Statistics
| Branch: | Revision:

iof-tools / mrai_setter @ 62798f6e

Name Size Revision Age Author Comment
Readme.md 1.47 KB 0a478326 5 months tiamilani possibility to plot multiple cdf on the same pl...
milaniBGPLoad.py 4.86 KB c1397947 5 months tiamilani corrected DPC calculation, based on which node ...
mrai_setter.py 12.6 KB 62798f6e 4 months tiamilani intermidiate state
test_mrai_setter.py 8.22 KB 62798f6e 4 months tiamilani intermidiate state

Latest revisions

# Date Author Comment
62798f6e 11/26/2019 01:40 PM tiamilani

intermidiate state

cfe4c6c3 11/12/2019 10:52 AM tiamilani

addresses bug fixed

fb99df25 11/12/2019 10:50 AM tiamilani

Possibilit√° di disabilitare l'mrai medio impostandolo a 0

0a478326 11/11/2019 08:32 AM tiamilani

possibility to plot multiple cdf on the same plot, implemented the mean_mrai on mrai_setter

c1397947 10/31/2019 03:14 PM tiamilani

corrected DPC calculation, based on which node share a network and not on the type of the node, there could be for example nodes of Type M that does not share destiantions, this before whas a problem because every node different than a T was sharing a contribute, but this should happen only if they really share one internal network

28c0971b 10/25/2019 12:19 PM tiamilani

new mrai settings, test done

2a73c73a 10/02/2019 08:34 AM tiamilani

Fixed mrai setter, customer swapped

55aabef7 09/17/2019 02:21 PM Luca Baldesi

fix compatibility bug on customer attribute type (from integer to string)

3a4e2b49 09/13/2019 02:05 PM Luca Baldesi

add MRAI setter code

View revisions

README


MRAI setter

This software takes in input a graphml annotated according to the specifications of IoF and set the node MRAI values.

Usage

./mrai_setter.py <graphml_file> <strategy> <outputDir> <mean_mrai> [<advertising_node>]

The output dir will be created if it does not exist

The optional option indicates which node is advertising. This can be of paramount importance for some strategies (e.g., Fabrikant gadgets).

Strategies

Strategies are MRAI setting policies. At the time of writing available strategies include: * 30secs: set all timers to 30 seconds * none: set all timers to 0 seconds * fabrikant: set timers according to worst case gadget configuration (see paper) * inversefabrikant: inverted timers of the previous case (should lead to good case) * milanicent: set timers according to our theorecal derived model based on Milani centrality (mice) * milanicent2: variation of the previous one, with a different normalization factor * uniformdistrmrai: Set timers randomly following a uniform distribution btween 'defaultmrai'%'percentageconstant' and defaultmrai * constantfabrikant: Set timers following Fabrikant policies, but with a constant increment, the constant percentage is given by 'percentageconstant' * constantinversefabrikant: Set timers following inverse Fabrikant polices, but with a constant decrement, the constant percentage is given by 'percentage_constant'

Tests

Just type python3 -mpytest

Also available in: Atom