Re: Crypto Hardware for Loop-aes?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Nov 8, 2005, at 4:28 AM, Venkat Manakkal wrote:
http://hifn7751.sourceforge.net/


I tried the Soekris card on FreeBSD; both OpenBSD and FreeBSD use a kernel-level cryptographic framework that can use these PCI hardware accelerators (hifn chips).

They are not appropriate for on-disk encryption.

I don't have the details correct, but it is a performance issue: data must transferred from main memory via the CPU to the PCI card for processing, and then the data is transferred back, and then it is written to disk.

The accelerator cards are more appropriate for network packet encryption: the network stack could tell a network interface to perform direct memory transfer to the crypto card on the PCI bus before ever hitting the CPU or the main memory. In that application, one could achieve "wire-speed" encryption of the network packets, without CPU overhead.

My experience is that for disk I/O, crypto operations of the CPU are still faster than the PCI-based crytpo accelerators.

This almost certainly does NOT apply to the VIA "padlock" crypto acceleration, which adds instructions to the x86 ABI and performs crypto operations on the CPU.

{Open, Free}BSD is a fun system to use if one is concerned about security and robustness. But I have not found a disk-encryption solution that rivals loop-aes on Linux for performance and stability.

~ boyd




-
Linux-crypto:  cryptography in and on the Linux system
Archive:       http://mail.nl.linux.org/linux-crypto/


[Index of Archives]     [Kernel]     [Linux Crypto]     [Gnu Crypto]     [Gnu Classpath]     [Netfilter]     [Bugtraq]
  Powered by Linux