alphabetic (or whole character set) variable shift
consider the phrase: "THE QUICK BROWN FOX JUMPS THE LAZY DOGS BACK".
If this phrase is repeatedly used to shift characters in your message...
then the first message character is shifted by 20 (because "T" is
the twentieth
letter in the alphabet)
then the second message character is shifted by 8 (because "H" is
the eighth letter of the alphabet)
then the third message character is shifted by 5 (because "E" is
the fifth letter of the alphabet)
etc.
Alternatively, if the message is in
ASCII format
then you might wish to
XOR (exclusive OR) it with the ASCII value of the passphrase...
then the first message character is XOR'd with 84 which is the ASCII value of
"T"
then the second message character is XOR'd with 72 which is the ASCII value of
"H"
then the third message character is XOR'd with 70 which is the ASCII value of
"E"
etc.
If you hit the end of your pass phrase, just wrap around to the
beginning (therefore longer phrases introduce more scrambling).
If someone guesses your pass phrase, then they will be able to
decrypt your message
You use the same phrase to decrypt (shift back or just XOR a second
time)
Seemingly random passphrases
in essence, this is what the Germans used with their Enigma Machine.
Changing rotors (wiring), and starting settings (initialization value) produced a different code every
day. The standard machine only possessed three rotors while the Navy
version possessed four.
Transposition Basics
I won't waste too much time here exception to mention that substitution
ciphers still have a common problem: the information still contains positional
information which means cipher-text may still be subjected to
certain kinds of frequency analysis.
Transpositions change character position in order to hide positional
information thus defeating most kinds of frequency analysis.
Keys and other stuff
Symmetric Keys
used to encrypt and decrypt
this means that either party can encode or decode)
in the case of SSL, each end maintains their own private key whilst
the public key is distributed by a trusted third party like Verisign or
Thawte to only name two.
If I encrypt a message with my private key, and you then decrypt it
with my public key, then you know the encrypted message came from me
(electronic signature) To continue...
If I encrypt this signed message with your public key before I send
it to you, then I will be sure that only you can read it (with your
private key). The message is now secured and verified
in each direction.
Computational Load
Due to a reliance upon a pair of large prime numbers, asymmetric keys require much more computer power to encrypt/decrypt
than symmetric keys.
So before each communication session:
large random numbers are used to generate a set of symmetric keys (which
are known as session keys)
asymmetric keys are used to encrypt then exchange symmetric keys
symmetric keys (session keys) are then used to encrypt/decrypt the message
this method allows two parties that have no prior knowledge of
each other to jointly establish a shared secret key over an insecure
communications channel. This key can then be used to encrypt
subsequent communications using a symmetric key cipher.
Example Math (caveat:
implemented in JavaScript with usual
JS limitations so stick to small numbers)
Don't be confused by the fancy math. As far as I am concerned,
any formula could be used including simple addition or
multiplication. Consider these non-modulus examples:
Addition:
Assume: y=7
Alice chooses: A=3
Alice computes: a=(7+3)=10 which she sends to Bob
Bob chooses: B=6
Bob computes: b=(7+6)=13 which he sends to Alice
Alice now computes: key=(3+13)=16
Bob now computes: key=(6+10)=16
Multiplication:
Assume: y=7
Alice chooses: A=3
Alice computes: a=(7x3)=21 which she sends to Bob
Bob chooses: B=6
Bob computes: b=(7x6)=42 which he sends to Alice
Alice now computes: key=(3x42)=126
Bob now computes: key=(6x21)=126
So why the fancy math?
What do you want to do?
If you want to encrypt a file or hard drive
you encrypt with your
asymmetric public key (which is publically available to everyone) and will
decrypt with your private key (which is only known to you)
If you want to prove who you are on the net (a good example is
Secure DNS)
then you encrypt with your private key.
others will use your public key then say: "yes, he was the only one
who could have encrypted that message"
If Alice wants to send a message to Bob which can only be read by Bob
then Alice encrypts using Bob's public key (only he will be able to decrypt
with his private key)
this is what happens during web banking (you use a password to prove
who you are; by encrypting with your banks public key, only your bank
will be able to decrypt)
If Alice wants to send a message to Bob but also wants to prove it came
from Alice
then Alice encrypts using Alice's private key (Bob will decrypt
using Alice's public key)
Caveat: eves dropper "Eve" will also have
access to Alice's public key so this communication is not considered
secure; in fact, it is only known as a "digital signature"
If Alice wants to send a message to Bob which can only be read by Bob
but wishes to ensure that it only came from Alice
Alice encrypts using Alice's private key ("digital signature")
Alice encrypts again using Bob's public key
Bob will decrypt using Bob's private key (only he can do this)
Bob will decrypt again but this time will use Alice's public key
(which only she could have encrypted)
In SSL/TLS you would always encrypt with the other guys' public key (which
everyone has access to via a trusted third party)
In SSH/SSH2, you create a public/private pair, then copy the public key
to the remote end. This means you would always encrypt with your private
key. The far end always decrypts with your public key.
The UNIX/Linux/Enterprise Server community sort of went the way of SSH2 while the
public web went the way of SSL/TLS. Today, most systems employ both
Today, encrypting with large asymmetrical keys is too computational
intensive so they are only used to encrypt randomly generated asymmetric
session keys. Session keys are then used to drive cyphers (like
TripleDES,
BlowFish,
TwoFish, etc.)
What's in a name?
Usually, if "S" is the first character then the protocol employs
SSH/SSH2:
Steganography (hiding information
in plain sight (like inside GIFs and JPEGs) comments:
consider a totally red pixel with
the HTML color code of "255,0,0". Would your eye be able to see the
difference between this color and "254,0,0"? (probably not)
Now consider a picture in
which all the colors have be pre-adjusted to have even color numbers.
Next, a binary code (zeros and ones) could be serially inserted into the
picture just by adjusting the color of each addressable pixel (binary 0 = even
color; binary 1 = odd color).
The code would be right before you eyes but you
would not see it. But a computer could.
Back
to OpenVMS Back
to Home
Neil Rieck
Kitchener - Waterloo - Cambridge, Ontario, Canada.
OpenVMS
Notes:
Cryptography, etc.
Back
to OpenVMS
Back
to Home