Encryption and Compression

Encryption and compression are interrelated. Compression can be used legitimately to
improve delivery and storage bandwidth utilization when transmitting or storing
voice, text, images, or video content. Compression or steganography (for example,
voice files embedded in image files) can be used nonlegitimately as a form of encryption.
Highly compressed files have high entropy. It can be hard to distinguish whether
files are encrypted or heavily compressed.
We said earlier that compression ratios increase by an order of magnitude every 5
years. This is the result of additional processor and memory bandwidth. Similarly, the
gap between encryption and cryptoanalysis increases with time, that is, encryption
distance (and, potentially, encryption value) increases with time. This is because
increasing computing/processing power confers greater advantage to the cipher user
because longer key lengths take polynomially more time to use but exponentially more
time to break. Ten years ago, a 56-bit length key would have been considered secure.
Today, RSA encryption typically uses 1024- or 2048-bit keys.
Encryption performance can be measured in terms of time to break in MIPS/years
(how much computing power in terms of million of instructions per second × how long
it would take theoretically to break the code). An RSA 1024-bit key takes theoretically
1011 MIPS/years to break. A 2048-bit key takes 1020 MIPS/years to break.

So as processor bandwidth increases, it becomes possible to use progressively
longer keys. However, as keys get longer, encryption/decryption delay increases,
processor overheads and power consumption increases, and code and memory space
occupancy increases.