Lossy vs lossless compression impact on customer (1 mark each, max 5):
• Lossy compression reduces file size significantly by permanently removing some data (1 mark)
• This means the video will stream faster / use less bandwidth / buffer less, but the quality will be reduced (1 mark)
• Lossless compression reduces file size without losing any data / the original quality is preserved (1 mark)
• The video will be higher quality but the file size will be larger, requiring more bandwidth and potentially causing buffering on slower connections (1 mark)
• The customer must choose between quality and performance based on their internet speed and preferences (1 mark)

Two uses of hashing in a database (2 marks each, max 4):

Use 1 (2 marks):
• Hashing can be used to store passwords securely (1 mark)
• The password is run through a hash function to produce a hash value which is stored instead of the plaintext (1 mark)
• if unauthorised access happens then they can't read the passwords

Use 2 (2 marks):
• Hashing can be used to create hash table indexes for faster data retrieval (1 mark)
• A hash function is applied to a key field (e.g. CustomerID) to calculate the storage location / index, allowing direct access to records without searching the entire database (1 mark)
• fast access (O1)

Encryption vs hashing for robot data (9 mark extended response):

Level 3 (7-9 marks): Thorough comparison with clear application to the scenario
Level 2 (4-6 marks): Good comparison covering most aspects
Level 1 (1-3 marks): Basic points

Types of encryption:
• Symmetric encryption - uses the same key for encryption and decryption. Faster but the key must be shared securely between robots
• Asymmetric encryption - uses a public key for encryption and a private key for decryption. More secure for communication as the private key is never shared
• Symmetric is suitable for encrypting stored data; asymmetric for initial key exchange between robots

How hashing secures data:
• Hashing converts data into a fixed-length hash value using a one-way function
• Cannot be reversed / decrypted back to original data
• Suitable for storing passwords - the robot can verify user passwords without storing them in plaintext
• Used for data integrity checking - if data is modified during transmission, the hash will change

Why each is used:
• Encryption for communications - data sent between robots wirelessly needs to be encrypted so if intercepted, it cannot be read. Asymmetric encryption is ideal for this
• Encryption for stored data - sensitive data stored on the robot should be encrypted in case the robot is physically accessed
• Hashing for passwords - the wireless network password should be stored as a hash, not encrypted, because it never needs to be decrypted back to plaintext
• Hashing for integrity - hashing transmitted data ensures it has not been tampered with during wireless communication

Why lossless for thin client data (1 mark each, max 3):
• Text input data must be transmitted accurately / exactly as the user typed it (1 mark)
• Lossless will not permanently remove data
• Lossy compression permanently removes data which could corrupt the text / change characters or commands (1 mark)
• Lossless compression preserves all the original data so the virtual machine receives exactly what the user input / data can be perfectly reconstructed (1 mark)
• Even a small loss of data in text input could make it unusable ( 1mark)

Way 1 - Password storage:
• Security (1 mark)
• Hashing can be used to securely store user passwords (1 mark)
• prevents unauthorised access if data is breached (1 mark)
• When a user logs in, the entered password is hashed and compared to the stored hash (1 mark)

Way 2 - Indexing / fast lookup:
• Hashing can be used to create a hash table / index for quick data retrieval (1 mark)
• A hash function converts a key (e.g. student ID) into an index/memory location for fast lookup rather than searching through all records (1 mark)

Other valid ways:
• Data integrity checking - hashing records to detect if data has been tampered with or corrupted
• Unique identifier generation - creating unique references for records

Collision identification:
• All three keys (24, 54, 84) produce the same hash index: 24 MOD 10 = 4, 54 MOD 10 = 4, 84 MOD 10 = 4 (1 mark)
• This means all three keys collide at index 4 (1 mark)
• Collisions must be resolved- only 1 value can be stored at each index (1 mark)
Linear probing:
• With linear probing, 24 is placed at index 4, 54 is placed at index 5, and 84 at index 6 (1 mark)
• This causes clustering, where consecutive slots are filled, degrading performance for future insertions and lookups (1 mark)
• Searching for key 84 would require checking index 4, then 5, then 6 - meaning O(n) lookup instead of O(1) (1 mark)
Chaining:
• With chaining, all three keys are stored in a linked list at index 4 (1 mark)
• This avoids clustering but searching within the chain is O(n) for the number of items in the chain (1 mark)

Evaluation:
• Both methods degrade from O(1) to O(n) performance due to collisions (1 mark)
• The hash function is poor for this data set because all keys share the same last digit (1 mark)

Max 6 marks

Symmetric encryption description:
• Symmetric encryption uses a single/same key to encrypt and decrypt data (1 mark)
• Both the sender and receiver must share this key (1 mark)

Risk of symmetric for transmission:
• If the key is intercepted during transmission, all encrypted data can be decrypted (1 mark)
• There is no secure way to share the key over an insecure channel without risk of interception (1 mark)

Asymmetric encryption description:
• Asymmetric encryption uses a pair of keys: a public key for encryption and a private key for decryption (1 mark)
• The public key can be shared openly without compromising security (1 mark)

Why asymmetric is more secure for transmission:
• Only the private key can decrypt data encrypted with the public key, and the private key is never transmitted (1 mark)
• Even if the public key and encrypted data are intercepted, the data cannot be decrypted without the private key (1 mark)

Evaluation:
• For internal backups, symmetric is suitable because the key can be stored securely without transmission (1 mark)
• For external transmission, asymmetric removes the key distribution problem, making it more secure (1 mark)

Max 6 marks

Asymmetric encryption for connection establishment:
• Asymmetric encryption is used initially to securely exchange a symmetric session key (1 mark)
• The client encrypts the session key using the server's public key, so only the server can decrypt it with its private key (1 mark)
• This solves the key distribution problem as the symmetric key can be shared securely (1 mark)

Symmetric encryption for session data:
• Once the session key is securely exchanged, symmetric encryption is used for the rest of the communication (1 mark)
• Symmetric encryption is significantly faster than asymmetric encryption, making it suitable for large volumes of session data (1 mark)
• This is important for e-commerce where many transactions and data exchanges occur during a session (1 mark)

Justification of hybrid approach:
• Using only asymmetric would be too slow for real-time data exchange due to computational overhead (1 mark)
• Using only symmetric would be insecure because there is no safe way to share the key initially (1 mark)
• The hybrid approach combines the security of asymmetric key exchange with the speed of symmetric encryption (1 mark)
• This is the basis of protocols like TLS/SSL used in HTTPS (1 mark)

Max 6 marks

RLE (Run Length Encoding) effectiveness:
• RLE replaces consecutive repeated values with a count and the value (1 mark)
• The large block of identical blue pixels would compress very well with RLE, as many identical pixels can be represented as a single count-value pair (1 mark)
• However, the detailed landscape section would have many different pixel values with few consecutive repeats (1 mark)
• RLE could actually increase the file size for the landscape section, as each unique pixel still needs a count value stored alongside it (1 mark)

Dictionary Coding effectiveness:
• Dictionary coding builds a dictionary of patterns/sequences and replaces repeated patterns with shorter codes/references (1 mark)
• The blue pixel block would compress well as the repeated pattern would have a single dictionary entry (1 mark)
• The landscape section may contain some repeated colour patterns or sequences that could be added to the dictionary (1 mark)
• Dictionary coding would likely be more effective overall for the landscape section as it can identify repeated patterns even if they are not consecutive (1 mark)
• Dictionary has an overhead because the dictionary must be stored alongside the data (1 mark)

Analysis:
• RLE is most effective for the uniform blue block but ineffective for the detailed landscape (1 mark)
• Dictionary coding provides more consistent compression across both sections of the image (1 mark)
• A combination of both methods, or dictionary coding alone, would likely give the best overall compression for this image (1 mark)

Max 6 marks

RLE analysis:
• RLE works by replacing consecutive repeated characters/values with a count and the character (1 mark)
• Text rarely contains long runs of the same character, so RLE would be largely ineffective for a technical manual (1 mark)

Dictionary Coding analysis:
• Dictionary coding builds a table of repeated patterns/phrases and replaces them with shorter codes (1 mark)
• A technical manual with many repeated long phrases is ideal for dictionary coding, as each phrase only needs to be stored once in the dictionary (1 mark)
• Subsequent occurrences of the phrase are replaced with a short reference/index to the dictionary entry (1 mark)

Evaluation:
• Dictionary coding would give significantly better compression for this scenario (1 mark)
• The longer and more frequently repeated the phrases are, the greater the compression ratio achieved by dictionary coding (1 mark)
• RLE is designed for consecutive repeated values (e.g. pixels in images) and is not suited to text with varied characters (1 mark)
• Dictionary coding requires more processing to build the dictionary (1 mark) and must store the dictionary alongside the data (1 mark)

Max 4 marks