Laurentiu,

I am cc-ing analytics@ public e-mail lists where you can direct this type of questions.

>for this goal we need time series of the http requests (pagecount, traces, and so on) the resolution of milliseconds.

This is really not super clear. I think you need to provide a more through description of the data you need.

For example, this describes internal data that we retain for ad most 90 days:

https://wikitech.wikimedia.org/wiki/Analytics/Data/Webrequest

Given that you need ms resolution it is not likely we might have a dataset you can use but maybe someone on this list can point you to a released dataset that might be suitable.

Thanks,

Nuria

On Fri, Jan 27, 2017 at 3:00 PM, Laurentiu Checiu <lchec097@uottawa.ca> wrote:

Dear Ms. Nuria Ruiz,

I am a PhD student at the University of Ottawa and my research is focused on the cloud computing stochastic models.

I found "Page view statistics for Wikimedia projects": http://dumps.wikimedia.org/other/pagecounts-raw and from this source I can construct time series of http requests on a hourly base. Based on these time series we can estimate a model for a cloud computing system. However, this hourly rate of the requests is not quite suitable for our intended model. We are aiming to a model able to react at level of seconds or even faster and for this goal we need time series of the http requests (pagecount, traces, and so on) at the resolution of milliseconds. We are interested only on the number of the requests on the time unit (ms) and not on the actual source or the destination of these http requests.

May I ask you for help in this matter ?

Best regards,
Laurentiu Checiu