[SLL] WRT54GL --> wireless to wired repeater?

[Robert Woodcock]

On Thu, Dec 25, 2008 at 10:04:47PM -0500, Jarod Wilson wrote:
> It'd require some investigation into the computational complexity of
> AES vs. blowfish. Offhand, I have no idea how they compare.

They're not too different - you can benchmark OpenSSL's implementation of
them just by running "openssl speed".

My results on an Athlon XP 2500+ are:

The 'numbers' are in 1000s of bytes per second processed.
type             16 bytes    64 bytes    256 bytes   1024 bytes  8192 bytes
[...]
des cbc          42910.58k   45357.53k   45967.87k   46149.29k   46046.66k
des ede3         15559.86k   15860.45k   15985.38k   15985.36k   16012.63k
[...]
blowfish cbc     67574.89k   73725.16k   75315.20k   75577.36k   75986.26k
[...]
aes-128 cbc      44625.12k   69858.65k   81771.67k   85795.84k   86955.35k
aes-192 cbc      40406.79k   61185.22k   70736.93k   73755.99k   74711.04k
aes-256 cbc      37157.79k   54430.68k   62146.30k   64428.37k   65120.94k

Note that 3DES is much, much slower than both.

A couple other factors to consider:

* You are far more likely to see AES hardware acceleration support than
  Blowfish support, especially in low-power processors that could use it,
  such as the AMD Geode LX and the VIA C3/C5/C7 (Soekris net5501, all
  manner of mini-ITX boards)

* It's probably safe to say that AES has been vetted more comprehensively
  than Blowfish (although it's a fair bet that both will age gracefully.)
-- 
Robert Woodcock - rcw at blarg.net
"When failure is not an option, success can get expensive."
	-- Peter Stibrany