Re: [Patch] tcpdump probabilistic sampling

[Milosz Marian Hulboj <mhulboj () hulboj org>]

On Wednesday 02 April 2008, Jesse Kempf wrote:
> I meant to say, "What does deterministic sampling give you that  
> random sampling does not?"

Hello,

Deterministic sampling is trivial to implement (but it was rather an issue 
in the hardware, in the past). It is vulnerable to bias if the metric being 
measured itself exhibits a period which is rationally related to the 
sampling interval. Systematic sampling can be easily anticipated, hence is 
open to susceptible to manipulation.
In some cases it can provide better estimates that other sampling methods 
(pointers to the literature in the document).

It is good to have different methods of sampling available for various 
experiments...

Link:
https://openlab-mu-internal.web.cern.ch/openlab-mu-internal/openlab-II_Projects/SamplingReport.pdf

Cheers,
Milosz

-- 
Milosz Marian Hulboj
http://www.linkedin.com/in/mhulboj

Attachment: signature.asc
Description: This is a digitally signed message part.