosure, modification of information, and/or denial of service
What is aThreat model (or adversary model) ?
Technical term that defines the adversary’s capabilities (more about it next)
-
- 1 / 491
-
Flashcards
What can be the reasons of a DoD attack ? (motivation)
• Financial extortion
• Hacktivism
• E.g., operation Avenge Assange in 2010: takedown of the websites of Visa, Mastercard, PostFinance, …
• State-sponsored attacks on opponents
• Diversion from the real attack
• E.g., bank systems where the real attackison their payment switches or ATM networks
What are the 4 Categories of resources that can be attacked?
1. Network bandwidth
• Relates to the capacity of the network links connecting a server to the Internet
• For most organizations, this is their connection to their ISP
• Lower capacity of this connection than the links within and between ISP routers
↳ High volume traffic will deny users access to the server
• E.g., flooding attack
2. System resources
• Aims to overload the network handling software
• Rather that consuming the bandwidth, consuming the limited resources available on the system, e.g., temporary buffers, tables of open connections, …
• E.g., SYN spoofing attack
3. Application resources
• Typically involves many valid requests, each of which consumes significant resources
→ limiting the ability of the server to respond to requests from other users
• E.g., Skype kittens DoS
What was the SKYPE KITTENS DOS ATTACK ?
sending about 800 kitten emojis at once → Skype for Business client not responding for a few seconds
If a sender keeps sendings emojis → the Skype for Business client not usable until the attack ends
What is a FLOODING ATTACK and it's goal ?
• Goal: overwhelm the capacity of the network connection to the target organization
• Traffic handled by higher capacity links on the path…
… but packets discarded as capacity decreases (e.g., in smaller-size company)
• E.g.: flooding ping command directed at the Web server in the target company
What is the problem of a FLOODING ATTACK for a attacker ?
• Problem: source of the attack identifiable since its IP address is used as the source address in the ICMP echo request packets (used in ping)
• Increases the chance that the attacker is identified, and that legal action is taken in response against him/her
• Response to each ICMP echo request packet with an ICMP echo response packet directed back to the attacker → network performance affected
↳ for both reasons, the attacker may want to hide the source address
How adress this problem ?
SOURCE ADDRESS SPOOFING
What is a SOURCE ADDRESS SPOOFING and why to use it ?
Source address spoofing = use of a forged source address
• Easy to do via the raw socket interface on operating systems
• Attacker much harder to identify
What is the core difference between them ?
Destination address: target system & source addresses: random
• Same congestion as in the original flooding attack at the target router
• However, the ICMP echo response packets are scattered across the
Internet to all the random (forged) source addresses
• Identifying the source requires the cooperation of network engineers
managing all routers on the path to specifically query flow information
from their routers
What is a TCP SYN SPOOFING ATTACK ?
Common DoS attack (like flooding attacks)
What TCP SYN SPOOFING ATTACK and how ?
• Attacking the ability of a network server to respond to TCP connection requests by overflowing the tables used to manage them
• Legitimate users are denied access to the server
• Attack on system resources
What is the TCP THREE-WAY HANDSHAKE protocole, give steps?
1. The client initiates the request for a TCP connection by sending a SYN packet to the server.
The server records all the details about this request in a table of known TCP connections.
2. The server responds to the client with a SYN-ACK packet.
3. The client sends an ACK packet to the server and marks the connection as established.
When the server receives the ACK packet, it also marks the connection as established.
Give now the pipeline of a TCP SYN SPOOFING
• The attacker generates many SYN connection request packets with forged source addresses (1).
• For each of these the server records the details of the TCP connection request and sends the SYN-ACK packet to the claimed source address (2).
• If there is a valid system at this address, it will respond with an RST (reset) packet to cancel this unknown connection request.
When the server receives this packet, it cancels the connection request and removes the saved information.
• If the source system is too busy, or there is no system at the forged address, then no reply will return.
In these cases, the server will resend the SYN-ACK packet several times before finally assuming the connection request has failed and deleting the information saved about it.
• In this period between when the original SYN packet is received and when the server assumes the request has failed, the server is using an entry in its table of known TCP connections.
What is the problem with table of known TCP connections ?
The table of known TCP connections has a limited siz
What is the core of a TCP SYN SPOOFING atck and what does the attacker do ? ?
• In a SYN spoofing attack, the attacker sends a very large number of forged connection requests at the targeted server
→ rapidly fill the table of known TCP connections
• Once this table is full, any future requests, incl. legitimate requests from other users, are rejected
• If the attacker keeps a sufficient volume of forged requests flowing, this table will be constantly full → the server will be unable to respond to legitimate connection requests
What are the difference bwtween a TCP SYN SPOOFING VS. BASIC FLOODING ?
Significant difference in the volume of network traffic between a SYN spoofing attack and the basic flooding attack
• Volume of SYN traffic lower than flooding attack, nowhere near the maximum capacity of the network link to the server
• It simply must be high enough to keep the known TCP connections table filled → the attacker does not need access to a high-volume network connection
• Even a medium-sized organization, or even a broadband home user, could successfully attack a large company using a TCP SYN spoofing attack
What is the goal of a FLOODING ATTACKS and how is it performed ?
• Goal: overload the network capacity on some link to a target
• Flooding the network link with a torrent of malicious packets usually overwhelming valid traffic flowing to the server
• Congestion → many packets dropped
• Valid traffic has low probability of accessing the server
→ server’s ability to respond to network connection requests either severely degraded or failing entirely
What type of network packet can be used ?
Virtually any type of network packet can be used
Give the 3 types of PACKET TYPES IN FLOODING ATTACKS
• ICMP flood
• UDP flood
• TCP SYN flood
Explain ICMP flood
• Ping command flood using ICMP echo request packets
• Network administrators allow such packets into their networks because ping is a useful network diagnostic tool
• Remaining one of the most common types of DoS attacks
Explain UDP flood
• UDP packets directed to some port number on the target system
• Spoofed source addresses are normally used if the attack is generated using a single source system, for the same reasons as with ICMP attacks
Explain TCP SYN flood
• Sends TCP packets to the target system
• Most likely normal TCP SYN connection requests with real or spoofed source addresses
• Aim: total volume of packets rather than the system resources
• TCP SYN flooding attacks ≠ TCP SYN spoofing attacks!
What is a DISTRIBUTED DENIAL-OF-SERVICE ATTACKS?
• Flooding attacks with a single system:
• Limited in the total volume of traffic that can be generated
• Easier to trace the attacker
→ Use of multiple systems to perform a DoS attack
→ Distributed denial-of-service (DDoS) attack
How to perform a SSoS attack ?
Systems are typically compromised workstations (i.e., bots)
• Compromised with some malware
• Large collection of bots → botnet
• Botnets can then be used/rented for DDoS attacks
Why a DoS attacks cannot be prevented entirely ?
• An attacker can direct a large enough volume of traffic to your system
• High traffic volumes may also be legitimate
• High publicity about a specific site
• Activity on a very popular site
What are the Four lines of defense against DoS attacks ?
• Attack prevention (before the attack)
• Attack detection and filtering (during the attack)
• Attack source traceback and identification (during and after the attack)
• Attack reaction (after the attack)
What are the 2 element for PREVENTING DOS ATTACKS ?
Block spoofed source addresses
Reverse path filtering (RPF)
How to Block spoofed source addresses ?
• Needs to be done as close to the source as possible, i.e., by routers or gateways knowing the valid address ranges of incoming packets
• Typically by the ISP providing the network connection for the source
• Best placed to ensure that valid source addresses are used from its customers since it knows which addresses are allocated to all its customers
• Using explicit access control rules in a router to ensure that the source address is one allocated by the ISP
How to Reverse path filtering (RPF) ?
• Filters to ensure that the path back to the claimed source address is the one being used by the current packet
• When a packet arrives at a router, the router checks if it were to send a reply to this source IP, whether it would go through the same interface it just came in from
• Filters must be applied to traffic before it leaves the ISP’s network, or at ISP entry points
How to DEFENSE AGAINST TCP SYN SPOOFING ? Give the 2 steps
• Modify the TCP connection handling code
1. Instead of saving the connection details in a table, cryptographically encode session information in a SYN cookie that is sent as the server’s
initial sequence number in the SYN-ACK packet
2. If the client is legitimate, it will return the SYN cookie (with incremented sequence number) in the final ACK packet
3. The server verifies the cookie, reconstructs the connection state, and establishes the session
What is the advantage and disadventage of this defense ? (Modify the TCP connection handling code)
• Advantage: not consuming any memory resources on the server until the three-way TCP connection handshake is completed
• Disadvantage: take computational resources on the server to calculate the SYN cookie
What is a Good incident response plan for RESPONDING TO DOS ATTACKS ? (4)
• Clear roles and responsibilities
• Step-by-step procedures for identifying, mitigating, and recovering from an attack
• Communication plan with internal and external stakeholders (customers, partners, …)
• Details on how to contact technical personal for the ISP
• Needed to impose traffic filtering upstream
• Non-network means of contact must be available (e.g., phone number)
What can be preparated for Respondinf to DOS Attck ?
• Antispoofing, directed broadcast filtering, and rate limiting
• Ideally, automated network monitoring and intrusion detection system (IDS) to detect and notify abnormal traffic patterns
In RESPONDING TO DOS ATTACKS, what is the best defense against it?
• Capture and analyze packets
• Design filters to block attack traffic upstream
• Or identify and correct system/application bug
How an ISP can help you ?
ISP trace packet flow back to the source
What are the 2 other measures in RESPONDING TO DOS ATTACKS ?
• Implement contingency plan
• Either to switch to alternate backup servers
• Or to commission new servers at a new site with new addresses
• Update incident response plan
• Analyze the attack and the response for future handling
Why Internet access creates a threat to modern organizations?
it creates a threat since it enables the outside world to reach and interact with local network assets
What is an effective means of protecting local area networks (LANs) ?
Firewall
Where Firewall are insered and how it works ? ?
• Inserted between the internal network and the external network to establish a controlled link
• Checks all traffic and implements some access control rules to decide which flows should enter in the local network
Why do we say that it is Used as a perimeter defense ?
• Single choke point to impose security and auditing
• Insulates the internal systems from external networks
• Following the defense-in-depth security principle
What are the 4 Design goal of Firewall ?
• All traffic from inside to outside, and vice versa, must pass through the firewall
• Achieved by physically blocking all access to the local network except via the firewall
• Only authorized traffic, as defined by the local security policy, will be allowed to pass
• Various types of firewalls can be used
• The firewall itself must be immune to penetration
→ Secured operating system
