[clug] how long to encrypt a 1T exernal disk using LUKS

jhock jhock at iinet.net.au
Mon Jun 28 16:47:28 MDT 2010


Hi All,
Thanks for the prompt replies and suggestions. The encryption finally
finished last night. Thank goodness I only have to do that once.
Hopefully. ;--)

John

> steve jenkin <sjenkin at canb.auug.org.au> writes:
> > jhock wrote on 28/06/10 8:35 AM:
> >>
> >> I have an eeepc 10HD with Ubuntu 10.0 and Western Digital 1 terabyte
> >> external hard disk. I started encrypting the WD disk yesterday using
> >> gdecrypt and luks. Does anyone know how long this could take. It is
> >> still encrypting this morning. So far top tells me the kcrypt command
> >> has been running for "1075:00". I expect that it minutes.
> >
> > A useful benchmark for crypto performance is ssh/scp.
> 
> It is?
> 
> > [Which is RSA/DSA to start, then uses lighter symmetric keys, IIRC]
> 
> All sane people and protocols do, both because asymmetric encryption is so
> slow, and because it is so relatively weak compared to symmetric
> encryption.[1]
> 
> > I've rarely seen anything that pushes 100Mbps (~10MB/s).
> 
> Relatively modest, old hardware can comfortably hit that target; my old core2
> 2.6GHz system can manage a flat 125 megabytes (yes, *bytes*) a second of
> AES-256-IGE 8K blocks, and 138 of AES-256-CBC 8K blocks.  On a single core.
> 
> > Without evidence/knowledge, I'm going to guess the Atom does crypto like
> > a SPARC 300Mhz: ~10Mbps (1MB/sec == 3.5Gb/hour == 200-300 Hrs/Tb)
> 
> About one sixth the speed of my core2 machine, IIRC, so around 20MB a second
> or so for large-block encryption.  Roughly.
> 
>         Daniel
> 
> Footnotes: 
> [1]  See also 1K key lengths being weak in asymmetric encryption now, while
>      256 bit symmetric keys remain strong.
> 
> -- 
> ✣ Daniel Pittman            ✉ daniel at rimspace.net            ☎ +61 401 155 707
>                ♽ made with 100 percent post-consumer electrons




More information about the linux mailing list