Bug#750546: ITP: sluice -- rate limiting data piping tool

[Colin Ian King <colin.king@canonical.com>]

On 05/01/15 18:11, Henrique de Moraes Holschuh wrote:
> On Mon, 05 Jan 2015, Colin Ian King wrote:
>> Since my original email, I have added more features and more
>> sophisticated rate limiting features, namely:
>>
>> * constant delay time between each write (rate limiting by changing
>> buffer sizes)
>> * constant buffer sizes, (rate limiting by changing write times)
>> * rate limiting by changing buffer sizes and write times
>> * -s option to tweak rate limiting throttling
>>
>> The -s option controls the damping behaviour, which can be tweaked for
>> different kinds of variable rate inputs. Some analysis of this can be
>> seen on the sluice project page:
>>
>> http://kernel.ubuntu.com/~cking/sluice/
>>
>> I wonder if this justifies sluice being reconsidered?
> 
> Sure it does.  I can see those two rate-limiters being useful for people
> playing with network testing.
> 
> Can you also take a look at pv and see if any of its features are worthwhile
> for sluice?
> 
I've added a bunch of extra features into sluice that similar to those
found in pv and cstream, namely:

-e skip read errors,
-p progress and ETA statistics
-I input file (rather than just read from stdin)


plus:
SIGUSR1 - toggle on/off verbose mode
SIGUSR2 - toggle overrun, underrun modes

Also, sluice has a -S summary mode that gives some in-depth stats on the
streaming rates including drift from the desired target rate.

I think the various date rate control mechanisms (e.g. fixed block size,
dynamic variable block size or hybrid of both), plus the ability to
tweak these with the -s option along with the -S stats info are the
unique selling points of sluice.  Sluice was designed for steady and
accurate streaming rates for network and data copying benchmarking etc.

Colin