Control Plan & User Plan reconciliation¶
This Atom is able to find data mobile customers usages that have experienced network dysfunctions. Customers usages are defined by the customers TCP session provided by Otarie probes, network dysfunctions correspond to suspicious signaling messages, provided by Astellia probes, that occured during a TCP session (between Start & Stop). The suspicious signaling messages are defined thanks to specific values of the three following files of astellia Raw data : 'Cause', 'Final sate' and 'LastEvent'.
|Cause||Final Sate||Last Event|
|Insuficient ressources||Access Bearer Asignment Failed||PDP context rejected|
|Missing or unknown APN||/||PDP context rejected|
|No remaining RAB||Access Bearer Asignment Failed||/|
|No ressource available||Access Bearer Asignment Failed||/|
|Network Failure||Access Bearer Asignment Failed||PDP context rejected|
The file RES provided with this Atom, located in /src, is the result of a data treatment on Raw data coming from Otarie and Astellia probes. These data have been collected between the 23rd of august 2016 and the 28th of November 2016 in the area of Lyon on the 3G network. This data treatment consists to link the records in the two database that have the same value on the three following fields : Cell_name, Time(same day, same hour, same minute) and same customerID. Then we selected only the records that contain suspicious signaling message that occured between the start and stop of the customer TCP session. These specific data treatments will be the object of a futur Atom. The result file RES is then used by R scripts, located in /src, to provide a friendly user interface that allows the user to filter the data by date, usage and type of signaling message
To use CPUP Atom in your own environment (linux or windows no matter), retrieve all the materials stored in gitlab under the repository "src" and put them in a same local directory.
The only software prerequisite is to have RStudio tool installed on your system. Once it is installed, click on "File/Open File" browse up to your local directory and click on CP_UP_reconciliation.Rproj to load all the necessary files in RStudio.
Then modify the file app.R : line 11, in read.csv2 instruction, replace the folder in quote by your own local folder where you previously stored all the files. Like that : 'you local folder/RES', don't forgot to mention the file RES. Then modify the file Launch_appR.r in the same way : line 3, in appDir instruction change the path by the path of your local directory '.
Then you have to install the following packages under R by typing this command
and finaly launch the file "launch_appR.r" (more quickly, you can have a look on http://10.192.226.15:8111/ to have a clear idea of what CPUP Atom look likes, available only under ERMES network)
Once CPUP is launched, your browser open and you can see a map where the number of suspicious messages appears in little circle, if you zoom in suficiently, there are more an more circle and finaly, with a certain level of zoom, each circle represent a nodeB with the number of suspicious message. Then you can put some filter on customer usage, type of suspicious signaling message and also on Date.