 |
 |
|
|
MISTY has a 128-bit encryption
key, a length that ensures its safety against the exhaustive
key search method of cryptanalysis.
Table 1. Computer function and cost of deciphering an encryption
algorithm within a year.
| Length
of Encryption Key |
1995 |
2000 |
2005 |
| 56 bits |
$64,000 |
$16,000 |
$2,000 |
| 64 bits |
$16 million |
$4.1 million |
$510,000 |
| 128 bits |
$3.0 x10 |
$7.5 x10 |
$9.4 x10 |
|
|
|
 |
|
|
Differential cryptanalysis
is a chosen, plain-text attack applied to DES encryption and
proposed by Biham and Shamir in 1990. It was the first method
faster than the exhaustive key search.
MISTY is provably safe against differential cryptanalysis.
Table 2 utilizes differential characteristics probability and
average differential probability to present DES’s and
MISTY's strength against differential cryptanalysis.
Table 2. Encryption strength against differential cryptanalysis.
| Algorithm |
Differential Characteristics
Probability |
Average Differential
Probability |
| DES |
2 |
Unknown |
| MISTY |
Below 2 |
Below 2 |
|
Differential characteristics probability summarizes encryption
strength against differential cryptanalysis. Encryptions that
remain secure against differential cryptanalysis have small
differential characteristics probability rates. Table 2 shows
that MISTY's differential characteristics probability is much
smaller than DES’s.
Average differential probability presents precise encryption
strength against differential cryptanalysis. Strong encryptions
possess a low average differential probability. As seen in Table
2, the DES average differential probability is unknown, whereas
MISTY’s average differential probability is extremely
low.
Because average differential probability is a summary and differential
characteristics probability is a set, encryption security against
differential cryptanalysis is not necessarily assured. This
is because the maximum differential characteristics probability
is merely low. With MISTY, the maximum average differential
probability is substantially low, or what is known as provably
secure, making MISTY the first realistically safe encryption.
|
|
 |
|
|
Linear cryptanalysis is a
known plain text attack applied to DES encryption and introduced
by Mitsubishi Electric's Mitsuru Matsui in 1993. It is the most
powerful method of deciphering an encryption, and with it we
became the first to succeed in deciphering DES using a computer.
MISTY, however, is provably secure against the linear cryptanalysis.
Table 3 utilizes linear characteristics probability and average
linear probability to present DES’s and MISTY's strength
against linear cryptanalysis.
Table 3. Encryption strength against linear cryptanalysis.
| Algorithm |
Linear Characteristics
Probability |
Average Differential
Probability |
| DES |
1.5 x 2 |
Unknown |
| MISTY |
Below 1.0x2 |
Below 1.0x2 |
|
Linear characteristics probability summarizes encryption strength
against linear cryptanalysis. Encryptions that remain secure
against linear cryptanalysis have small differential characteristics
probability rates. Table 3 shows that MISTY's linear characteristics
probability is much smaller than DES’s.
Average linear probability presents precise encryption strength
against linear cryptanalysis. Strong encryptions possess low
average linear probability. Table 3 reveals that DES’s
average linear probability is unknown, whereas MISTY’s
average linear probability is extremely low.
Because average linear cryptanalysis is a summary and linear
characteristics probability is a set, an encryption’s
security against linear cryptanalysis is not a given. The reason
for this is that the maximum linear characteristics probability
registers only as “low.” MISTY, however, demonstrates
a substantially low maximum average linear probability to what
is called a provably secure degree. This makes MISTY the first
realistically secure encryption. |
|
| |
 |
|
