Skip to content

Concerto-D/experiment_files

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

89 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Directories

Parameters

Transitions times files

Contains the random draws of transitions times for the synthetic use-case. Located in parameters/transitions_times.
The file name describes the content, i.e.: transitions_times-<min_duration>-<max_duration>-deps<nb_deps>-<draw_number>.json For the ICSOC, 2 draws 0 and 1 have been generated.

Transitions time generation

The random generation of the transitions times is done in the file transitions_times_generation.py. The parameters of the generation are put directly in the file, and the distribution follows a lognormal distribution.

Uptimes files

Contains the pseudo-random generated uptimes for the synthetic use case. Located in parameters/uptimes.
The file name describes the content, i.e.: uptimes-<nb_uptimes>-<duration>-<nb_deps>-<min_overlap>-<max_overlap>.json For the ICSOC, 4 uptimes corresponding to 4 overlap rates have been generated: 2-5% (0_02-0_05), 20-30% (0_2-0_3), 50-60% (0_5-0_6), 100% (1-1).

Uptimes pseudo-random generation

In all these scripts, the term dependency is used for all dependencies and for the server (which is the 0th dep).

The script uptimes/uptimes_generation_overview.py gives a graphic overview of an uptime generation for a given overlap. But the resulted scripts are not used for the experiment. It is only to preview the aspect of the different possibles overlaps rates.

The script that pseudo-random generate uptimes that are used for the experiment is uptimes/uptimes_generation_logic.py.
To use it:

  • Change the parameters inside the script (round: total nb of times to wake up, duration: duration of a wake up, nb_deps: total number of deps only, the server is considered implicit and is always here, overlap_taux: overlap rate, nb_generations: how many generation do you want)
  • Call the script with generate: python3 uptimes_generation_logic.py generate. It will generate as many files as <nb_generations>. These files contain the desired overlaps duration for each dependency, for each round.
  • Call the script with see: python3 uptimes_generation_logic.py see <file_name>. To have informations about a generated file.
  • Call the script with compute: python3 uptimes_generation_logic.py compute <file_name>. It will compute the actual uptimes for each dependency and for the server, by respecting the desired overlap rate for dependency and for each round. The desired overlap rate is written in on of the previously generated files, reference by <file_name>.

Results

The raw_results folder contains all the results of the experiment that was conducted for the ICSOC 2022. These results are not ordered.

The global_results_4_draws.json are the result of the execution of the script compute_globals_results.py on raw_results. In order to have the same amount of draws for every results, draws in excess are removed from the results (some results have 5 to 8 draws). However, removing these draws do not have a significant impact on the resulting std.

In order to get the full results, run the script python3 compute_globals_results.py.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published