Anuket Project

2021-07-02 AI/ML for NFV Meeting Minutes


Attendees

Sridhar Rao

Rohit Singh Rathaur

Girish

Al Morton

Ildiko


Sl. No.TopicPresenterNotes
1Request to create gitlab repo for this incubation project
Sridhar Rao to initiate on 6th of July.
2Presentation to EUAG

EUAG: End User Advisory Group.

6th of July - Present the plan, and request for support w.r.t Data.

3Framework Deployment Status

LF Acumos deployment is not stable - Ended up wasting too much time on it.

Move the "framework" issue to later phase - need volunteer(s) to get the issue with these frameworks deployment fixed.

Work independent of these framework - Python and ML-Libraries – Jupyter Notebooks.

Pod12 - Jumphost. Install Ubuntu 20.04 LTS on Jumphost.

4Model Deployment Status

RNN and LSTM – VM failure.

Data: https://github.com/amasend/SDN_with_Neural_Network/tree/master/data  - CPU

5Summary of "Gaps" in Existing Failure prediction works
  1. Container Failures are not considered.
  2. Node level failures are not considered.
  3. Application level failures are not considered.
6Model Enhancement Status

Plan will be based on - 5.

 Create separate models for each of the failures 

Girish/Rohit would like to share ideas on Model Enhancement Next week.

  • Node-Level .
  • VM - Current focus.
  • Container/Pod
  • Application/Workload
7Data Status

Sridhar Rao to send mail to Ildiko - explaining the requirements.


8. Failure Prediction Definition - Status
  • Classification of Failures
  • Mapping data and failures  - ToDo - Node and VM (Rohit), Container and App to Girish.
  •  Models and Failure category mapping.