Comparing Cryptosystems 1. 2. 3. 4. 5. Among Rotation, Affine, and Autokey, which is most resistant to a known-plaintext attack? Among Affine, Autokey, and Vigenere, which can encrypt the fastest? Among Affine, Autokey, and Vigenere, which can decrypt the fastest? Which has a larger keyspace? Affine over or Vigenere over ? Which has a larger keysize? Substitution over or rotation over

Comparing Cryptosystems Which is more secure? Affine or a Vigenere ciphers? 7. Which is more secure? Vigenere or Hill ciphers? 8. Could a computer with 4 processors perform 200 Rotation encryptions or 100 Autokey encryptions faster? 9. Could a computer with 4 processors perform 200 Rotation encryptions or 100 Affine encryptions faster? 10. Which has a larger keyspace? Hill over or Vigenere over ? 6.

Substitution Cryptosystems How many possible keys does an affine cipher on have? 7. Encrypt using a rotation cipher over with . 8. Encrypt using an affine cipher over with 9. Cipher X consists of a rotation, and then an affine cipher. What type of cipher is X? 10. Cipher Y is a substitution cipher over . Cipher consists of applying Y twenty-four times. What type of cipher is X? Be as specific as possible. 6.

Decryption Encryption P1 P2 b K K Encrypt b K Encrypt b Encrypt b C1 C2 C1 C2 b K Pn b b Cn Cn b K Decrypt b K Decrypt b Decrypt b P1 b P2 Pn (a) Block cipher encryption (electronic codebook mode) Key K Key K Pseudorandombyte generator (key streamgenerator) Pseudorandombyte generator (key streamgenerator) k Plaintext bytestream M ENCRYPTION k Ciphertext bytestream C DECRYPTION (b) Streamencryption Figure2.2 Types of Symmetric Encryption Plaintext bytestream M

Isolate 4 CONCLUSIONS BACKGROUND Acinetobacter spp. Has been documented as the cause of numerous nosocomial outbreaks and has been associated with increased lengths of stay in the ICU as well as increased mortality in infected ICU patients. In our institution the number of isolates cultured nearly doubled from 1998-2000 compared to 2001 with over one-third being multi-drug resistant. Susceptibility to commonly used antibiotics including cefepime piperacillin/tazobactam, and ampicillin/ sulbactam decreased while imipenem/cilistatin and tobramycin susceptibility remained stable. Antimicrobial resistance makes drug selection difficult. Inappropriate antimicrobial selection can lead to increased mortality, while exposure to broad spectrum antimicrobials is associated with development of resistance. Initial selection of antimicrobials at optimal dosing may have a significant effects on patient outcome antimicrobial resistance. Carbapenems alone or in combination with other antimicrobials are considered to be a therapy-ofchoice for Acinetobacter infections. The purpose of this study was to compare the bactericidal effects of imipenem/cilistatin (IC) and meropenem (M) against AC clinical isolates using an in vitro PK bacteremia model. Meropenem 3/4 runs Meropenem 1/4 runs Imipenem 8 7 6 5 4 Table 2. MIC/MBC for Meropenem and Imipenem Against Isolate 4 After Exposure to Meropenem 3 2 1 0 2 4 6 8 10 1224 30 36 42 48 MIC/MBC (mcg/ml) 48 hr. 24 hr. 48 hr. antibiotic 48 hr. Isolate Model Drug Baseline drug-free model plate drug-free plate sample (4x plate MIC) Meropenem 1/1 1/>16 >16/ND 128/ND 128/ND A Imipenem 0.5/0.5 1/1 >16/ND >128/ND >128/ND 4 Meropenem 1/1 NG NG NG NG B Imipenem 0.5/0.5 NG NG NG NG Time (hrs) Isolate 7 METHODS Organisms/Susceptibilities. Four AC clinical isolates were tested. Baseline data collected for M an IC against isolate included: MIC/MBC by microtiter methodologies following NCCLS protocol. Ppresence of resistant subpopulations on antibiotic containing plates representing 2x and 4x the MIC. Frequency of resistance = the number of organisms growing on the antibiotic-containing medium divided by number of antibiotic-free agar plates. Meropenem 5/6 runs Meropenem 1/6 runs Imipenem 8 7 6 5 4 3 2 1 0 In vitro pharmacokinetic/pharmacodynamic model. Isolates were tested over 48 hours against M and IC in a glass one compartment in vitro PK model emulating a bloodstream infection. Drug regimens and human PK parameters to be simulated include: 1) Imipenem/cilistatin 500 mg every six hours simulating a peak concentration of 40 mcg/ml and a half-life of 1 hour. 2) Meropenem 1 gram every 8 hours simulating a peak concentration of 60 mcg/ml and a half-life of 1 hour. 2 4 6 8 10 1224 30 36 42 48 Table 3. MIC/MBC for Meropenem and Imipenem Against Isolate 7 After Exposure to Meropenem MIC/MBC (mcg/ml) 48 hr. 24 hr. 48 hr. antibiotic 48 hr. Isolate Model Drug Baseline drug-free model plate drug-free plate sample (4x plate MIC) Meropenem 1/1 2/2 >16/ND 4/64 128/ND C Imipenem 0.5/0.5 1/1 >16/ND 1/32 >128/ND 7 Meropenem 1/1 2/2 >16/ND 4/64 4/32 D Imipenem 0.5/0.5 1/1 >16/ND 0.5/32 1/16 Time (hrs) Resistant subpopulations were seen in all isolates for IC and M (Table 1). Both agents provided bactericidal killing against 3/4 isolates. Regrowth occurred earlier for M in 3/4 isolates with IC maintaining killing at or below 2 log for ≥ 30 hours in 2/4 isolates. Despite regrowth occurring, no changes in MIC/MBC or growth on antibiotic-containing media were noted for isolates 9 and 10 for either M or IC. In contrast, MIC/MBC changes did occur with isolates 4 and 7. Tables 2 and 3 describe examples of these changes. For isolate 4, selection of a resistant clone likely occurred leading to regrowth in 3/4 model runs. High level resistance was seen when sampling directly from the model as well as off the antibioticcontaining medium. One of 4 runs did not regrow and may be attributed to lack of selection of a resistant clone. MIC/MBC changes occurring in isolate 7 are more difficult to explain. Selection of a resistant clone explains the regrowth occuring in 5/6 runs. No change in MIC/MBC occurred when evauating growth from the 24 hr drug-free plates. Induction of resistance with the addition of the antimicrobial likely occurred followed by reversion back to the susceptible form with the removal of drug pressure. Highly resistant MICs at 48 hours from samples from the model are expected. Tolerance, low MICs and elevated MBCs, is seen when evaluating growth from the 48 hr antibiotic-containing medium as well as with one run of the 48 drug-free plates. A second example (model C) at 48 hrs. maintained high level resistance despite removal of drug pressure. M and IC had similar rates of killing with both agents being bactericidal in 3/4 isolates. IC had superior duration of bacterial killing compared to M against AC. Despite initial bactericidal activity, significant regrowth occurred for all isolates against both drugs by 48 hours. This suggests carbapenem monotherapy may select for resistant populations or induce resistance and coul lead to potential drug failure. The presence of resistant subpopulations may support the use of combination therapy in AC infections.

Attack on a Cryptosystem Attempt to decrypt encrypted message without knowledge of the decryption key. Four basic sorts of possible attacks: Ciphertext only: Eve has only a fragment of the encrypted message: to decrypt the specific message to obtain the key and to be able to decrypt all subsequent messages Known plaintext: Eve has a copy of the ciphertext and the corresponding plaintext. The goal is to deduce the key Chosen plaintext: Eve gains the access to the encryption machine, she doesn’t have a key, but she can use the some of the plaintexts and corresponded ciphertexts, the goal is to deduce the key Chosen ciphertext: Eve gains the access to the decryption machine.

Security and performance point for the first 9 correct point for the next 20 correct 1 point for each correct answer beyond 29 (Rotation does not count) (Assume you only ever store the encryption key and never modify it for quick decryption) Key siz e Substitution over Rotation over Autokey over One-time-pad over Affine over Vigenere over Hill over Keyspa ce Encrypti on runtime Decrypti on runtime pairs needed Resistant to Frequency Analysis?

Breaking an encryption scheme • Cipher-text only attack – Trudy has ciphertext that she can analyze – Two approaches: • Search through all keys: must be able to differentiate resulting plaintext from gibberish • Statistical analysis • Known-plaintext attack – trudy has some plaintext corresponding to some ciphertext • eg, in monoalphabetic cipher, trudy determines pairings for a,l,i,c,e,b,o, • Chosen-plaintext attack – trudy can get the cyphertext for some chosen plaintext 8

Breaking an encryption scheme Cipher-text only attack: Trudy has ciphertext that she can analyze Two approaches: Search through all keys: must be able to differentiate resulting plaintext from gibberish Statistical analysis Known-plaintext attack: trudy has some plaintext corresponding to some ciphertext eg, in monoalphabetic cipher, trudy determines pairings for a,l,i,c,e,b,o, Chosen-plaintext attack: trudy can get the cyphertext for some chosen plaintext 8

Encryption! 1. 2. 3. 4. 5. Encrypt 7 using a rotation cipher over with key . Encrypt 18 over using an affine cipher with key . Encrypt 3 using RSA with and . Encrypt using an autokey cipher with starting key over . XOR the string with .

Substitution and Affine Ciphers General Substitution Cipher: To encrypt: C = (P + K) (mod 26), K is the KEY To decrypt: P = (C - K) (mod 26), K is the SAME KEY Affine Cipher: To encrypt: C = (AP + B) (mod 26), A and B are KEYS. A is relatively prime to 26 To decrypt: P = A-1 (C - B) (mod 26) A-1 is multiplicative inverse of A mod 26

DES Modes • Electronic Code Book Mode (ECB) – Encipher each block independently • Cipher Block Chaining Mode (CBC) – Xor each plaintext block with previous ciphertext block – Requires an initialization vector for the first one – The initialization vector can be made public • Encrypt-Decrypt-Encrypt Mode (2 keys: k, k) • Encrypt-Encrypt-Encrypt Mode (3 keys: k, k, k) 11

Decryption! 6. 7. 8. 9. 10. Decrypt , which was encrypted with a vigenere cipher using the key over . Decrypt which is the output of box. Decrypt , which was encrypted using the key over . Decrypt , which was encrypted cipher Solve the discrete log problem the accompanying Swith a one-time-pad using the substitution mod 100. S 0 1 2 3 0 1 1 1 2 1 3 1 4 1 1 5 1 6 1 7 1 8 2 1 2 2 2

Cryptanalysis Overview Method used is based on the amount of information Brute Force: try all possibilities Dictionary Attack: run through a dictionary of words trying to find the key or plaintext Cipher text only Chosen Plaintext: Have the ability to find the cipher text relating to an arbitrary plaintext Chosen Cipher text: can choose an arbitrary cipher text and know the plaintext Adaptive chosen plaintext: determine cipher text based on plaintext using iteration

Tricky Questions 6. 7. 8. 9. 10. A substitution cipher over has a key sorted by plaintext. During decryption, how many tablelookups are required in the worst case? A substitution cipher over has a key sorted by plaintext. 4 different characters are encrypted. How many table-lookups are required in the worst case? How many possible Vigenere ciphers over are there that can encrypt exactly 3 characters? A one-time-pad is used on a 5 character message. Oscar figured out the keys, but not the order. In the worst case, how many guesses must be try before deciphering the message? How many factors does 1,000,000 have?