1. Cryptography This week we are going to use OpenSSL
We will begin to introduce some cryptographic concepts
Cryptography takes a long time to learn properly
There are many idiosyncrasies that must be considered (and we will not cover)
Just to compound the idiosyncrasies: They cover many topics from algorithm implementation to use…
This lab will not introduce everything - I find the topic immensely interesting and I hope you do as well
Further Reading (not expensive): Bruce Schneier’s Applied Cryptography
You will compare unencrypted data with encrypted data
You are going to evaluate Code Books for fault tolerance.
You will also look for patterns in encrypted material

2. Analysis Tools Available to you
C Programming language
Spreadsheet that can read a CSV format
Data Input, Encrypted Material, Date Output
Histograms
You know how to write text files – add commas and newlines where necessary
Comma-separated Values (CSV)
Generate plots to illustrate what you deem important
A tool for visualizing images
IJ can help with that
Do you need anything else?
Statistics is helpful, although you will not need for this assignment

3. Applied Cryptography Principles
What properties cryptography principles seem important?
What if you see patterns in encrypted material?
Those patterns could have interesting occurrences
Does the encrypted material change greatly if the input data is changes by a small amount?
What about the same data encrypted twice – does it look the same? Is it still “encrypted” if it is?
What about the time it takes to decrypt data?
Is faster always better?
Much more needs to be considered, but you are already equipped to analyze the topics above.

4. Actual Cryptography run Time
How would you analyze how long an algorithm takes?
What cryptography steps are the most important relative to time?
What steps do you think are necessary?
Where do you need to call the clock?
What math will you need to perform?
You should analyze timing information as a function of the input length.

5. Ciphers Block Ciphers: We are going to be looking here Stream Ciphers
DES: Developed back 70’s; people are starting to get away from vanilla DES.
IDEA: Designed in 1991 as a DES replacement. Algebraic weaknesses have been discovered.
AES: NIST adopted in Considered secure.
Many More…
Stream Ciphers
RC4: Created in 1997, should be avoided…
Many more…

6. Code Books
Typically used for block cipher algorithms – some can make a block cipher look more like a a stream cipher
Electronic Code Book (ECB)
Cipher-Block Chaining (CBC)
Cipher Feedback (CFB)
Output Feedback (OFB)
Initialization Vector (IV): block of bits to randomize the encryption and produce distinct cipher texts even if the same plaintext is encrypted multiple times

7. Avalanche Effect
For every input bit that is changed the encrypted output should change dramatically.
(Applied Math) Deviations and variances between Encrypted data should be high

8. Electronic Code Book (ECB)
Really simple design
The message is divided into blocks and each block is encrypted separately.

9. Cipher Block Chaining (CBC)
Starting to get more interesting.
IV is XORed with the input
What does an error in the encrypted material do to the output (post decryption)?

10. Cipher Feedback (CFB)
This code book doesn’t even encrypt the plaintext!
This looks like it could save a small amount of time in certain circumstances
What happens if the encrypted material is corrupted?

11. Output Feedback (OFB)
Okay this code book doesn’t run plaintext through the cipher at all…
Does it save time? This algorithm could be run in parallel in ways the others could not!
What happens if there is an error in the encrypted material?

12. Pulling it together
Good Cryptography means good algorithms, code books, keys, and IV
Output Feedback (OFB) does not work well with a bad key (Good IV is also not enough)
Idea does has since been removed from OpenSSL
Can you find a weak key that will give you similar results?

13. What do you need to do for the lab?
This Lab: two components that are not graded equally
Practical component
implementing the concepts discussed
Graded as 1 lab
Documentation Component
Describing what you are seeing with graphics you generated
Graded as 2 labs
You cannot complete the documentation without the code
We have not discussed histograms
Histograms should be an array that is as large is the dynamic range
The data type should be large enough to not roll over
You can write it out as a CSV file and use bar plot in excel
I’m not saying you need to use a bar plot, bar plots are just traditional

14. The analysis has three components…
Part 1: Analyze code books to see what happens if an encrypted material bit is corrupted
Part 2: Analyze the run time of specific code books
How does it change relative to the number of blocks passed in?
Part 3: Analyze encrypted spatial patterns
This should be done with imagery and histograms
So…
Verify you are writing out the correct amount of data
Size of data in / Block Size should not have a remainder
Why is this important?