Security for dummies @ishaikhspear-blog - Tumblr Blog

Week 6 lecture notes

AES

Stands for Advanced Encryption Standard. AES uses substitution permutation network. It seems to not be broken in feasible time. Block cipher using a symmetric key.

Block Modes

Allow block ciphers to work with large data streams without the risk of compromising the provided security. Initialization vector is randomly generated used to ensure distinct ciphertexts are produced even when the same plaintext is encrypted multiple times independently with the same key.

ECB (Electronic Codebook Mode)

Encryption: Each plaintext block is encrypted separately using a key to create a cipher text.

Decryption: Each cipher-text block is decrypted separately.

CBC (Cipher block chaining)

CBC is like ECB but with an initialisation vector. The first plaintext block is XOR’d with a random initialisation vector which is the same size as the plaintext block and then for each subsequent block then previous block is used as the next IV.

Block modes like CBC and ECB use padding to bring its length up to a multiple of the block size.

Disk Encryption

Generate a key to encrypt the disk. Encrypt the key (that can be decrypted with a password) and store it somewhere NOT ON THE DISK.

RAM can still store data even after shutting off power. If the RAM is frozen, the data leaks even slower and thus can be retrieved.

Cryptocurrency

What is a block chain? A chain of blocks where each block contains data of hundreds of transactions, a hash of the block and a hash of the previous block. The hashes make the block tamper resistant. Probably use a HMAC or something since we have already discussed how MACs and hashes can be tampered with.

How blocks are formed:

Transactions are grouped into a pool

Miners gather transactions to form a candidate block

Candidate block is given metadata known as a block header

Hash the block with a nonce - randomly generated number

If the hash is lower than a certain target value the block is added to the chain

Miners race to find the nonces

#lecture notes

Week 5 lecture notes

“Don’t roll your own” = never try to make your own cryptographic function. It will always break

Wired Equivalent Privacy (WEP)

Wired Equivalent Privacy is a security protocol that is designed to provide a wireless local area network (WLAN) with a level of security and privacy comparable to what is usually expected of a wired LAN.

WEP uses an initialisation vector (IV) to produce 2^24 keystreams. A key will be guaranteed to be used again, which is very bad. Using XOR, they can figure out the plaintext.

How it works:

A stream of bits (A) is sent from one point to another

A random stream of bits (B) is generated using an algorithm such as RC4

A XOR B -> C (encrypted)

C XOR B -> A

This gives the property of randomness because since something random XOR something else will also result in something random. The danger occurs when someone transmits the same data under the same key - same data in the same frame.

Hashes

Hash functions like MD5, SHA0, SHA1 are all broken. SHA2 and SHA3 have not been broken (as we know of). SHA3 only one not vulnerable to length extension attacks - a type of attack where an attacker can use Hash(message1) and the length of message1 to calculate Hash(message1 ‖ message2) for an attacker-controlled message2, without needing to know the content of message1. Since SHA3 has updated internal mechanisms that creates a large change in the produced output from a small change in input.

Birthday attacks

The birthday attack is used as an idea around how good hash functions should be designed by taking the example of birthday paradox. It says that in a room full of people, the probability of finding two people with the same birthday is root(n), where n is the total number of people. This concept can be applied to hash collisions as well, where the probability of find 2 messages, M1 & M2 that produce the same output is considerably lower than mathematically advertised.

If the hash function has less probability than sqrt(n) of getting the same hash(where n is the number of bits in the hash function output - similar to the number of people in the room), the hash function is considered a strong hash function.

Merkle Damgard Construction

Merkle Damgard Construction - is a method of building collision-resistant cryptographic hash functions from collision-resistant one-way compression functions (used in the design of MD5, SHA1, SHA2)

Break the message into a series of blocks

Start with IV (Initialisation vector)

Given hash function f, it takes in IV and block 1, output f(B1) goes into f again with block 2 etc.

After doing this with all blocks you have your hash.

This would not prevent against length extension attacks. Since the hashing of separate blocks in the message would leave the recalculation of newly added blocks easily exploited.

Digital signatures + RSA

Digital signature is a mathematical scheme for verifying the authenticity of digital messages or documents. A valid digital signature, where the prerequisites are satisfied, gives a recipient very strong reason to believe that the message was created by a known sender (authentication), and that the message was not altered in transit (integrity)

RSA relies on people not being able to factorise the large number that is the mod, if someone is able to efficiently find the factors of a number, RSA will break.

Hash Collisions

Collisions are finding 2 messages M1&M2 such that they both produce the same hash.

Collision resistance: It should be hard to find any two things that have the same hash

Key Stretching

Key stretching techniques are used to make a possibly weak password more secure against a brute-force attack by increasing the bits it takes to test each possible password.

Online: Brute forcing the website

Offline: Steal the hard drive and try to crack the passwords at home.

Salting is the process of appending a byte of data to each password and hashing that total. The salt is then stored in the table next to the password but a potential attacker needed to add the salt to each M and then hash that and compare it to the value, it takes significantly longer to do that for every password.

Key stretching is designing a hash to be slow, such that it adds even more work for a hacker to decrypt a table of passwords. The added time for an individual user would be negligible by comparison.

#lecture notes

Diffie-hellman

So in the context of end-to-end encryption, we need some kind of symmetric key to encrypt/decrypt our message to communicate privately and solve any man-in-the-middle problems. Diffie-Hellman algorithm is how we create that secret key and exchange it between users in a network.

This secret key exchange was first published in 1976 and has quickly become a staple for most cryptography techniques with the same problem. The problem is whenever we need a secret key it’s very had to exchange this key because sending across networks would immediately expose it to anyone watching the network. So to start communications we need two values, g & n, where g is a value and n is a large prime number. Alice and Bob in this example both need private variables, lets calls them a and b, where a&b are some number between 1..n.

So the next step is to use the public g raised to a, b to produce 2 new public variables g^a mod n and g^b mod n. Now because of the nature of the computation it’s very hard to discern what these new variables are given just their values. Alice and Bob will then exchange these 2 variables between each other, so g^a mod n and g^b mod n are now public for the sake of security. Now Alice raises g^b mod n to a which gives g^(ab) mod n because of exponent laws. Bob does the same and gets g^(ab) mod n. So now Alice and Bob both have a secret key that they can use.

#self study

Length extension attacks

So to understand length extension attacks we need to understand message authentication codes. Message authentication codes (MACs) are a way to verify the authenticity of a message. In the more naive implementation of a MAC, the server generates a hash based off the message to use as its MAC value. Then it appends this value to the end of the encrypted message.

The problem with this is you can change the contents of the message and recompute the hash. So what next? We need a secret key, a shared secret. We take out key, K, append that to the message and hash that. So now if they alter the message they can’t recompute the message. But it’s now susceptible to length extension attacks. The problem is that some hash functions have a predictable way of hashing a message such that if you append something to the message the hash value will only be the hashed value of the original message + the new message. So that’s why it’s called a length extension attack.

The way around this is to use HMACs. In which we have 2 sub keys, k1 & k2. You produce a hash of h(k1 + M) and then hash that with k2, so h(k2+h(K1+M)). This new HMAC is resilient to length extension attacks.

#self study

BEAST attack

BEAST is short for Browser Exploit Against SSL/TLS. This vulnerability is an attack against the confidentiality of a HTTPS connection, it does this is a relatively feasible amount of time. That is, it provides a way to extract the unencrypted plaintext from an encrypted session.

BEAST has three conditions that must be met for this attack to take place:

JavaScript or applet injection into the same origin of the web site

Network sniffing of the connection must be possible

A vulnerable version of SSL must be used which is using a block cipher

Taking a look at these conditions it seems that when you have all 3 you aren’t limited to only performing a BEAST attack. If you had the ability to inject code into the website you have access to the page content, you can perform any action the user can do, you can steal cookies and hijack the session.

You can prevent these attacks for sure, one example is the HTTPOnly property that can be set on cookies. With this property, a browser will not allow applets or JavaScript to gain access to the cookie value. This can prevent session hijacking. So if you are worried about the BEAST attack first worry about the HTTPOnly property.

#security everywhere

POODLE attack

A vulnerability was discovered in the SSL 3.0 protocol, which can allow an attacker to have access to personal information such as passwords and cookies.

SSLv3 has a feature where it provides a protocol downgrade dance in the interest of working around server-side interoperability bugs and improving user experience. You can probably see where this is going.

POODLE attack (Padding Oracle On Downgraded LegacyEncryption) is another example of a downgrade attack that allows hackers to, for example steal “secure” HTTP cookies.

The recommendation to stop this attack from happening is just to disable SSLv3 protocols on either the client or the server side or both if possible. If either side only communicates with SSLv3 then “all hope is lost”.

The authors' description of the POODLE attack brings attention back to the the BEAST attack, essentially binding the discovered vulnerability with the latter. Yet, even though the new attack looks fairly serious and might produce really bad results for the affected parties, it would not be correct to collate POODLE and BEAST, as both attacks have different targets, use different methods, and exploit different vulnerabilities. What we can say for sure about the relations between those two is that POODLE can be used to prepare the environment to conduct BEAST.

#security everywhere

DROWN attack

DROWN stands for 'Decrypting RSA using Obsolete and Weakened Encryption'. SSLv2 was first released in 1995 and depreciated in 2011. It was found that 33% of all HTTPs servers and 22% of those with browser trusted certificates are vulnerable to the attack.

So how does the DROWN attack work? Because most people will not buy multiple certificates, a server will use the same RSA private key for both TLS and SSLv2 protocols meaning that any bugs from SSLv2 could easily affect the TLS. The DROWN attack uses this exact method to break encryption. Sites that support SSLv2 and EXPORT cipher suites are vulnerable to this attack.

Solution

It is recommended to use Mozilla as it has an SSL Configuration Generator that will generate SSL server configurations for common web servers. This includes guidance on cipher suites, HTTP Strict Transport Security (HSTS), OCSP Stapling, Session resumption, HPKP Public Key Pinning, etc.

Disable SSLv2 and/or update OpenSSL.

And finally, regularly change your system passwords and update your SSL Certificates with new keys. This will limit the the impact that a future hack will cause on your system.

#security everywhere

Difference between SSL & TSL

What’s the difference between SSL (Secure Socket Layers) and TLS (Transport Layer Security)? SSL and TLS are both cryptographic protocols that provide authentication and data encryption between servers, machines and applications operating over a network. SSL is the predecessor to TLS.

SSL was originally developed by Netscape in 1995. There were a lot of vulnerabilities with Version 2.0 and thus needed to be replaced with Version 3.0 very quickly. TLS was introduced in 1999 as a new version of SSL and was based on SSL 3.0.

Vulnerabilities have been discovered in the deprecated SSL protocols, for which I will link some of them to my related Blog posts:

POODLE

DROWN

BEAST

On top of that, most modern browsers will show a degraded user experience when they encounter a web server using the old protocols. For these reasons it is recommended to use TSL protocols.

#Week7 #self study

Vulnerabilities in TSL

I discovered there are are some "serious weaknesses" with the Transport Layer Security (TLS) protocol, which provides security for many sites dealing with valuable information -- including online banks, email databases and social networks like Facebook.

Admittedly, the conditions of the attack are pretty limited. It only works when connected to another computer over a LAN network, not on the web. The specific encryption needed also needs to be an "HMAC-SHA1 MAC algorithm".

Normally man-in-the-middle attacks are relatively well protected against but the problem that the found is that at the very end of each piece of info that's sent using TLS there's a tiny bit of unencrypted information. It's only up to four bits of the last byte of each encrypted piece of data. If the attacker is able to repeatedly see the encrypted information they'll be able to see a different few bits of it each time thanks to the inherent "jitter and other effects" across a network. In their tests, they found that they could recover any piece of TLS-encrypted plaintext with around 2^23 transmissions of information, or only 2^19 if they know in advance that the encryption is in base 64. For some TLS networks, an attack caused the connection to drop, but then try to reconnect and send another encrypted cookie, letting the researchers run their hack over and over again until it built up the full picture of what was encrypted. They called this the "Lucky Thirteen" attack, because it's apparently only possible as an unintended consequence of the 13 bytes in a TLS header.

Now this is pretty specific and with most communications this wouldn’t be a problem but if it is true that hackers are always a couple of steps ahead then what else could they know? Also, reassuringly, the researchers write that "OpenSSL, NSS, GnuTLS, yaSSL, PolarSSL, Opera and BouncyCastle are preparing patches to protect TLS against our attacks" and that they've "informed several other vendors about the work, including Apple, Cisco and Microsoft".

#security everywhere #Week8

Week 4 - Lecture notes

Protocols - Integrity

Umbrellas make a good defence against cameras substitution cipher example on English alphabet

26! is very big ~ 100x(10^3)^8, 100~7 bits, 10^3 ~10 bits, total ~90 bits of work.

For a 16GHz computer it would take a whole year of running to compute => 16 x 10^9, 2^34 operations per second * 2^12 seconds in an hour * 2^5 hours in day * 2^9 days in year = 2^60 = 60 bits of work.

Theoretically it would be infeasible to crack a substitution cipher, yeah because of the patterns in the English language it takes considerably less time than advertised. This is because English has low Entropy.

Passwords are more likely to use common english phrases. Guided brute force algorithms takes advantage of this.

Telegraph example: possible attacks

Replay Attacks An attacker uses replay messages from a different context into the intended or original context. This exploits the predictability or protocols and fooling participants that think they are successfully completing task up to protocol.

Main-in-the-middle An attacks secretly relays and/or possibly altering message between 2 parties communicating in a network that believe the are only communicating to each other.

Hashing

A method of cryptography that generates a seemingly random value given distinct messages/data. The hash function must be consistent in assigning these values.

They must return the same y given an x (deterministic)

Avoid Hash collisions - where 2 distinct x’s map to the same y value

Clustering in the data causes clustering in the hash function

Birthday attack exploits hash collisions

To check if a file has been tampered with, compare hash’s of the original file and the file you received to see if anything was changed. The site will normally put up the hash of their file on the website.

Putting it all together, the features of a hash are:

Deterministic

Fast to hash, infeasibly slow to decipher

Avalanche property

Hard to find 2 messages with the same hash (Collisions)

Social Engineering

Social engineering is the art of learning and lying to gain access to physical systems and into software vulnerabilities. Mostly relies on human manipulation. 95% of attacks use some sort of social engineering cause people are so trusting. It usually takes 146 days to detect a breach cause by social engineering. E.g. phishing, ransomware, USB baiting etc.

Life Cycle

Investigation - learn as much as possible about victims

Hook - initiate the conversation with the target and build rapport

Play - obtain the information

Exit - leave the conversation without seeming suss

Social engineering vectors

Pretexting - invented scenario to get information out from someone

Baiting - taking advantage of curiosity or greed

Quid Pro Quo - getting something for something

Tailgating e.g. heavy box technique

Phishing - like emailing someone to get credit card details or sensitive information

Principles of Persuasion

People can be easily exploited and/manipulated.

Reciprocity - people like to return acts of generosity with acts of generosity. Taking advantage of obligation essentially.

Liking - Getting someone to like you will make it easier for you to manipulate or take advantage of them

Social cues - peer pressuring and exploitation of purely social functioning in the victim and his/her circle.

Authority - people often blindly follow orders without asking too many questions.

Hashing

Setting the hash output to the biggest it can possibly be is the only way of preventing most attacks. That way even the square root of a number is too much work to do.

Preimage attacks - given an h(M) find an M. Usually takes 2^n bits and on average 2^(n-1).

2nd preimage attack - given M find M’ such that h(M)=h(M’). Usually only 2^(n) and on average 2^(n-1)

Collision attack - find both M and M’ such that h(M)=h(M’). Usually only 2^(n/2) and on average 2^(n/2-1)

Examples of hashes include:

Fingerprinting - check that the hash generated from the fingerprint scan matches the one save on file

Proof of Work - in bitcoin you only get a new block when you solve a very difficult puzzle. It’s hard to produce and easy to verify.

Message Authentication Codes (MAC) - append a password to the end of a message to be hashed.

Cryptographic Hashes

Has some extra properties:

Deterministic

Needs to be quick to compute

Hard to reverse (no preimage attacks)

Hard to find collisions

A small change in the text results in a large change in the hash

#lecture notes #Week4

Week 3 - Lecture notes

Kahoot Answers

Kasiski's method

brute force 6 digit launch code -> 20 bits (220)

brute force 6 digit launch code (on average) -> 19 bits (219)

Risk

David Hume - is ought : confuse what is for what ought to be. "Break the glass!"

Laptop repair High/Low probability, low/High impact low prob, low impact - people don't care

Invisible Risk is invisible. People only care when the risk goes wrong, they don't consider they took the risk in the first place. Taking a risk is just as bad [morally] as the worst outcome.

Humans are bad at assessing risk Security is all about assessing risk. Human nature is about gettting obsessed about a risk or forget about it completely.

Terrorism If you worry about it then you focus on it. You then focus on the low prob risk.

Centralisation (not done yet)

Cloud

Google

One player

Hong Kong attacks

Patients in hospitals have had records accessed by government -> arrested because of protesting against

Trust is broken

Backdoor in record system

Time/space tradeoff

You can often do more work with more storage. If you can store what you did prev you can speed up your work. Sometimes you can halve the number of bits. 60 (with more storage) -> 30.

Insiders/Traitors - the biggest problem in terms of catastrophic failure - but often overlooked

Every contact leaves a trace

Dealing with risk

Prevention Try to prevent the event from actually happening by removing vulnerabilities which allow the event to occur.

Limitation Allows have a backup when you can’t prevent the event from happening. Limit what the attacker can do/access if he did get through the defence.

Passing the risk of a 3rd party Try to limit the effects of the risk by passing on responsibility to a 3rd party.

Wearing the risk If the risk can not be fully prevented it’s time for damage control. What measures do you have in place that can limit the impact of a breakthrough?

Public key Cryptography

Confidential

Need a key for every person you are talking to

for 100 people, n2

-> Public Key cryptog

1 key to decode, 1 to encode

encode is public

Merkle puzzles

Use a cipher, key is a 4 letter word. If I gave you the cipher text you could brute force it. The sender sends thousands of these tiny puzzles and the receiver only needs to crack the correct one to get the message. The middleman will have to crack all of them to be sure he has the correct message.

RSA

Solves the problem by having 2 keys. 1 private & 2 public - for decrypting and encrypting respectively. We don’t care who has access to the public key. The private key is retained by the both parties. Based on modular arithmetic (ie A^B mod C = ( (A mod C)^B ) mod C).

Lock Picking

Lock bumping Bump key inserted one notch short of full insertion, then you quickly inser the key and all pins should “bump“ into all the correct places

Tension Wrench Applies torque to the plug of a lock to hold picked pins in place

Rake pick Long picks with a rakes at the end. Used to insert and take out multiple time.

Shim Thin piece of metal, slide it into the shackle housing and pull it from side to side

#lecture notes #Week3

Week 2 - Lecture notes

Security by Design

If one thing fails then the whole system shouldn’t.

There are 4 primary colours of security engineering:

Trust Well there is none. Trust no one and don’t place complete trust in any one thing.

Secrets

Humans

Engineering

Defence in depth

The system should be able to work if something fails

Try to avoid single point of failure

Walls of a castle analogy - if one thing fails there are other layers of defence

Segmentation of confidential data

Security by Design

Design the system with security in mind. Don’t just add on layers when an attack occurs.

Apes vs ants - 2 different paradigms

Apes - complexity between individuals Every ape is independent of each other, yet have a necessity to work together. This facilitates cohesion between parts in a system.

Ants - each ant is essentially dumb on it’s own The colony works as a whole to create solutions and use pieces in a greater game. No single ant has freedom to do whatever it wants. The colony is rigid and demands obedience.

Bell LaPadua - levels of confidentiality

Restricting access of specific information

Complete human obedience

High control = Large penalties to deter

Could have have an idiot at the top, meaning a single point of failure

Physical Security

If someone can access the hardware they can do anything

We programmers take for granted physical security

Consider hardware security measures before software security measures

Can easily capture passwords by setting up cameras or looking for imprints

Vigenere (Caesar + passwords)

Shift the encrypted message by the alphabetical order of the key. e.g ABBAA would five a shift of (0,1,1,0) on repeate, so “Hello” would become: H+0(A) = H E+1(B) = F L+1(B) = M L+0(A) = L O+0(A) = O

The Kasiski test

Used to crack ciphers where the key repeats like the Vigenere

Find the length of the key by looking for repetitions (longer repetitions are better)

Line up each section, then solve the lined up letters as single substitution

The index of Coincidence

Makes it possible to evaluate the global distribution of letters in encrypted messages for a given alphabet.

High index - message probably been encrypted with transposition cipher monoalphabetic substitution.

Low index - probably been encrypted using a polyalphabetic cipher.

Estimation of index:

Line up text with copy of itself shifted slightly

Count the number of times letter in first copy matches the second copy

Divide by number of aligned pairs of letters

Multiply by 26

Enigma Machine

Enigma Machine was an encryption device used by Nazi Germany in WW2 to protect commercial, diplomatic and military communication. Enigma has an electromechanical rotor mechanism that scrambles the 26 letters of the alphabet. It’s like the substitution cipher by with a theoretically infinite password.

One Time Pad

OTP is an encryption technique that cannot be cracked, but requires the use of a one-time pre-shared key the same size as, or longer than, the message being sent. In this technique, a plain text is paired with a random secret key (also referred to as a one-time pad). Then, each bit or character of the plaintext is encrypted by combining it with the corresponding bit or character from the pad using modular addition.

Type 1&2 Errors

Type 1: False positive A medical test saying you do have a disease when in reality you do not

Type 2: False negative A medical test saying you do not have a disease when in reality you do

Insiders

Basically a security violation that can only be done from the inside. Spies, espionage or generally just greed and corruption.

#lecture notes #Week2

Week 1 - Lecture Notes

Security Engineering

What makes a good security engineer:

Good at analysis Don’t just rely on public opinion or what your gut feeling says. Use logic to deduct; think from an attackers perspective as well as a defenders. Identify strengths, weaknesses and recommendations.

Need an awareness and be able to farm other disciplines Learning from history to analyse and learn from past mistakes. Build upon previous solutions and identify features that look familiar.

Case studies Whenever something goes wrong in the security field a good security engineer should break down the scenario to points of failure. Don’t just allocate blame like the masses. Use logic to breakdown events; find reasons as to why things fail; find out how they failed/exploited.

Decisions We need to be able to make pro’s and con’s for all viable options.

Identify what’s important Sift through the mess of complexity and break solutions down to its most important bits.

Certainty You can never be too sure of yourself, there could always be a way through your defences. The job is never done, keep building on your solutions

Humans To understand the systems we inherited we need to understand ourselves (humans). Many issues occur between the communication between humans and tech, so just focusing on tech will never give you a complete solution.

Vulnerabilities vs. Exploits vs. Threats Vulnerabilities are weaknesses in a system. Exploits are methods of taking advantage of a vulnerability. Threat models are basically a description of what you are up against.

Security Through Obscurity

Security through Obscurity: Relying on secrecy or complexity for security.This was an issue for a variety of reasons including:

Increasing complexity or obscurity means that it is very difficult to build upon the solution or analyse its efficacy when changes are made.

This added obscurity might contribute to other unforeseen vulnerabilities cause it is so hard to put together.

Kerckhoff's principles

The system must be practically, if not mathematically, indecipherable.

It should not require secrecy and should not be a problem if it falls into enemy hands.

It should be possible to communicate and remember the key without using written notes; it should also be possible to change the key at will

It must be applicable to telegraph communications

it must be portable

Lastly, It should be easy to use and not require a long list of rules of use.

We should not rely on obscurity to make things secure. Good security happens when the enemy knows everything you know but they can’t decode what you can.

Cryptography Principles

Confidentiality - Only authorised personnel can access restricted information. Integrity - Maintaining the “trustworthiness” of a message. Authentication - The act of verifying if the person presented to you is the person you think it is

Steganography - hiding the existence of a message

Issue - as soon as the method is known/cracked the message and all past messages are compromised.

Example - Greeks used to tattoo messages on the bald head of slaves. The message would be concealed when the hair grew back

Codes

Caesar cipher shift the letters by the amount of the key

Substitution cipher Use the key to substitute for other letters

Transposition keep the letters the same but change the order

Jefferson’s wheel Set of wheels with an alphabet written on them. Order the wheels and line up the plaintext to code. Write down a lined up gibberish line. To decrypt, recipient arranges disks in agreed order.

Playfair cipher Encrypt a pair of letters instead of single letters

Polybius square Put the letters of the alphabet onto a grid and replace each letter with its coordinates on the grid

Case Study: The Halifax Explosion

The incident caused the largest man made explosion prior to the atomic bomb

SS Imo was carrying a large shipment of explosives through port Halifax

A misunderstanding between SS Imo and SS Mont-Blanc caused a collision course

Sailors abandoned ship prior to collision and escaped the explosion

Resulting explosion killed 2000 people and injured 9000

Vince Coleman, a local train dispatcher was notified of the explosion about to occur and stayed at the dispatch station to prevent trains from entering Halifax saving 1000s and allowing stopped trains to be used for the recovery effort

#lecture notes #Week1

Week 8 Lectures notes

Errors

A fallback attack is an exploitation of the negotiability of a protocol to abandon a higher quality mode of operation (e.g. an encrypted connection) in favour of weaker security, maybe for reasons like backward compatibility.

China syndrome - Threat of nuclear disaster where a reactor melts through the ground a reaches china.

Root cause analysis - working out what went wrong

could lead to improvements, might not

humans prefer to have a single explanation

First culprit is usually referring to user error

We like to blame people

Cyber attacks will always be the result of human error.

Second culprit usually refers to culture

We all are to blame for bad culture

Third culprit is the whole system being used

Human weakness

Honour codes

Misdirection and limited focus

In any situation there are more noticeable things and less noticeable things.

We only focus on a few things at once.

We have a series of heuristic we follow to tell us which factors to look at Similarity matching - see how similiar a situation is to something you’ve experienced, seen, or heard of before. Frequency gambling - trying to recall a pattern of things you’ve seen before and follow the steps of that pattern.

We are thus extremely susceptible to misdirection. This is exploited by social engineers and magicians.

Accident vs. Attack: An intelligent agent is responsible for an attack. Murphys Law is behind an accident. “Satan’s computer”, says that for everything I don’t directly control I have to assume will be turned against me.

Confirmation bias

Satisficing - getting something “good enough”

Bounded rationally We have this small amount of attention that we can put into analysing things that we don’t want to waste.

People prefer to prove a generalisation that we come up with - self fulfilling prophecies.

Group-think When you prioritise group membership over getting things right.

System error

Highly complex system usually won’t have a single point of failure.

Culture likes to blame the last cause, but the flawed system is where a solution lies.

High/low coupling (finally a primarily programming topic) High coupling defers to objects or conceptually separated elements in code have a strong codependence with one another - if one changes then the other will likely be affected. Low coupling is when objects are very independent of one another - each object is conceptually very separated from the rest.

High/low coherence The ability of objects to work together in an orderly and succinct fashion.

System with high coherency and low coupling have better resilience

Never be 100% confident

Due to oversimplifications of problems we tend to account for fewer contingencies than actually occur.

Hindsight bias Things that happen seem inevitable and things that don’t happen seem impossible

Defence in depth can go wrong when the measures fail invisibly

Automatic security measures deskill the human operators.

#lecture notes #Week8

Case Study Review - Snoop

This week in your tutorial class you'll be debating privacy - specifically should the government or government agencies collect and have access to your data for good purposes, or should citizens, e.g. you, have a right to privacy which stops them?

I really thought a lot about this question since my tutor had made a point in talking about hope the Australian government passed controversial laws designed to compel technology companies to grant police and security agencies access to encrypted messages. This has resulted in some companies blacklisting Australia.

Vault Cloud, which provides online storage to parts of the federal government including the Department of Defence, said companies were abandoning Australia because of its small market and the "perceived compliance burden of the jurisdiction".

So let’s get into the topic. Should the government have access to your data? In our tute we posed the question as “should the government have increased surveillance?” and that covered cyber territory and also more physical forms of surveillance.

I was on the side against such actions because the more and more we become integrated into an “internet of things” society the more of a privacy invasion it’ll really be. Also it’s a possibility that with vague and ill defined laws as a foundation for how the government deals with privacy leading us into the future where new technologies and new instances of data collection will occur. How’s to say they will use their growing superpower wisely? Who will police the government?

There was an interesting article referencing how most the Australian population didn’t really care about their privacy. They were on average mostly on the fence. Maybe they believed they had nothing to hide and this meant that their privacy wasn’t that much of a priority. It does seem like if the laws tighten up or alternatively stay the same in an increasingly data driven society that most Australians will not be on the fence anymore.

https://theconversation.com/australians-accept-government-surveillance-for-now-110789

Another point to make would be determining how efficient would surveillance be in catching terrorist or criminals or more vaguely people that pose a threat to public order. There was an article that indicated that it’s a lot more complex than it seems and it swayed in the other direction of, maybe it’s not that effect as we think?

https://www.tandfonline.com/doi/full/10.1080/01972243.2017.1414721

So I am pretty much against increasing the extent of surveillance.

#Case Study

CTF: Mistake

We all make mistakes, let's move on. (don't take this too seriously, no fancy hacking skill is required at all)

This task is based on real event Thanks to dhmonkey

hint : operator priority