[tahoe-dev] encoding times / power costs on small devices

[Johannes Nix]

Hello Shawn,

excuse the delay, I had a bit of stomach flu the last days :-P

> > Yes, I see that encryption times don't really matter on a desktop
> > CPU.
> 
> 
> Are you sure they matter on a NAS box?  If you haven't measured them,
> it would be a good thing to do.  You might be surprised how fast they
> are.


Well, on the DNS-323 NAS (which has a 500 MHz Marvell Orion 88F5182 CPU)
the time for encoding displayed under "recent uploads" is commonly about
10 % of network upload times, which is arguably not much.

However, there are at least two more aspects: 

1) encoding times could affect network response latencies, if
lengthy operations block progress of server responses.

2) on smartphones and similar devices, CPU speed itself is much less
a concern than power consumption. Because waiting idle on network
responses does not consume much power, the relative cost of encoding can
be way larger than the fractional time.

But in order to guess less and know more, I think it's really 
better to make some benchmarks and measurements as Zooko did suggest.

Regards,

Johannes


On Fri, 16 Mar 2012 07:30:15 -0600
Shawn Willden <shawn at willden.org> wrote:

> On Fri, Mar 16, 2012 at 1:19 AM, Johannes Nix <Johannes.Nix at gmx.net>
> wrote:
> 
> > Hello Brian,
> >
> > > Upload and download some large files and then look at the "Recent
> > > Upload/Download Status" pages for them (linked from the Welcome
> > > Page): that should give you some idea of how much time is spent
> > > for various operations. Zfec encoding and AES encryption are
> > > pretty minimal: they run at 10s or 100s of megabytes per second,
> > > at least on regular laptop/desktop CPUs. SHA256 hashing is also
> > > really fast.
> >
> > Yes, I see that encryption times don't really matter on a desktop
> > CPU.
> 
> 
> Are you sure they matter on a NAS box?  If you haven't measured them,
> it would be a good thing to do.  You might be surprised how fast they
> are.
>