Re: Block cipher modes of operations?

From: Allan Herriman (allan.herriman.hates.spam_at_ctam.com.au.invalid)
Date: 05/17/04 

Date: Mon, 17 May 2004 16:24:59 +1000

On 16 May 2004 21:31:13 -0700, ggr@qualcomm.com (Gregory G Rose)
wrote:

>In article <6fega0pcm993bh1r1aemjfsp7a25aubt55@4ax.com>,
>Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid> wrote:
>>For me, the major advantage of CTR mode is that it can be run on
>>parallel hardware, e.g. E(IV + i), E(IV + i + 1), E(IV + i + 2), etc.
>>can all be calculated at the same time, giving throughput faster than
>>any other mode.
>
>Do you have such a piece of hardware? Very few
>people do.

Yes, I have such a piece of hardware.

>While I like stream ciphers (and hence Counter
>Mode) I find this argument uncompelling.

Modes other than CTR become difficult above a few Gbps using cheap
technology. If you need to go faster, the argument becomes
compelling.
As you suggest, most users do not need that sort of speed.

Regards,
Allan.

Relevant Pages

  • Re: Block cipher modes of operations?
    ... Allan Herriman wrote: ... I have such a piece of hardware. ... break it into X pipeline stages you can process X counters at the same ... ;-) Most users of CTR mode ...
    (sci.crypt)
  • Re: Block cipher modes of operations?
    ... You have a per-session unique IV which is used to ... >create a key stream. ... the major advantage of CTR mode is that it can be run on ... can all be calculated at the same time, giving throughput faster than ...
    (sci.crypt)
Quantcast